{ "schema": 1, "generatedAt": "2026-08-14T02:16:50Z", "host": { "cpu": "Apple M1 Max", "platform": "macOS-26.5.2-arm64-arm-64bit", "onnxruntime": "1.28.0", "threads": 1 }, "stt": { "id": "faster-whisper:base.en", "detail": "CTranslate2 int8 CPU, greedy, model=base.en" }, "vad": { "id": "silero-vad", "threshold": 0.5, "block_size": 512 }, "inputSiSdrDb": 10.92073222479158, "reportingNote": "siSdrDb is the absolute output; siSdrDeltaDb is improvement over inputSiSdrDb. Improvement is bounded by input headroom, so quote both.", "testset": { "version": 1, "seed": 20260101, "sample_rate": 16000, "utterances": 10, "clips": 180, "total_audio_seconds": 1406.3, "licence_note": "LibriSpeech is CC-BY-4.0. ESC-50 is CC-BY-NC-3.0 \u2014 the noise conditions are research/evaluation only. The competing-speaker and babble conditions are built from LibriSpeech alone and carry no NC restriction." }, "conditions": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10", "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "backends": [ { "id": "raw", "label": "Raw (no processing)", "license": "\u2014", "shippable": true, "rtf": null, "blockMsP99": null, "realtimeHeadroomP99": null, "latencyMs": 0.0, "siSdrDb": 10.92073222479158, "siSdrDeltaDb": 0.0, "pesqWb": 1.817549209462272, "stoi": 0.8436932337195054, "estoi": 0.7371305398229724, "wer": 0.1476608187134503, "werSubstitutions": 344, "werInsertions": 101, "werDeletions": 60, "werRefWords": 3420, "vadF1": 0.9503607044937831, "vadFalseAlarm": 0.17156970094457827, "conditions": { "clean": { "si_sdr_db": 150.0, "pesq_wb": 4.643888473510742, "stoi": 0.9999999999999998, "vad_f1": 0.9756248909923848, "vad_false_alarm": 0.07023743685253026, "wer": 0.04736842105263158, "insertions": 2, "substitutions": 6, "deletions": 1, "ref_words": 190 }, "noise_snr-5": { "si_sdr_db": -5.018447138557539, "pesq_wb": 1.104841959476471, "stoi": 0.7198304186475334, "vad_f1": 0.9284106039361377, "vad_false_alarm": 0.06810271419067943, "wer": 0.18421052631578946, "insertions": 4, "substitutions": 29, "deletions": 2, "ref_words": 190 }, "noise_snr0": { "si_sdr_db": -0.008532567221562221, "pesq_wb": 1.2508283972740173, "stoi": 0.8199818537088367, "vad_f1": 0.9650457996816915, "vad_false_alarm": 0.07585414368555106, "wer": 0.11052631578947368, "insertions": 2, "substitutions": 18, "deletions": 1, "ref_words": 190 }, "noise_snr5": { "si_sdr_db": 4.994365358650382, "pesq_wb": 1.661296832561493, "stoi": 0.8595580950060532, "vad_f1": 0.9655544660175688, "vad_false_alarm": 0.06337917413073132, "wer": 0.08947368421052632, "insertions": 2, "substitutions": 15, "deletions": 0, "ref_words": 190 }, "noise_snr10": { "si_sdr_db": 9.996620019510184, "pesq_wb": 1.8911864519119264, "stoi": 0.9162243593607545, "vad_f1": 0.9715691872453027, "vad_false_alarm": 0.0521201836224206, "wer": 0.06315789473684211, "insertions": 2, "substitutions": 10, "deletions": 0, "ref_words": 190 }, "noise_snr20": { "si_sdr_db": 19.998494087473777, "pesq_wb": 2.7327469110488893, "stoi": 0.9804213536382772, "vad_f1": 0.9722639322945248, "vad_false_alarm": 0.0769268153848264, "wer": 0.042105263157894736, "insertions": 2, "substitutions": 6, "deletions": 0, "ref_words": 190 }, "babble_snr5": { "si_sdr_db": 4.993941571017191, "pesq_wb": 1.1675394892692565, "stoi": 0.7761497290169415, "vad_f1": 0.8446580505770814, "vad_false_alarm": 0.9770890756151038, "wer": 0.24736842105263157, "insertions": 10, "substitutions": 27, "deletions": 10, "ref_words": 190 }, "competing_snr0": { "si_sdr_db": 0.03889707927885379, "pesq_wb": 1.1588235735893249, "stoi": 0.7344131574857835, "vad_f1": 0.8978559481992789, "vad_false_alarm": 0.5817642991147487, "wer": 0.8473684210526315, "insertions": 45, "substitutions": 81, "deletions": 35, "ref_words": 190 }, "competing_snr5": { "si_sdr_db": 5.022236435769729, "pesq_wb": 1.2851489901542663, "stoi": 0.8219422931414545, "vad_f1": 0.8993866413420953, "vad_false_alarm": 0.5812582262402548, "wer": 0.2631578947368421, "insertions": 19, "substitutions": 28, "deletions": 3, "ref_words": 190 }, "reverb": { "si_sdr_db": -26.909354792747905, "pesq_wb": 1.3408596515655518, "stoi": 0.6332772949104204, "vad_f1": 0.9623023734900757, "vad_false_alarm": 0.07937271734601117, "wer": 0.05789473684210526, "insertions": 1, "substitutions": 10, "deletions": 0, "ref_words": 190 }, "reverb_snr10": { "si_sdr_db": -27.123846596241197, "pesq_wb": 1.1811597108840943, "stoi": 0.5884034302319419, "vad_f1": 0.9565243987100389, "vad_false_alarm": 0.07830285083543445, "wer": 0.1368421052631579, "insertions": 0, "substitutions": 24, "deletions": 2, "ref_words": 190 }, "telephony": { "si_sdr_db": 16.26187008926712, "pesq_wb": 3.488973093032837, "stoi": 0.9918412060061893, "vad_f1": 0.9761642766503543, "vad_false_alarm": 0.05308086433332433, "wer": 0.05789473684210526, "insertions": 1, "substitutions": 9, "deletions": 1, "ref_words": 190 }, "telephony_snr10": { "si_sdr_db": 8.954470364257293, "pesq_wb": 1.557040011882782, "stoi": 0.8990446864832322, "vad_f1": 0.9688306828446971, "vad_false_alarm": 0.05104087749954463, "wer": 0.10526315789473684, "insertions": 3, "substitutions": 16, "deletions": 1, "ref_words": 190 }, "telephony_alaw": { "si_sdr_db": 8.728991488815906, "pesq_wb": 1.5973321437835692, "stoi": 0.9067809556983153, "vad_f1": 0.9649250943221878, "vad_false_alarm": 0.04056293999789538, "wer": 0.07368421052631578, "insertions": 2, "substitutions": 12, "deletions": 0, "ref_words": 190 }, "telephony_g722": { "si_sdr_db": 9.79645826361485, "pesq_wb": 1.917847216129303, "stoi": 0.9225672085418355, "vad_f1": 0.9718918630456168, "vad_false_alarm": 0.07682802264269852, "wer": 0.05263157894736842, "insertions": 1, "substitutions": 9, "deletions": 0, "ref_words": 190 }, "telephony_opus12k": { "si_sdr_db": 4.226187239589424, "pesq_wb": 1.7387397050857545, "stoi": 0.8722374918789457, "vad_f1": 0.9700698810886056, "vad_false_alarm": 0.06495112394829519, "wer": 0.06842105263157895, "insertions": 2, "substitutions": 9, "deletions": 2, "ref_words": 190 }, "telephony_loss3": { "si_sdr_db": 7.542881470294612, "pesq_wb": 1.5541536569595338, "stoi": 0.8820020724030957, "vad_f1": 0.9585494127803207, "vad_false_alarm": 0.045096973984416894, "wer": 0.11052631578947368, "insertions": 2, "substitutions": 18, "deletions": 1, "ref_words": 190 }, "telephony_loss10": { "si_sdr_db": 5.077947673477316, "pesq_wb": 1.4434795022010802, "stoi": 0.8618026007914867, "vad_f1": 0.9568651776701316, "vad_false_alarm": 0.052286177577941614, "wer": 0.1, "insertions": 1, "substitutions": 17, "deletions": 1, "ref_words": 190 } }, "nativeSampleRate": 16000, "resampled": false }, { "id": "passthrough", "label": "Passthrough (control)", "license": "\u2014", "shippable": true, "rtf": 5.941857038447201e-05, "blockMsP99": 0.008084231521934283, "realtimeHeadroomP99": 1979.16152655803, "latencyMs": 0.0, "siSdrDb": 10.92073222479158, "siSdrDeltaDb": 0.0, "pesqWb": 1.817549209462272, "stoi": 0.8436932337195054, "estoi": 0.7371305398229724, "wer": 0.1476608187134503, "werSubstitutions": 344, "werInsertions": 101, "werDeletions": 60, "werRefWords": 3420, "vadF1": 0.9503607044937831, "vadFalseAlarm": 0.17156970094457827, "conditions": { "clean": { "si_sdr_db": 150.0, "pesq_wb": 4.643888473510742, "stoi": 0.9999999999999998, "vad_f1": 0.9756248909923848, "vad_false_alarm": 0.07023743685253026, "wer": 0.04736842105263158, "insertions": 2, "substitutions": 6, "deletions": 1, "ref_words": 190 }, "noise_snr-5": { "si_sdr_db": -5.018447138557539, "pesq_wb": 1.104841959476471, "stoi": 0.7198304186475334, "vad_f1": 0.9284106039361377, "vad_false_alarm": 0.06810271419067943, "wer": 0.18421052631578946, "insertions": 4, "substitutions": 29, "deletions": 2, "ref_words": 190 }, "noise_snr0": { "si_sdr_db": -0.008532567221562221, "pesq_wb": 1.2508283972740173, "stoi": 0.8199818537088367, "vad_f1": 0.9650457996816915, "vad_false_alarm": 0.07585414368555106, "wer": 0.11052631578947368, "insertions": 2, "substitutions": 18, "deletions": 1, "ref_words": 190 }, "noise_snr5": { "si_sdr_db": 4.994365358650382, "pesq_wb": 1.661296832561493, "stoi": 0.8595580950060532, "vad_f1": 0.9655544660175688, "vad_false_alarm": 0.06337917413073132, "wer": 0.08947368421052632, "insertions": 2, "substitutions": 15, "deletions": 0, "ref_words": 190 }, "noise_snr10": { "si_sdr_db": 9.996620019510184, "pesq_wb": 1.8911864519119264, "stoi": 0.9162243593607545, "vad_f1": 0.9715691872453027, "vad_false_alarm": 0.0521201836224206, "wer": 0.06315789473684211, "insertions": 2, "substitutions": 10, "deletions": 0, "ref_words": 190 }, "noise_snr20": { "si_sdr_db": 19.998494087473777, "pesq_wb": 2.7327469110488893, "stoi": 0.9804213536382772, "vad_f1": 0.9722639322945248, "vad_false_alarm": 0.0769268153848264, "wer": 0.042105263157894736, "insertions": 2, "substitutions": 6, "deletions": 0, "ref_words": 190 }, "babble_snr5": { "si_sdr_db": 4.993941571017191, "pesq_wb": 1.1675394892692565, "stoi": 0.7761497290169415, "vad_f1": 0.8446580505770814, "vad_false_alarm": 0.9770890756151038, "wer": 0.24736842105263157, "insertions": 10, "substitutions": 27, "deletions": 10, "ref_words": 190 }, "competing_snr0": { "si_sdr_db": 0.03889707927885379, "pesq_wb": 1.1588235735893249, "stoi": 0.7344131574857835, "vad_f1": 0.8978559481992789, "vad_false_alarm": 0.5817642991147487, "wer": 0.8473684210526315, "insertions": 45, "substitutions": 81, "deletions": 35, "ref_words": 190 }, "competing_snr5": { "si_sdr_db": 5.022236435769729, "pesq_wb": 1.2851489901542663, "stoi": 0.8219422931414545, "vad_f1": 0.8993866413420953, "vad_false_alarm": 0.5812582262402548, "wer": 0.2631578947368421, "insertions": 19, "substitutions": 28, "deletions": 3, "ref_words": 190 }, "reverb": { "si_sdr_db": -26.909354792747905, "pesq_wb": 1.3408596515655518, "stoi": 0.6332772949104204, "vad_f1": 0.9623023734900757, "vad_false_alarm": 0.07937271734601117, "wer": 0.05789473684210526, "insertions": 1, "substitutions": 10, "deletions": 0, "ref_words": 190 }, "reverb_snr10": { "si_sdr_db": -27.123846596241197, "pesq_wb": 1.1811597108840943, "stoi": 0.5884034302319419, "vad_f1": 0.9565243987100389, "vad_false_alarm": 0.07830285083543445, "wer": 0.1368421052631579, "insertions": 0, "substitutions": 24, "deletions": 2, "ref_words": 190 }, "telephony": { "si_sdr_db": 16.26187008926712, "pesq_wb": 3.488973093032837, "stoi": 0.9918412060061893, "vad_f1": 0.9761642766503543, "vad_false_alarm": 0.05308086433332433, "wer": 0.05789473684210526, "insertions": 1, "substitutions": 9, "deletions": 1, "ref_words": 190 }, "telephony_snr10": { "si_sdr_db": 8.954470364257293, "pesq_wb": 1.557040011882782, "stoi": 0.8990446864832322, "vad_f1": 0.9688306828446971, "vad_false_alarm": 0.05104087749954463, "wer": 0.10526315789473684, "insertions": 3, "substitutions": 16, "deletions": 1, "ref_words": 190 }, "telephony_alaw": { "si_sdr_db": 8.728991488815906, "pesq_wb": 1.5973321437835692, "stoi": 0.9067809556983153, "vad_f1": 0.9649250943221878, "vad_false_alarm": 0.04056293999789538, "wer": 0.07368421052631578, "insertions": 2, "substitutions": 12, "deletions": 0, "ref_words": 190 }, "telephony_g722": { "si_sdr_db": 9.79645826361485, "pesq_wb": 1.917847216129303, "stoi": 0.9225672085418355, "vad_f1": 0.9718918630456168, "vad_false_alarm": 0.07682802264269852, "wer": 0.05263157894736842, "insertions": 1, "substitutions": 9, "deletions": 0, "ref_words": 190 }, "telephony_opus12k": { "si_sdr_db": 4.226187239589424, "pesq_wb": 1.7387397050857545, "stoi": 0.8722374918789457, "vad_f1": 0.9700698810886056, "vad_false_alarm": 0.06495112394829519, "wer": 0.06842105263157895, "insertions": 2, "substitutions": 9, "deletions": 2, "ref_words": 190 }, "telephony_loss3": { "si_sdr_db": 7.542881470294612, "pesq_wb": 1.5541536569595338, "stoi": 0.8820020724030957, "vad_f1": 0.9585494127803207, "vad_false_alarm": 0.045096973984416894, "wer": 0.11052631578947368, "insertions": 2, "substitutions": 18, "deletions": 1, "ref_words": 190 }, "telephony_loss10": { "si_sdr_db": 5.077947673477316, "pesq_wb": 1.4434795022010802, "stoi": 0.8618026007914867, "vad_f1": 0.9568651776701316, "vad_false_alarm": 0.052286177577941614, "wer": 0.1, "insertions": 1, "substitutions": 17, "deletions": 1, "ref_words": 190 } }, "nativeSampleRate": 16000, "resampled": false }, { "id": "aic-quail", "label": "ai-coustics Quail L", "license": "proprietary", "shippable": false, "rtf": 0.12367526480992147, "blockMsP99": 12.941659858333843, "realtimeHeadroomP99": 0.7726984103635248, "latencyMs": 30.0, "siSdrDb": 8.548828976429423, "siSdrDeltaDb": -2.371903248362157, "pesqWb": 2.0337085286776224, "stoi": 0.8777677532236395, "estoi": 0.7796906463633915, "wer": 0.15555555555555556, "werSubstitutions": 349, "werInsertions": 126, "werDeletions": 57, "werRefWords": 3420, "vadF1": 0.9449486118098753, "vadFalseAlarm": 0.18669395332035799, "conditions": { "clean": { "si_sdr_db": 27.84007563846679, "pesq_wb": 4.549348592758179, "stoi": 0.9979469321108599, "vad_f1": 0.9639893357243585, "vad_false_alarm": 0.098763985449553, "wer": 0.06315789473684211, "insertions": 2, "substitutions": 9, "deletions": 1, "ref_words": 190 }, "noise_snr-5": { "si_sdr_db": 3.006117660273734, "pesq_wb": 1.3319878101348877, "stoi": 0.80270478473238, "vad_f1": 0.9477738242853431, "vad_false_alarm": 0.0845464844086952, "wer": 0.14210526315789473, "insertions": 4, "substitutions": 21, "deletions": 2, "ref_words": 190 }, "noise_snr0": { "si_sdr_db": 7.759026923111756, "pesq_wb": 1.7395087718963622, "stoi": 0.881004514557446, "vad_f1": 0.9604395874001461, "vad_false_alarm": 0.09786018692279255, "wer": 0.06315789473684211, "insertions": 1, "substitutions": 10, "deletions": 1, "ref_words": 190 }, "noise_snr5": { "si_sdr_db": 11.852503242003932, "pesq_wb": 1.9032573103904724, "stoi": 0.8986237918868076, "vad_f1": 0.9607153562156581, "vad_false_alarm": 0.08036451327419945, "wer": 0.11578947368421053, "insertions": 3, "substitutions": 19, "deletions": 0, "ref_words": 190 }, "noise_snr10": { "si_sdr_db": 15.479623247295502, "pesq_wb": 2.323303830623627, "stoi": 0.9409131483264656, "vad_f1": 0.9607995627662849, "vad_false_alarm": 0.07517488945140421, "wer": 0.06315789473684211, "insertions": 2, "substitutions": 10, "deletions": 0, "ref_words": 190 }, "noise_snr20": { "si_sdr_db": 22.243370432972224, "pesq_wb": 3.232144832611084, "stoi": 0.9855479705245193, "vad_f1": 0.9634842371764554, "vad_false_alarm": 0.0989882280312728, "wer": 0.05789473684210526, "insertions": 2, "substitutions": 9, "deletions": 0, "ref_words": 190 }, "babble_snr5": { "si_sdr_db": 5.649188370699831, "pesq_wb": 1.1958783388137817, "stoi": 0.7957007016140093, "vad_f1": 0.847271888651999, "vad_false_alarm": 0.9596433700403029, "wer": 0.2736842105263158, "insertions": 6, "substitutions": 30, "deletions": 16, "ref_words": 190 }, "competing_snr0": { "si_sdr_db": -0.09394895592496648, "pesq_wb": 1.1568449854850769, "stoi": 0.7304902045451132, "vad_f1": 0.8976787826711222, "vad_false_alarm": 0.5766324244238711, "wer": 0.7947368421052632, "insertions": 57, "substitutions": 80, "deletions": 14, "ref_words": 190 }, "competing_snr5": { "si_sdr_db": 4.938847182207267, "pesq_wb": 1.2767481446266173, "stoi": 0.8203741761712635, "vad_f1": 0.8986293123244862, "vad_false_alarm": 0.5756152417780034, "wer": 0.3684210526315789, "insertions": 30, "substitutions": 29, "deletions": 11, "ref_words": 190 }, "reverb": { "si_sdr_db": -9.645195696813321, "pesq_wb": 1.4569910526275636, "stoi": 0.7794941009639246, "vad_f1": 0.9579868395928663, "vad_false_alarm": 0.10152439665663182, "wer": 0.06842105263157895, "insertions": 1, "substitutions": 12, "deletions": 0, "ref_words": 190 }, "reverb_snr10": { "si_sdr_db": -9.870202129246783, "pesq_wb": 1.2710994005203247, "stoi": 0.721157074157111, "vad_f1": 0.9523859457055025, "vad_false_alarm": 0.0979205227484904, "wer": 0.15789473684210525, "insertions": 3, "substitutions": 24, "deletions": 3, "ref_words": 190 }, "telephony": { "si_sdr_db": 15.062950734561266, "pesq_wb": 3.5198292970657348, "stoi": 0.9887910857033656, "vad_f1": 0.9614507398427781, "vad_false_alarm": 0.08403608096882278, "wer": 0.07368421052631578, "insertions": 2, "substitutions": 11, "deletions": 1, "ref_words": 190 }, "telephony_snr10": { "si_sdr_db": 12.150491716836187, "pesq_wb": 1.9379297375679017, "stoi": 0.919801650915758, "vad_f1": 0.9574210139093473, "vad_false_alarm": 0.060592939216135976, "wer": 0.12105263157894737, "insertions": 4, "substitutions": 18, "deletions": 1, "ref_words": 190 }, "telephony_alaw": { "si_sdr_db": 11.437681209532018, "pesq_wb": 1.8422489762306213, "stoi": 0.9225692406813317, "vad_f1": 0.955208853759897, "vad_false_alarm": 0.05951805488700949, "wer": 0.07368421052631578, "insertions": 2, "substitutions": 12, "deletions": 0, "ref_words": 190 }, "telephony_g722": { "si_sdr_db": 15.267170563413654, "pesq_wb": 2.4224007964134215, "stoi": 0.9486216705334882, "vad_f1": 0.9633286137928344, "vad_false_alarm": 0.09278590360631025, "wer": 0.06842105263157895, "insertions": 3, "substitutions": 10, "deletions": 0, "ref_words": 190 }, "telephony_opus12k": { "si_sdr_db": 5.02978607067792, "pesq_wb": 2.131618559360504, "stoi": 0.8980314267560605, "vad_f1": 0.9605505231284232, "vad_false_alarm": 0.09029396961269336, "wer": 0.07368421052631578, "insertions": 2, "substitutions": 11, "deletions": 1, "ref_words": 190 }, "telephony_loss3": { "si_sdr_db": 9.45176908478857, "pesq_wb": 1.738975727558136, "stoi": 0.8958776111455252, "vad_f1": 0.950505258663141, "vad_false_alarm": 0.06229282988880808, "wer": 0.11578947368421053, "insertions": 1, "substitutions": 17, "deletions": 4, "ref_words": 190 }, "telephony_loss10": { "si_sdr_db": 6.319666280874034, "pesq_wb": 1.5766373515129088, "stoi": 0.8721694727000827, "vad_f1": 0.9494553369671153, "vad_false_alarm": 0.0639371384014472, "wer": 0.10526315789473684, "insertions": 1, "substitutions": 17, "deletions": 2, "ref_words": 190 } }, "nativeSampleRate": 16000, "resampled": false }, { "id": "aic-quail-vf", "label": "ai-coustics Quail VF 2.2 L", "license": "proprietary", "shippable": false, "rtf": 0.07551169552295382, "blockMsP99": 6.330404886975879, "realtimeHeadroomP99": 2.3695166846058884, "latencyMs": 30.0, "siSdrDb": 5.020177632795128, "siSdrDeltaDb": -5.900554591996452, "pesqWb": 1.889695609278149, "stoi": 0.8672816722336409, "estoi": 0.7715433769809489, "wer": 0.18099415204678362, "werSubstitutions": 383, "werInsertions": 161, "werDeletions": 75, "werRefWords": 3420, "vadF1": 0.9451311774451506, "vadFalseAlarm": 0.1528947454682677, "conditions": { "clean": { "si_sdr_db": 10.686513902549086, "pesq_wb": 3.7941840171813963, "stoi": 0.9845755693515995, "vad_f1": 0.9642642935054493, "vad_false_alarm": 0.09843234304981663, "wer": 0.05789473684210526, "insertions": 1, "substitutions": 9, "deletions": 1, "ref_words": 190 }, "noise_snr-5": { "si_sdr_db": 4.171784805732267, "pesq_wb": 1.3114071249961854, "stoi": 0.7924695917715265, "vad_f1": 0.9388323659781698, "vad_false_alarm": 0.08067543587989572, "wer": 0.24736842105263157, "insertions": 7, "substitutions": 32, "deletions": 8, "ref_words": 190 }, "noise_snr0": { "si_sdr_db": 6.870490167873546, "pesq_wb": 1.7109012603759766, "stoi": 0.8749545147105188, "vad_f1": 0.9541062681659322, "vad_false_alarm": 0.08695552061901465, "wer": 0.1, "insertions": 1, "substitutions": 15, "deletions": 3, "ref_words": 190 }, "noise_snr5": { "si_sdr_db": 8.55534047579613, "pesq_wb": 1.8149235367774963, "stoi": 0.8899136330852446, "vad_f1": 0.9585920604028434, "vad_false_alarm": 0.08098640311755946, "wer": 0.09473684210526316, "insertions": 1, "substitutions": 15, "deletions": 2, "ref_words": 190 }, "noise_snr10": { "si_sdr_db": 9.053240518586914, "pesq_wb": 2.0865408301353456, "stoi": 0.9251405171546254, "vad_f1": 0.9575286185992142, "vad_false_alarm": 0.08155190805646076, "wer": 0.05789473684210526, "insertions": 1, "substitutions": 10, "deletions": 0, "ref_words": 190 }, "noise_snr20": { "si_sdr_db": 10.285153389650736, "pesq_wb": 2.7600326776504516, "stoi": 0.9704965590416401, "vad_f1": 0.9608867606416368, "vad_false_alarm": 0.09897604266658552, "wer": 0.05789473684210526, "insertions": 1, "substitutions": 9, "deletions": 1, "ref_words": 190 }, "babble_snr5": { "si_sdr_db": 6.414789620780034, "pesq_wb": 1.3471896767616272, "stoi": 0.8381941198963625, "vad_f1": 0.8690845654840278, "vad_false_alarm": 0.7463232636699273, "wer": 0.20526315789473684, "insertions": 7, "substitutions": 24, "deletions": 8, "ref_words": 190 }, "competing_snr0": { "si_sdr_db": -1.7133309543356483, "pesq_wb": 1.1623432993888856, "stoi": 0.6651142087930103, "vad_f1": 0.8955673227016481, "vad_false_alarm": 0.4431741328075541, "wer": 1.168421052631579, "insertions": 109, "substitutions": 90, "deletions": 23, "ref_words": 190 }, "competing_snr5": { "si_sdr_db": 5.080606487156081, "pesq_wb": 1.383172297477722, "stoi": 0.8275947087900635, "vad_f1": 0.9204030792116544, "vad_false_alarm": 0.37171796917161315, "wer": 0.20526315789473684, "insertions": 10, "substitutions": 23, "deletions": 6, "ref_words": 190 }, "reverb": { "si_sdr_db": -13.24554059809119, "pesq_wb": 1.4499569058418273, "stoi": 0.7627670795455301, "vad_f1": 0.9536769965561019, "vad_false_alarm": 0.08770041193161836, "wer": 0.15263157894736842, "insertions": 5, "substitutions": 22, "deletions": 2, "ref_words": 190 }, "reverb_snr10": { "si_sdr_db": -14.59282052885861, "pesq_wb": 1.289806079864502, "stoi": 0.7007605316818591, "vad_f1": 0.9425579524995179, "vad_false_alarm": 0.07576065480987951, "wer": 0.23157894736842105, "insertions": 6, "substitutions": 33, "deletions": 5, "ref_words": 190 }, "telephony": { "si_sdr_db": 11.232591614391113, "pesq_wb": 3.024996542930603, "stoi": 0.9750662538542361, "vad_f1": 0.9626392421208759, "vad_false_alarm": 0.08560959061088316, "wer": 0.08421052631578947, "insertions": 3, "substitutions": 11, "deletions": 2, "ref_words": 190 }, "telephony_snr10": { "si_sdr_db": 10.092926850677953, "pesq_wb": 1.861033058166504, "stoi": 0.9115778288842649, "vad_f1": 0.9571519498721031, "vad_false_alarm": 0.057651468730805844, "wer": 0.09473684210526316, "insertions": 3, "substitutions": 14, "deletions": 1, "ref_words": 190 }, "telephony_alaw": { "si_sdr_db": 9.769707357687391, "pesq_wb": 1.736561918258667, "stoi": 0.9197724409707865, "vad_f1": 0.9584714101325235, "vad_false_alarm": 0.0535979419445999, "wer": 0.10526315789473684, "insertions": 3, "substitutions": 16, "deletions": 1, "ref_words": 190 }, "telephony_g722": { "si_sdr_db": 9.535822728618594, "pesq_wb": 2.1940940618515015, "stoi": 0.9344314872081163, "vad_f1": 0.961230532167449, "vad_false_alarm": 0.08843995543495607, "wer": 0.06842105263157895, "insertions": 1, "substitutions": 10, "deletions": 2, "ref_words": 190 }, "telephony_opus12k": { "si_sdr_db": 4.080929458425428, "pesq_wb": 1.8920276045799256, "stoi": 0.8922807044737331, "vad_f1": 0.9605288807028897, "vad_false_alarm": 0.08400093121227807, "wer": 0.09473684210526316, "insertions": 1, "substitutions": 14, "deletions": 3, "ref_words": 190 }, "telephony_loss3": { "si_sdr_db": 8.147163883045357, "pesq_wb": 1.6926931738853455, "stoi": 0.8885439603958997, "vad_f1": 0.9489408846529939, "vad_false_alarm": 0.06050294787293735, "wer": 0.11052631578947368, "insertions": 0, "substitutions": 17, "deletions": 4, "ref_words": 190 }, "telephony_loss10": { "si_sdr_db": 5.937828210627113, "pesq_wb": 1.5026569008827209, "stoi": 0.8574163905965213, "vad_f1": 0.9478980106176815, "vad_false_alarm": 0.07004849684243278, "wer": 0.12105263157894737, "insertions": 1, "substitutions": 19, "deletions": 3, "ref_words": 190 } }, "nativeSampleRate": 16000, "resampled": false }, { "id": "gtcrn", "label": "GTCRN", "license": "MIT", "shippable": true, "rtf": 0.04976707822388168, "blockMsP99": 4.660299685783687, "realtimeHeadroomP99": 3.4332556013099835, "latencyMs": 16.0, "siSdrDb": 8.785021147165201, "siSdrDeltaDb": -2.135711077626379, "pesqWb": 2.090960569514168, "stoi": 0.8570012199935738, "estoi": 0.7638815064264679, "wer": 0.2026315789473684, "werSubstitutions": 443, "werInsertions": 210, "werDeletions": 40, "werRefWords": 3420, "vadF1": 0.9472170326540659, "vadFalseAlarm": 0.16376826535043865, "conditions": { "clean": { "si_sdr_db": 27.926269166103133, "pesq_wb": 3.984705018997192, "stoi": 0.9911377012132366, "vad_f1": 0.9660543915503998, "vad_false_alarm": 0.08697693497142549, "wer": 0.06315789473684211, "insertions": 2, "substitutions": 9, "deletions": 1, "ref_words": 190 }, "noise_snr-5": { "si_sdr_db": 8.553455391030614, "pesq_wb": 1.6495023846626282, "stoi": 0.8216203160412027, "vad_f1": 0.9552931392660063, "vad_false_alarm": 0.06843675237866058, "wer": 0.19473684210526315, "insertions": 4, "substitutions": 33, "deletions": 0, "ref_words": 190 }, "noise_snr0": { "si_sdr_db": 12.902761713389978, "pesq_wb": 2.186266005039215, "stoi": 0.8950200689378802, "vad_f1": 0.9624445080427307, "vad_false_alarm": 0.05563103296541287, "wer": 0.08421052631578947, "insertions": 3, "substitutions": 13, "deletions": 0, "ref_words": 190 }, "noise_snr5": { "si_sdr_db": 15.119459828576254, "pesq_wb": 2.373152422904968, "stoi": 0.9131658011342271, "vad_f1": 0.9654063506734911, "vad_false_alarm": 0.047388994708232095, "wer": 0.1, "insertions": 3, "substitutions": 16, "deletions": 0, "ref_words": 190 }, "noise_snr10": { "si_sdr_db": 17.439785157915942, "pesq_wb": 2.711093711853027, "stoi": 0.9451584335656407, "vad_f1": 0.964118704379777, "vad_false_alarm": 0.0550873866267928, "wer": 0.08421052631578947, "insertions": 1, "substitutions": 14, "deletions": 1, "ref_words": 190 }, "noise_snr20": { "si_sdr_db": 23.448350132653935, "pesq_wb": 3.4268387079238893, "stoi": 0.9774955198664728, "vad_f1": 0.9661604272190638, "vad_false_alarm": 0.07713223243072398, "wer": 0.06315789473684211, "insertions": 2, "substitutions": 9, "deletions": 1, "ref_words": 190 }, "babble_snr5": { "si_sdr_db": 3.8176879504595176, "pesq_wb": 1.2009369492530824, "stoi": 0.7462700905565257, "vad_f1": 0.8492274596155684, "vad_false_alarm": 0.9190328053687009, "wer": 0.6105263157894737, "insertions": 63, "substitutions": 53, "deletions": 0, "ref_words": 190 }, "competing_snr0": { "si_sdr_db": -0.5258054924236761, "pesq_wb": 1.1608852982521056, "stoi": 0.7030534609256999, "vad_f1": 0.8983963703575533, "vad_false_alarm": 0.5711054910377819, "wer": 0.8789473684210526, "insertions": 64, "substitutions": 93, "deletions": 10, "ref_words": 190 }, "competing_snr5": { "si_sdr_db": 4.664923262529486, "pesq_wb": 1.2809460639953614, "stoi": 0.801085091114541, "vad_f1": 0.8999027489965339, "vad_false_alarm": 0.5657796308855336, "wer": 0.4473684210526316, "insertions": 46, "substitutions": 30, "deletions": 9, "ref_words": 190 }, "reverb": { "si_sdr_db": -16.52225361688118, "pesq_wb": 1.3613831639289855, "stoi": 0.6885026433680478, "vad_f1": 0.9628998075243203, "vad_false_alarm": 0.09150975804156156, "wer": 0.06842105263157895, "insertions": 1, "substitutions": 12, "deletions": 0, "ref_words": 190 }, "reverb_snr10": { "si_sdr_db": -15.79689192086567, "pesq_wb": 1.3095790505409242, "stoi": 0.658753746269473, "vad_f1": 0.961264056628323, "vad_false_alarm": 0.07846796133865161, "wer": 0.14736842105263157, "insertions": 3, "substitutions": 24, "deletions": 1, "ref_words": 190 }, "telephony": { "si_sdr_db": 14.742226594898815, "pesq_wb": 2.7711402654647825, "stoi": 0.9604645235470931, "vad_f1": 0.9602097278763091, "vad_false_alarm": 0.056051811379656355, "wer": 0.08947368421052632, "insertions": 2, "substitutions": 14, "deletions": 1, "ref_words": 190 }, "telephony_snr10": { "si_sdr_db": 12.393333772731399, "pesq_wb": 1.9177796244621277, "stoi": 0.8890268887613324, "vad_f1": 0.9566377472547558, "vad_false_alarm": 0.037106203972974955, "wer": 0.11052631578947368, "insertions": 3, "substitutions": 17, "deletions": 1, "ref_words": 190 }, "telephony_alaw": { "si_sdr_db": 12.152632002779788, "pesq_wb": 1.8421189785003662, "stoi": 0.8913310431177095, "vad_f1": 0.9555179803091851, "vad_false_alarm": 0.04011145552226702, "wer": 0.1368421052631579, "insertions": 2, "substitutions": 23, "deletions": 1, "ref_words": 190 }, "telephony_g722": { "si_sdr_db": 16.617958649287523, "pesq_wb": 2.8151506662368773, "stoi": 0.9515643739025968, "vad_f1": 0.9656159755941424, "vad_false_alarm": 0.06969920953477984, "wer": 0.05789473684210526, "insertions": 2, "substitutions": 8, "deletions": 1, "ref_words": 190 }, "telephony_opus12k": { "si_sdr_db": 5.110675882475176, "pesq_wb": 2.305869090557098, "stoi": 0.8946813200508016, "vad_f1": 0.9666243797228073, "vad_false_alarm": 0.05336690244052579, "wer": 0.12105263157894737, "insertions": 4, "substitutions": 17, "deletions": 2, "ref_words": 190 }, "telephony_loss3": { "si_sdr_db": 9.428832518833037, "pesq_wb": 1.776776623725891, "stoi": 0.8618725670912977, "vad_f1": 0.9520409864804803, "vad_false_alarm": 0.04274530171733586, "wer": 0.14210526315789473, "insertions": 1, "substitutions": 22, "deletions": 4, "ref_words": 190 }, "telephony_loss10": { "si_sdr_db": 6.656979655479537, "pesq_wb": 1.5631662249565124, "stoi": 0.8358183704205503, "vad_f1": 0.9420918262817392, "vad_false_alarm": 0.03219891098687841, "wer": 0.24736842105263157, "insertions": 4, "substitutions": 36, "deletions": 7, "ref_words": 190 } }, "nativeSampleRate": 16000, "resampled": false }, { "id": "fastenhancer-l", "label": "FastEnhancer-L", "license": "MIT", "shippable": true, "rtf": 0.3761973454467899, "blockMsP99": 4.933790845097972, "realtimeHeadroomP99": 1.2667744126627831, "latencyMs": 25.75, "siSdrDb": 10.598903888438464, "siSdrDeltaDb": -0.32182833635311603, "pesqWb": 2.198379112614526, "stoi": 0.8504416784853068, "estoi": 0.7776768038832961, "wer": 0.2043859649122807, "werSubstitutions": 433, "werInsertions": 159, "werDeletions": 107, "werRefWords": 3420, "vadF1": 0.9417271235541121, "vadFalseAlarm": 0.1474487272599934, "conditions": { "clean": { "si_sdr_db": 39.18288707568924, "pesq_wb": 4.469868230819702, "stoi": 0.9985691734150979, "vad_f1": 0.9629243179178815, "vad_false_alarm": 0.09624337669876035, "wer": 0.05263157894736842, "insertions": 1, "substitutions": 9, "deletions": 0, "ref_words": 190 }, "noise_snr-5": { "si_sdr_db": 12.616845174315284, "pesq_wb": 1.765555477142334, "stoi": 0.8811480252234196, "vad_f1": 0.9579620400813467, "vad_false_alarm": 0.04727318113617144, "wer": 0.1368421052631579, "insertions": 2, "substitutions": 22, "deletions": 2, "ref_words": 190 }, "noise_snr0": { "si_sdr_db": 17.128747602967206, "pesq_wb": 2.383464682102203, "stoi": 0.9393521210962877, "vad_f1": 0.9627543916240529, "vad_false_alarm": 0.07848709604461479, "wer": 0.07894736842105263, "insertions": 1, "substitutions": 13, "deletions": 1, "ref_words": 190 }, "noise_snr5": { "si_sdr_db": 19.112763043833553, "pesq_wb": 2.774562680721283, "stoi": 0.9426231419351577, "vad_f1": 0.9649830801016552, "vad_false_alarm": 0.06291504690619273, "wer": 0.06842105263157895, "insertions": 1, "substitutions": 11, "deletions": 1, "ref_words": 190 }, "noise_snr10": { "si_sdr_db": 20.817571382688477, "pesq_wb": 2.9158797740936278, "stoi": 0.966353839520892, "vad_f1": 0.9630539185795298, "vad_false_alarm": 0.07026693016822186, "wer": 0.05263157894736842, "insertions": 3, "substitutions": 7, "deletions": 0, "ref_words": 190 }, "noise_snr20": { "si_sdr_db": 26.73180128035229, "pesq_wb": 3.722787380218506, "stoi": 0.9892948566868757, "vad_f1": 0.9658739810576314, "vad_false_alarm": 0.08561768613224802, "wer": 0.03684210526315789, "insertions": 1, "substitutions": 6, "deletions": 0, "ref_words": 190 }, "babble_snr5": { "si_sdr_db": 5.395945380158565, "pesq_wb": 1.2278975129127503, "stoi": 0.7354607769389095, "vad_f1": 0.8511730293294211, "vad_false_alarm": 0.6824619387440339, "wer": 0.3684210526315789, "insertions": 12, "substitutions": 40, "deletions": 18, "ref_words": 190 }, "competing_snr0": { "si_sdr_db": -3.012576792253197, "pesq_wb": 1.122466790676117, "stoi": 0.5320367998448507, "vad_f1": 0.8826170709857722, "vad_false_alarm": 0.5451808126594629, "wer": 0.8421052631578947, "insertions": 31, "substitutions": 91, "deletions": 38, "ref_words": 190 }, "competing_snr5": { "si_sdr_db": 4.913332986908968, "pesq_wb": 1.317231547832489, "stoi": 0.736221421645394, "vad_f1": 0.9059799600495037, "vad_false_alarm": 0.3904156960002697, "wer": 0.7631578947368421, "insertions": 85, "substitutions": 54, "deletions": 6, "ref_words": 190 }, "reverb": { "si_sdr_db": -17.63390407751499, "pesq_wb": 1.278448188304901, "stoi": 0.6013940434420364, "vad_f1": 0.9127343755383647, "vad_false_alarm": 0.09703355738966754, "wer": 0.1736842105263158, "insertions": 3, "substitutions": 22, "deletions": 8, "ref_words": 190 }, "reverb_snr10": { "si_sdr_db": -15.913789017980614, "pesq_wb": 1.2272470951080323, "stoi": 0.6030266831454774, "vad_f1": 0.9474239824936153, "vad_false_alarm": 0.08594514993159044, "wer": 0.22631578947368422, "insertions": 1, "substitutions": 31, "deletions": 11, "ref_words": 190 }, "telephony": { "si_sdr_db": 14.99243760140448, "pesq_wb": 2.850234866142273, "stoi": 0.9673641558179616, "vad_f1": 0.959613620904966, "vad_false_alarm": 0.07541109124065136, "wer": 0.07368421052631578, "insertions": 1, "substitutions": 10, "deletions": 3, "ref_words": 190 }, "telephony_snr10": { "si_sdr_db": 13.22983460020851, "pesq_wb": 2.0140509247779845, "stoi": 0.917541804187856, "vad_f1": 0.9532800954240364, "vad_false_alarm": 0.05091177459836592, "wer": 0.1368421052631579, "insertions": 3, "substitutions": 22, "deletions": 1, "ref_words": 190 }, "telephony_alaw": { "si_sdr_db": 12.681412001544459, "pesq_wb": 1.934655487537384, "stoi": 0.8985760074692383, "vad_f1": 0.9500018687942842, "vad_false_alarm": 0.05337487974278775, "wer": 0.12105263157894737, "insertions": 1, "substitutions": 18, "deletions": 4, "ref_words": 190 }, "telephony_g722": { "si_sdr_db": 19.11680772609362, "pesq_wb": 2.9838998079299928, "stoi": 0.9690114339841692, "vad_f1": 0.9651000357824729, "vad_false_alarm": 0.07506196752001078, "wer": 0.05263157894736842, "insertions": 1, "substitutions": 8, "deletions": 1, "ref_words": 190 }, "telephony_opus12k": { "si_sdr_db": 5.276003409717047, "pesq_wb": 2.337374782562256, "stoi": 0.9163541509111368, "vad_f1": 0.9659973396042536, "vad_false_alarm": 0.05232077297435804, "wer": 0.09473684210526316, "insertions": 4, "substitutions": 13, "deletions": 1, "ref_words": 190 }, "telephony_loss3": { "si_sdr_db": 9.564346199356404, "pesq_wb": 1.7554494857788085, "stoi": 0.8771451079136099, "vad_f1": 0.9377517967544996, "vad_false_alarm": 0.05672220228888042, "wer": 0.2, "insertions": 5, "substitutions": 28, "deletions": 5, "ref_words": 190 }, "telephony_loss10": { "si_sdr_db": 6.579804414403062, "pesq_wb": 1.4897493124008179, "stoi": 0.8364766695571533, "vad_f1": 0.9418633189507318, "vad_false_alarm": 0.0484339305035929, "wer": 0.2, "insertions": 3, "substitutions": 28, "deletions": 7, "ref_words": 190 } }, "nativeSampleRate": 16000, "resampled": false }, { "id": "fastenhancer-s", "label": "FastEnhancer-S", "license": "MIT", "shippable": true, "rtf": 0.03364957248664088, "blockMsP99": 1.043815444572829, "realtimeHeadroomP99": 15.32838020666368, "latencyMs": 16.0, "siSdrDb": 10.233805867854713, "siSdrDeltaDb": -0.6869263569368673, "pesqWb": 2.2054898540178933, "stoi": 0.853893223683037, "estoi": 0.7725298573839887, "wer": 0.2064327485380117, "werSubstitutions": 475, "werInsertions": 159, "werDeletions": 72, "werRefWords": 3420, "vadF1": 0.9452405508174968, "vadFalseAlarm": 0.16179167096077154, "conditions": { "clean": { "si_sdr_db": 37.56554413430047, "pesq_wb": 4.511321878433227, "stoi": 0.9982137897260062, "vad_f1": 0.9666243625184097, "vad_false_alarm": 0.0872933030935098, "wer": 0.05789473684210526, "insertions": 1, "substitutions": 9, "deletions": 1, "ref_words": 190 }, "noise_snr-5": { "si_sdr_db": 11.01258338251217, "pesq_wb": 1.7659117102622985, "stoi": 0.8503449900669813, "vad_f1": 0.9543291432211548, "vad_false_alarm": 0.050790979530382696, "wer": 0.22631578947368422, "insertions": 9, "substitutions": 33, "deletions": 1, "ref_words": 190 }, "noise_snr0": { "si_sdr_db": 15.615540372427613, "pesq_wb": 2.3970736265182495, "stoi": 0.9252274660568001, "vad_f1": 0.9627685527161255, "vad_false_alarm": 0.06696549623758156, "wer": 0.08947368421052632, "insertions": 3, "substitutions": 13, "deletions": 1, "ref_words": 190 }, "noise_snr5": { "si_sdr_db": 18.17582506101028, "pesq_wb": 2.6875906467437742, "stoi": 0.9325193314053587, "vad_f1": 0.9693534128979128, "vad_false_alarm": 0.04713787126123235, "wer": 0.08421052631578947, "insertions": 1, "substitutions": 15, "deletions": 0, "ref_words": 190 }, "noise_snr10": { "si_sdr_db": 19.268514833674235, "pesq_wb": 2.925475537776947, "stoi": 0.9591335800473271, "vad_f1": 0.9657387351921551, "vad_false_alarm": 0.05911004697911254, "wer": 0.07368421052631578, "insertions": 2, "substitutions": 12, "deletions": 0, "ref_words": 190 }, "noise_snr20": { "si_sdr_db": 25.13889618109814, "pesq_wb": 3.6609259843826294, "stoi": 0.9877846453152171, "vad_f1": 0.9664454793934116, "vad_false_alarm": 0.07605903222268466, "wer": 0.05263157894736842, "insertions": 1, "substitutions": 8, "deletions": 1, "ref_words": 190 }, "babble_snr5": { "si_sdr_db": 5.87496200341801, "pesq_wb": 1.2744722366333008, "stoi": 0.7580606167825745, "vad_f1": 0.8577665455520223, "vad_false_alarm": 0.8505172318248697, "wer": 0.3736842105263158, "insertions": 13, "substitutions": 35, "deletions": 23, "ref_words": 190 }, "competing_snr0": { "si_sdr_db": -1.4980791722368703, "pesq_wb": 1.1203026294708252, "stoi": 0.6130913149667717, "vad_f1": 0.8979486609373313, "vad_false_alarm": 0.5681011160616604, "wer": 0.8631578947368421, "insertions": 63, "substitutions": 100, "deletions": 1, "ref_words": 190 }, "competing_snr5": { "si_sdr_db": 4.61460515844626, "pesq_wb": 1.2845615148544312, "stoi": 0.7577854437605377, "vad_f1": 0.9010474911433507, "vad_false_alarm": 0.5559165919852933, "wer": 0.5105263157894737, "insertions": 35, "substitutions": 53, "deletions": 9, "ref_words": 190 }, "reverb": { "si_sdr_db": -14.828905543128116, "pesq_wb": 1.2738083243370055, "stoi": 0.6380293231196266, "vad_f1": 0.9411470356453613, "vad_false_alarm": 0.08884428522578725, "wer": 0.22631578947368422, "insertions": 9, "substitutions": 29, "deletions": 5, "ref_words": 190 }, "reverb_snr10": { "si_sdr_db": -14.665252147462477, "pesq_wb": 1.256077241897583, "stoi": 0.6256058396499627, "vad_f1": 0.9488516914877356, "vad_false_alarm": 0.08334507746767365, "wer": 0.2578947368421053, "insertions": 5, "substitutions": 35, "deletions": 9, "ref_words": 190 }, "telephony": { "si_sdr_db": 15.560167266289785, "pesq_wb": 3.051128697395325, "stoi": 0.9744370283187063, "vad_f1": 0.9629450687541201, "vad_false_alarm": 0.06770453006248747, "wer": 0.05263157894736842, "insertions": 1, "substitutions": 8, "deletions": 1, "ref_words": 190 }, "telephony_snr10": { "si_sdr_db": 12.066630295771436, "pesq_wb": 1.9201302170753478, "stoi": 0.891965400987323, "vad_f1": 0.952131310382198, "vad_false_alarm": 0.04433800863550697, "wer": 0.14210526315789473, "insertions": 3, "substitutions": 21, "deletions": 3, "ref_words": 190 }, "telephony_alaw": { "si_sdr_db": 11.48439321478298, "pesq_wb": 1.8805826783180237, "stoi": 0.8878264649385518, "vad_f1": 0.9457094682914681, "vad_false_alarm": 0.049777509449397364, "wer": 0.1631578947368421, "insertions": 2, "substitutions": 23, "deletions": 6, "ref_words": 190 }, "telephony_g722": { "si_sdr_db": 18.26168213696321, "pesq_wb": 3.0631088256835937, "stoi": 0.9642998007788165, "vad_f1": 0.966584863318619, "vad_false_alarm": 0.07172363649743699, "wer": 0.05789473684210526, "insertions": 1, "substitutions": 10, "deletions": 0, "ref_words": 190 }, "telephony_opus12k": { "si_sdr_db": 5.237359882435156, "pesq_wb": 2.4052021622657778, "stoi": 0.9106634409698554, "vad_f1": 0.9678500739831397, "vad_false_alarm": 0.050673430949711475, "wer": 0.06842105263157895, "insertions": 3, "substitutions": 10, "deletions": 0, "ref_words": 190 }, "telephony_loss3": { "si_sdr_db": 8.952451549334475, "pesq_wb": 1.7135145306587218, "stoi": 0.8663022517588962, "vad_f1": 0.9464638375279874, "vad_false_alarm": 0.04881442287730438, "wer": 0.16842105263157894, "insertions": 3, "substitutions": 24, "deletions": 5, "ref_words": 190 }, "telephony_loss10": { "si_sdr_db": 6.371587011748082, "pesq_wb": 1.5076289296150207, "stoi": 0.8287872976453526, "vad_f1": 0.9406241817524394, "vad_false_alarm": 0.045137506932255246, "wer": 0.24736842105263157, "insertions": 4, "substitutions": 37, "deletions": 6, "ref_words": 190 } }, "nativeSampleRate": 16000, "resampled": false }, { "id": "fastenhancer-b", "label": "FastEnhancer-B", "license": "MIT", "shippable": true, "rtf": 0.021207815833411067, "blockMsP99": 5.443321244092665, "realtimeHeadroomP99": 2.9393819108809556, "latencyMs": 16.0, "siSdrDb": 9.840327678488405, "siSdrDeltaDb": -1.0804045463031748, "pesqWb": 2.163641287883123, "stoi": 0.8478673437873312, "estoi": 0.7615835740414477, "wer": 0.21140350877192982, "werSubstitutions": 485, "werInsertions": 159, "werDeletions": 79, "werRefWords": 3420, "vadF1": 0.9442287866040439, "vadFalseAlarm": 0.16169900929669306, "conditions": { "clean": { "si_sdr_db": 35.453957813284525, "pesq_wb": 4.479747796058655, "stoi": 0.9962926255606804, "vad_f1": 0.9665544167255531, "vad_false_alarm": 0.0848434576737768, "wer": 0.05789473684210526, "insertions": 1, "substitutions": 9, "deletions": 1, "ref_words": 190 }, "noise_snr-5": { "si_sdr_db": 10.512947261058834, "pesq_wb": 1.6973995685577392, "stoi": 0.8384307353820336, "vad_f1": 0.9486474458583054, "vad_false_alarm": 0.04696256044091943, "wer": 0.18421052631578946, "insertions": 3, "substitutions": 31, "deletions": 1, "ref_words": 190 }, "noise_snr0": { "si_sdr_db": 14.36052669903989, "pesq_wb": 2.224401664733887, "stoi": 0.9097550295163888, "vad_f1": 0.9631194721873815, "vad_false_alarm": 0.06304611532347819, "wer": 0.07368421052631578, "insertions": 2, "substitutions": 12, "deletions": 0, "ref_words": 190 }, "noise_snr5": { "si_sdr_db": 17.437656969105156, "pesq_wb": 2.6657270193099976, "stoi": 0.9286764998130842, "vad_f1": 0.9677787893171903, "vad_false_alarm": 0.04610694342618081, "wer": 0.06842105263157895, "insertions": 1, "substitutions": 12, "deletions": 0, "ref_words": 190 }, "noise_snr10": { "si_sdr_db": 18.830058937312238, "pesq_wb": 2.8779727697372435, "stoi": 0.9532192973289616, "vad_f1": 0.9664037535861152, "vad_false_alarm": 0.060846861774346575, "wer": 0.05789473684210526, "insertions": 1, "substitutions": 10, "deletions": 0, "ref_words": 190 }, "noise_snr20": { "si_sdr_db": 24.282107363030786, "pesq_wb": 3.6355909585952757, "stoi": 0.9841035510826623, "vad_f1": 0.968372309057336, "vad_false_alarm": 0.06914094507860882, "wer": 0.05263157894736842, "insertions": 1, "substitutions": 9, "deletions": 0, "ref_words": 190 }, "babble_snr5": { "si_sdr_db": 6.7661774319426415, "pesq_wb": 1.2877558469772339, "stoi": 0.7757449313372502, "vad_f1": 0.8539578737993498, "vad_false_alarm": 0.8737395622775651, "wer": 0.3684210526315789, "insertions": 25, "substitutions": 33, "deletions": 12, "ref_words": 190 }, "competing_snr0": { "si_sdr_db": -1.1385512194560388, "pesq_wb": 1.1292998313903808, "stoi": 0.6362220976253457, "vad_f1": 0.8984464534886838, "vad_false_alarm": 0.5720027589612217, "wer": 0.9263157894736842, "insertions": 57, "substitutions": 98, "deletions": 21, "ref_words": 190 }, "competing_snr5": { "si_sdr_db": 4.9656645648945155, "pesq_wb": 1.283970308303833, "stoi": 0.7676229029923501, "vad_f1": 0.9000794486343897, "vad_false_alarm": 0.5631822984264705, "wer": 0.5052631578947369, "insertions": 41, "substitutions": 46, "deletions": 9, "ref_words": 190 }, "reverb": { "si_sdr_db": -15.389190193330517, "pesq_wb": 1.2573384642601013, "stoi": 0.6132840226544805, "vad_f1": 0.9386444498238232, "vad_false_alarm": 0.0975312019452212, "wer": 0.2578947368421053, "insertions": 3, "substitutions": 42, "deletions": 4, "ref_words": 190 }, "reverb_snr10": { "si_sdr_db": -15.528447882609942, "pesq_wb": 1.2271236062049866, "stoi": 0.6030811009957959, "vad_f1": 0.9434659771081435, "vad_false_alarm": 0.08586169665095951, "wer": 0.34210526315789475, "insertions": 9, "substitutions": 44, "deletions": 12, "ref_words": 190 }, "telephony": { "si_sdr_db": 14.963951818398005, "pesq_wb": 2.95293071269989, "stoi": 0.9664349840436517, "vad_f1": 0.9617762772896791, "vad_false_alarm": 0.06606518580019238, "wer": 0.06842105263157895, "insertions": 2, "substitutions": 10, "deletions": 1, "ref_words": 190 }, "telephony_snr10": { "si_sdr_db": 12.757138149149148, "pesq_wb": 1.9234709024429322, "stoi": 0.8973493049515977, "vad_f1": 0.9511253080366208, "vad_false_alarm": 0.03909554636293228, "wer": 0.1368421052631579, "insertions": 2, "substitutions": 22, "deletions": 2, "ref_words": 190 }, "telephony_alaw": { "si_sdr_db": 10.826631851655765, "pesq_wb": 1.8088917970657348, "stoi": 0.8700585257583748, "vad_f1": 0.9506472775000845, "vad_false_alarm": 0.04273317783109369, "wer": 0.16842105263157894, "insertions": 3, "substitutions": 27, "deletions": 2, "ref_words": 190 }, "telephony_g722": { "si_sdr_db": 17.89000333779208, "pesq_wb": 3.003855514526367, "stoi": 0.9586726515636732, "vad_f1": 0.9668796232189438, "vad_false_alarm": 0.0695962794586116, "wer": 0.05263157894736842, "insertions": 1, "substitutions": 8, "deletions": 1, "ref_words": 190 }, "telephony_opus12k": { "si_sdr_db": 5.229467961736965, "pesq_wb": 2.384562051296234, "stoi": 0.903322204051417, "vad_f1": 0.9682160061569374, "vad_false_alarm": 0.049905613891517944, "wer": 0.08421052631578947, "insertions": 2, "substitutions": 12, "deletions": 2, "ref_words": 190 }, "telephony_loss3": { "si_sdr_db": 8.918200055334228, "pesq_wb": 1.6709478616714477, "stoi": 0.8527160477021735, "vad_f1": 0.9457321066331812, "vad_false_alarm": 0.03829010256246379, "wer": 0.16842105263157894, "insertions": 2, "substitutions": 26, "deletions": 4, "ref_words": 190 }, "telephony_loss10": { "si_sdr_db": 5.987597294453027, "pesq_wb": 1.43455650806427, "stoi": 0.8066256758120405, "vad_f1": 0.9362711704510704, "vad_false_alarm": 0.04163185945491459, "wer": 0.23157894736842105, "insertions": 3, "substitutions": 34, "deletions": 7, "ref_words": 190 } }, "nativeSampleRate": 16000, "resampled": false }, { "id": "fastenhancer-m", "label": "FastEnhancer-M", "license": "MIT", "shippable": true, "rtf": 0.11015943624908402, "blockMsP99": 3.208531528362057, "realtimeHeadroomP99": 3.1166905830920606, "latencyMs": 22.0, "siSdrDb": 10.63291946780459, "siSdrDeltaDb": -0.28781275698698927, "pesqWb": 2.120756770504846, "stoi": 0.8487749873037099, "estoi": 0.7708327837777946, "wer": 0.2239766081871345, "werSubstitutions": 466, "werInsertions": 217, "werDeletions": 83, "werRefWords": 3420, "vadF1": 0.9405634506008924, "vadFalseAlarm": 0.14912635065900548, "conditions": { "clean": { "si_sdr_db": 39.65844603864674, "pesq_wb": 4.376759338378906, "stoi": 0.9972736316886717, "vad_f1": 0.9646538637898979, "vad_false_alarm": 0.0910952605066407, "wer": 0.042105263157894736, "insertions": 2, "substitutions": 6, "deletions": 0, "ref_words": 190 }, "noise_snr-5": { "si_sdr_db": 11.954445764825433, "pesq_wb": 1.7187994360923766, "stoi": 0.8609145465163304, "vad_f1": 0.9472517306837446, "vad_false_alarm": 0.04154623326386513, "wer": 0.15263157894736842, "insertions": 6, "substitutions": 19, "deletions": 4, "ref_words": 190 }, "noise_snr0": { "si_sdr_db": 16.079098031035933, "pesq_wb": 2.2763553261756897, "stoi": 0.9289298874128361, "vad_f1": 0.9628264939790558, "vad_false_alarm": 0.07001115798376083, "wer": 0.07894736842105263, "insertions": 2, "substitutions": 12, "deletions": 1, "ref_words": 190 }, "noise_snr5": { "si_sdr_db": 19.048462566174493, "pesq_wb": 2.6307727336883544, "stoi": 0.9397480304506358, "vad_f1": 0.9660458156554125, "vad_false_alarm": 0.04938579582784024, "wer": 0.09473684210526316, "insertions": 1, "substitutions": 14, "deletions": 3, "ref_words": 190 }, "noise_snr10": { "si_sdr_db": 20.208909829852114, "pesq_wb": 2.8404269456863402, "stoi": 0.9609166514102506, "vad_f1": 0.9665975500045096, "vad_false_alarm": 0.058201537637794364, "wer": 0.05789473684210526, "insertions": 2, "substitutions": 9, "deletions": 0, "ref_words": 190 }, "noise_snr20": { "si_sdr_db": 26.41206541114976, "pesq_wb": 3.6216189861297607, "stoi": 0.986186727974595, "vad_f1": 0.9664991502166893, "vad_false_alarm": 0.07626595923666148, "wer": 0.031578947368421054, "insertions": 1, "substitutions": 5, "deletions": 0, "ref_words": 190 }, "babble_snr5": { "si_sdr_db": 6.133714674702732, "pesq_wb": 1.2688073873519898, "stoi": 0.770617786128212, "vad_f1": 0.8587428616865319, "vad_false_alarm": 0.7461100741696446, "wer": 0.3, "insertions": 9, "substitutions": 32, "deletions": 16, "ref_words": 190 }, "competing_snr0": { "si_sdr_db": -2.4268588320415576, "pesq_wb": 1.1271017789840698, "stoi": 0.5694455410506627, "vad_f1": 0.8949618104404795, "vad_false_alarm": 0.5473616995996899, "wer": 0.9631578947368421, "insertions": 71, "substitutions": 98, "deletions": 14, "ref_words": 190 }, "competing_snr5": { "si_sdr_db": 4.855496647655906, "pesq_wb": 1.2915028572082519, "stoi": 0.7619792298857027, "vad_f1": 0.9045856994266565, "vad_false_alarm": 0.4828302424810251, "wer": 0.46842105263157896, "insertions": 30, "substitutions": 50, "deletions": 9, "ref_words": 190 }, "reverb": { "si_sdr_db": -13.028290550752136, "pesq_wb": 1.2497697353363038, "stoi": 0.6207072627106849, "vad_f1": 0.9155334128372689, "vad_false_alarm": 0.09937079113985055, "wer": 0.2736842105263158, "insertions": 13, "substitutions": 37, "deletions": 2, "ref_words": 190 }, "reverb_snr10": { "si_sdr_db": -14.687220702888306, "pesq_wb": 1.239696443080902, "stoi": 0.6188126139981137, "vad_f1": 0.9511791968018667, "vad_false_alarm": 0.08211856272974793, "wer": 0.22105263157894736, "insertions": 4, "substitutions": 35, "deletions": 3, "ref_words": 190 }, "telephony": { "si_sdr_db": 14.963327088931655, "pesq_wb": 2.8068102836608886, "stoi": 0.9567813753750244, "vad_f1": 0.956563721091612, "vad_false_alarm": 0.05917274923662395, "wer": 0.05789473684210526, "insertions": 2, "substitutions": 7, "deletions": 2, "ref_words": 190 }, "telephony_snr10": { "si_sdr_db": 12.611677429964661, "pesq_wb": 1.8633883237838744, "stoi": 0.9001329422720643, "vad_f1": 0.9459190574496201, "vad_false_alarm": 0.04643360249724299, "wer": 0.12105263157894737, "insertions": 2, "substitutions": 18, "deletions": 3, "ref_words": 190 }, "telephony_alaw": { "si_sdr_db": 11.691927871933952, "pesq_wb": 1.7295814871788024, "stoi": 0.8737740927344824, "vad_f1": 0.9362633185965639, "vad_false_alarm": 0.041002393693906944, "wer": 0.48947368421052634, "insertions": 61, "substitutions": 28, "deletions": 4, "ref_words": 190 }, "telephony_g722": { "si_sdr_db": 18.580305233863008, "pesq_wb": 2.858854961395264, "stoi": 0.9641207778116307, "vad_f1": 0.9652604647870209, "vad_false_alarm": 0.06459809121692303, "wer": 0.06842105263157895, "insertions": 1, "substitutions": 12, "deletions": 0, "ref_words": 190 }, "telephony_opus12k": { "si_sdr_db": 5.270005307925586, "pesq_wb": 2.2206263542175293, "stoi": 0.9106134863892326, "vad_f1": 0.966958273782544, "vad_false_alarm": 0.04406370625512908, "wer": 0.08421052631578947, "insertions": 2, "substitutions": 12, "deletions": 2, "ref_words": 190 }, "telephony_loss3": { "si_sdr_db": 8.218723669207131, "pesq_wb": 1.6319350481033326, "stoi": 0.8530068047542269, "vad_f1": 0.9350656543242144, "vad_false_alarm": 0.04241791165623176, "wer": 0.2, "insertions": 2, "substitutions": 26, "deletions": 10, "ref_words": 190 }, "telephony_loss10": { "si_sdr_db": 5.8483149402955465, "pesq_wb": 1.4208144426345826, "stoi": 0.8039883829034217, "vad_f1": 0.9252340352623738, "vad_false_alarm": 0.04228854272952001, "wer": 0.3263157894736842, "insertions": 6, "substitutions": 46, "deletions": 10, "ref_words": 190 } }, "nativeSampleRate": 16000, "resampled": false }, { "id": "fastenhancer-t", "label": "FastEnhancer-T", "license": "MIT", "shippable": true, "rtf": 0.012719298236650675, "blockMsP99": 1.5898477664450195, "realtimeHeadroomP99": 10.063856639417002, "latencyMs": 16.0, "siSdrDb": 9.162791568404916, "siSdrDeltaDb": -1.7579406563866637, "pesqWb": 2.0369110067685443, "stoi": 0.8348230410899937, "estoi": 0.7383937833085584, "wer": 0.22456140350877193, "werSubstitutions": 514, "werInsertions": 145, "werDeletions": 109, "werRefWords": 3420, "vadF1": 0.9411334979174379, "vadFalseAlarm": 0.15766957819783697, "conditions": { "clean": { "si_sdr_db": 30.86589237932565, "pesq_wb": 4.3833733081817625, "stoi": 0.9954468352903098, "vad_f1": 0.9666838181089294, "vad_false_alarm": 0.08724581971269307, "wer": 0.06315789473684211, "insertions": 2, "substitutions": 9, "deletions": 1, "ref_words": 190 }, "noise_snr-5": { "si_sdr_db": 8.93559574968578, "pesq_wb": 1.5914408206939696, "stoi": 0.7985246095974191, "vad_f1": 0.9433551696343677, "vad_false_alarm": 0.05365350409425416, "wer": 0.24210526315789474, "insertions": 3, "substitutions": 39, "deletions": 4, "ref_words": 190 }, "noise_snr0": { "si_sdr_db": 13.382551989757554, "pesq_wb": 2.090251123905182, "stoi": 0.8870281067453533, "vad_f1": 0.9577605170422323, "vad_false_alarm": 0.05724875066252398, "wer": 0.10526315789473684, "insertions": 2, "substitutions": 17, "deletions": 1, "ref_words": 190 }, "noise_snr5": { "si_sdr_db": 15.732762847769184, "pesq_wb": 2.4338172316551208, "stoi": 0.9122563988357575, "vad_f1": 0.9648037870822685, "vad_false_alarm": 0.04392770615922425, "wer": 0.06842105263157895, "insertions": 1, "substitutions": 12, "deletions": 0, "ref_words": 190 }, "noise_snr10": { "si_sdr_db": 16.576306512328305, "pesq_wb": 2.7164655685424806, "stoi": 0.9435296213605116, "vad_f1": 0.9633761399270855, "vad_false_alarm": 0.05267319293479119, "wer": 0.05263157894736842, "insertions": 1, "substitutions": 9, "deletions": 0, "ref_words": 190 }, "noise_snr20": { "si_sdr_db": 22.609135641493836, "pesq_wb": 3.524846005439758, "stoi": 0.9820919434965794, "vad_f1": 0.9666896621410606, "vad_false_alarm": 0.07356986477149877, "wer": 0.042105263157894736, "insertions": 2, "substitutions": 6, "deletions": 0, "ref_words": 190 }, "babble_snr5": { "si_sdr_db": 6.362419950949018, "pesq_wb": 1.2362505197525024, "stoi": 0.7658401895695854, "vad_f1": 0.8559662388718697, "vad_false_alarm": 0.885740892003619, "wer": 0.3894736842105263, "insertions": 32, "substitutions": 33, "deletions": 9, "ref_words": 190 }, "competing_snr0": { "si_sdr_db": -0.6207483643258579, "pesq_wb": 1.1423205614089966, "stoi": 0.6766468753722773, "vad_f1": 0.899602895214499, "vad_false_alarm": 0.5668208226389345, "wer": 0.868421052631579, "insertions": 45, "substitutions": 87, "deletions": 33, "ref_words": 190 }, "competing_snr5": { "si_sdr_db": 5.141054406222373, "pesq_wb": 1.287438154220581, "stoi": 0.7940142357475412, "vad_f1": 0.9001131911973104, "vad_false_alarm": 0.5652997988267542, "wer": 0.4631578947368421, "insertions": 33, "substitutions": 33, "deletions": 22, "ref_words": 190 }, "reverb": { "si_sdr_db": -12.352974472508343, "pesq_wb": 1.2652860403060913, "stoi": 0.6425598867153429, "vad_f1": 0.9527805509568713, "vad_false_alarm": 0.09492847996937877, "wer": 0.21052631578947367, "insertions": 3, "substitutions": 34, "deletions": 3, "ref_words": 190 }, "reverb_snr10": { "si_sdr_db": -14.677006545492134, "pesq_wb": 1.2490818619728088, "stoi": 0.6262957155032878, "vad_f1": 0.9468638736513982, "vad_false_alarm": 0.083511584208375, "wer": 0.2894736842105263, "insertions": 5, "substitutions": 40, "deletions": 10, "ref_words": 190 }, "telephony": { "si_sdr_db": 14.495805757711071, "pesq_wb": 2.480261743068695, "stoi": 0.9209096002505861, "vad_f1": 0.9545402835712841, "vad_false_alarm": 0.05618961682560466, "wer": 0.12631578947368421, "insertions": 1, "substitutions": 19, "deletions": 4, "ref_words": 190 }, "telephony_snr10": { "si_sdr_db": 11.647561285068399, "pesq_wb": 1.6308165431022643, "stoi": 0.8398811569360122, "vad_f1": 0.9412494555526203, "vad_false_alarm": 0.023367166032038828, "wer": 0.1736842105263158, "insertions": 3, "substitutions": 29, "deletions": 1, "ref_words": 190 }, "telephony_alaw": { "si_sdr_db": 10.679982113752708, "pesq_wb": 1.5630768537521362, "stoi": 0.8212169593884658, "vad_f1": 0.9316559988106936, "vad_false_alarm": 0.024854343466531315, "wer": 0.21052631578947367, "insertions": 2, "substitutions": 35, "deletions": 3, "ref_words": 190 }, "telephony_g722": { "si_sdr_db": 16.6132873338346, "pesq_wb": 2.8592318058013917, "stoi": 0.952041437929849, "vad_f1": 0.9664999455237862, "vad_false_alarm": 0.06248539147997112, "wer": 0.05789473684210526, "insertions": 1, "substitutions": 10, "deletions": 0, "ref_words": 190 }, "telephony_opus12k": { "si_sdr_db": 5.124638171073222, "pesq_wb": 2.2783854722976686, "stoi": 0.8911895994188619, "vad_f1": 0.9653357140525506, "vad_false_alarm": 0.04857607922864995, "wer": 0.08947368421052632, "insertions": 1, "substitutions": 15, "deletions": 1, "ref_words": 190 }, "telephony_loss3": { "si_sdr_db": 8.468611336961834, "pesq_wb": 1.5348805069923401, "stoi": 0.8114282804604983, "vad_f1": 0.9318630451322856, "vad_false_alarm": 0.02819005151475672, "wer": 0.23684210526315788, "insertions": 2, "substitutions": 35, "deletions": 8, "ref_words": 190 }, "telephony_loss10": { "si_sdr_db": 5.9453721376812885, "pesq_wb": 1.3971740007400513, "stoi": 0.7659132870016488, "vad_f1": 0.9312626760427681, "vad_false_alarm": 0.029769343031466045, "wer": 0.3526315789473684, "insertions": 6, "substitutions": 52, "deletions": 9, "ref_words": 190 } }, "nativeSampleRate": 16000, "resampled": false }, { "id": "dfn3-hq", "label": "DeepFilterNet3 (96-frame warm-up)", "license": "MIT OR Apache-2.0", "shippable": true, "rtf": 0.22593318490966796, "blockMsP99": 20.75462536420671, "realtimeHeadroomP99": 0.48182030870313536, "latencyMs": 100.0, "siSdrDb": 8.317801769897711, "siSdrDeltaDb": -2.602930454893869, "pesqWb": 1.9597505331039429, "stoi": 0.8345496922461216, "estoi": 0.7467002222994543, "wer": 0.316374269005848, "werSubstitutions": 585, "werInsertions": 368, "werDeletions": 129, "werRefWords": 3420, "vadF1": 0.9142399530378845, "vadFalseAlarm": 0.22753456453369772, "conditions": { "clean": { "si_sdr_db": 24.030017023121644, "pesq_wb": 3.8402707815170287, "stoi": 0.9857279719777823, "vad_f1": 0.9360986893875168, "vad_false_alarm": 0.1692684998627181, "wer": 0.05263157894736842, "insertions": 2, "substitutions": 7, "deletions": 1, "ref_words": 190 }, "noise_snr-5": { "si_sdr_db": 9.648483539434165, "pesq_wb": 1.67253839969635, "stoi": 0.8350672104462034, "vad_f1": 0.9261450104900151, "vad_false_alarm": 0.1298462579001574, "wer": 0.34210526315789475, "insertions": 12, "substitutions": 48, "deletions": 5, "ref_words": 190 }, "noise_snr0": { "si_sdr_db": 12.979160287867966, "pesq_wb": 2.003598916530609, "stoi": 0.8944626866815609, "vad_f1": 0.9343800093823976, "vad_false_alarm": 0.13884014483050114, "wer": 0.14736842105263157, "insertions": 5, "substitutions": 23, "deletions": 0, "ref_words": 190 }, "noise_snr5": { "si_sdr_db": 14.958437243133218, "pesq_wb": 2.40952889919281, "stoi": 0.914274980607515, "vad_f1": 0.934412249993897, "vad_false_alarm": 0.13414231682950556, "wer": 0.14210526315789473, "insertions": 5, "substitutions": 22, "deletions": 0, "ref_words": 190 }, "noise_snr10": { "si_sdr_db": 15.700408943952095, "pesq_wb": 2.592140626907349, "stoi": 0.9405721834628608, "vad_f1": 0.9365538934717816, "vad_false_alarm": 0.12638295564827942, "wer": 0.06842105263157895, "insertions": 1, "substitutions": 12, "deletions": 0, "ref_words": 190 }, "noise_snr20": { "si_sdr_db": 19.66004934744376, "pesq_wb": 3.1675949811935427, "stoi": 0.9702427774477134, "vad_f1": 0.9355316846968706, "vad_false_alarm": 0.16237605213105896, "wer": 0.042105263157894736, "insertions": 1, "substitutions": 7, "deletions": 0, "ref_words": 190 }, "babble_snr5": { "si_sdr_db": 4.253545131754812, "pesq_wb": 1.217384946346283, "stoi": 0.7533000135832213, "vad_f1": 0.8500041792905947, "vad_false_alarm": 0.925784551273795, "wer": 0.4842105263157895, "insertions": 34, "substitutions": 55, "deletions": 3, "ref_words": 190 }, "competing_snr0": { "si_sdr_db": -1.2430980686788957, "pesq_wb": 1.140669572353363, "stoi": 0.6584787070667503, "vad_f1": 0.8919447790287733, "vad_false_alarm": 0.5804903911568184, "wer": 1.431578947368421, "insertions": 175, "substitutions": 91, "deletions": 6, "ref_words": 190 }, "competing_snr5": { "si_sdr_db": 4.416145385317329, "pesq_wb": 1.2846442818641663, "stoi": 0.783152932143565, "vad_f1": 0.8938889229328989, "vad_false_alarm": 0.5657146779539984, "wer": 0.7526315789473684, "insertions": 91, "substitutions": 41, "deletions": 11, "ref_words": 190 }, "reverb": { "si_sdr_db": -10.51527128348825, "pesq_wb": 1.1851179122924804, "stoi": 0.6084694566756635, "vad_f1": 0.8825847575856812, "vad_false_alarm": 0.14675663130372454, "wer": 0.43157894736842106, "insertions": 6, "substitutions": 51, "deletions": 25, "ref_words": 190 }, "reverb_snr10": { "si_sdr_db": -10.06076777615925, "pesq_wb": 1.178551983833313, "stoi": 0.5928682673304044, "vad_f1": 0.9006316937393082, "vad_false_alarm": 0.12369974154733454, "wer": 0.6, "insertions": 19, "substitutions": 67, "deletions": 28, "ref_words": 190 }, "telephony": { "si_sdr_db": 11.764231378932278, "pesq_wb": 2.346910309791565, "stoi": 0.9129413901674889, "vad_f1": 0.91882113721847, "vad_false_alarm": 0.1457169680366967, "wer": 0.15789473684210525, "insertions": 0, "substitutions": 24, "deletions": 6, "ref_words": 190 }, "telephony_snr10": { "si_sdr_db": 10.37913845562873, "pesq_wb": 1.764112377166748, "stoi": 0.8679266065089081, "vad_f1": 0.9184329136320004, "vad_false_alarm": 0.11763246572465899, "wer": 0.1631578947368421, "insertions": 2, "substitutions": 22, "deletions": 7, "ref_words": 190 }, "telephony_alaw": { "si_sdr_db": 9.580780982388367, "pesq_wb": 1.5953302025794982, "stoi": 0.848386875289606, "vad_f1": 0.9142942556059177, "vad_false_alarm": 0.11332284007301852, "wer": 0.1736842105263158, "insertions": 2, "substitutions": 28, "deletions": 3, "ref_words": 190 }, "telephony_g722": { "si_sdr_db": 15.173965222194761, "pesq_wb": 2.5773364186286924, "stoi": 0.943958462947046, "vad_f1": 0.9347436832223854, "vad_false_alarm": 0.17487761621319514, "wer": 0.06842105263157895, "insertions": 3, "substitutions": 10, "deletions": 0, "ref_words": 190 }, "telephony_opus12k": { "si_sdr_db": 4.888668899716424, "pesq_wb": 2.3249027967453, "stoi": 0.8962969953031396, "vad_f1": 0.9353605962179072, "vad_false_alarm": 0.13879215661680241, "wer": 0.1, "insertions": 3, "substitutions": 13, "deletions": 3, "ref_words": 190 }, "telephony_loss3": { "si_sdr_db": 8.625267225527562, "pesq_wb": 1.5988064050674438, "stoi": 0.8287161535917897, "vad_f1": 0.9097221457532649, "vad_false_alarm": 0.10226976829467996, "wer": 0.2, "insertions": 2, "substitutions": 23, "deletions": 13, "ref_words": 190 }, "telephony_loss10": { "si_sdr_db": 5.481269920072091, "pesq_wb": 1.3760697841644287, "stoi": 0.7870507891989709, "vad_f1": 0.9027685530322402, "vad_false_alarm": 0.09970812620961592, "wer": 0.3368421052631579, "insertions": 5, "substitutions": 41, "deletions": 18, "ref_words": 190 } }, "nativeSampleRate": 48000, "resampled": true }, { "id": "dfn3", "label": "DeepFilterNet3", "license": "MIT OR Apache-2.0", "shippable": true, "rtf": 0.13302028982254127, "blockMsP99": 13.651765545364476, "realtimeHeadroomP99": 0.7325059873589427, "latencyMs": 100.0, "siSdrDb": 7.581571113052651, "siSdrDeltaDb": -3.339161111738929, "pesqWb": 1.8693116068840026, "stoi": 0.8165857319385377, "estoi": 0.7275854841138771, "wer": 0.35175438596491226, "werSubstitutions": 626, "werInsertions": 380, "werDeletions": 197, "werRefWords": 3420, "vadF1": 0.9013407581938893, "vadFalseAlarm": 0.22185911710426998, "conditions": { "clean": { "si_sdr_db": 22.45828011726199, "pesq_wb": 3.667029309272766, "stoi": 0.977353111578281, "vad_f1": 0.9344634830661823, "vad_false_alarm": 0.1692684998627181, "wer": 0.04736842105263158, "insertions": 1, "substitutions": 7, "deletions": 1, "ref_words": 190 }, "noise_snr-5": { "si_sdr_db": 8.786601108017967, "pesq_wb": 1.527800941467285, "stoi": 0.8065894385235653, "vad_f1": 0.917273489908753, "vad_false_alarm": 0.10080445428748104, "wer": 0.32105263157894737, "insertions": 10, "substitutions": 48, "deletions": 3, "ref_words": 190 }, "noise_snr0": { "si_sdr_db": 12.136575288827661, "pesq_wb": 1.9170676827430726, "stoi": 0.8834410458614794, "vad_f1": 0.934640575037475, "vad_false_alarm": 0.12288608193714445, "wer": 0.1368421052631579, "insertions": 4, "substitutions": 20, "deletions": 2, "ref_words": 190 }, "noise_snr5": { "si_sdr_db": 14.53282294697107, "pesq_wb": 2.27543306350708, "stoi": 0.9047345141187281, "vad_f1": 0.931097136592846, "vad_false_alarm": 0.13011499502599427, "wer": 0.14210526315789473, "insertions": 4, "substitutions": 23, "deletions": 0, "ref_words": 190 }, "noise_snr10": { "si_sdr_db": 15.067988968507175, "pesq_wb": 2.4277612447738646, "stoi": 0.9300565695589865, "vad_f1": 0.9294361150908566, "vad_false_alarm": 0.13400594715946113, "wer": 0.1, "insertions": 1, "substitutions": 17, "deletions": 1, "ref_words": 190 }, "noise_snr20": { "si_sdr_db": 17.600708146077448, "pesq_wb": 2.929212486743927, "stoi": 0.9570989647068113, "vad_f1": 0.9320230637003544, "vad_false_alarm": 0.16500763107842736, "wer": 0.04736842105263158, "insertions": 1, "substitutions": 8, "deletions": 0, "ref_words": 190 }, "babble_snr5": { "si_sdr_db": 5.021942074116824, "pesq_wb": 1.2186425566673278, "stoi": 0.7555954696110951, "vad_f1": 0.8501436209011339, "vad_false_alarm": 0.9235918334215316, "wer": 0.35789473684210527, "insertions": 8, "substitutions": 35, "deletions": 25, "ref_words": 190 }, "competing_snr0": { "si_sdr_db": -1.228088847532652, "pesq_wb": 1.1413809657096863, "stoi": 0.6618649416642282, "vad_f1": 0.8930722311313269, "vad_false_alarm": 0.5851163493910241, "wer": 1.9947368421052631, "insertions": 271, "substitutions": 97, "deletions": 11, "ref_words": 190 }, "competing_snr5": { "si_sdr_db": 4.196297361483498, "pesq_wb": 1.255443274974823, "stoi": 0.7736127450790905, "vad_f1": 0.8944450385946545, "vad_false_alarm": 0.5745996540065439, "wer": 0.5, "insertions": 38, "substitutions": 47, "deletions": 10, "ref_words": 190 }, "reverb": { "si_sdr_db": -11.636951723924387, "pesq_wb": 1.128217387199402, "stoi": 0.5221157792751007, "vad_f1": 0.8101647187084605, "vad_false_alarm": 0.11145347037808477, "wer": 0.6, "insertions": 7, "substitutions": 67, "deletions": 40, "ref_words": 190 }, "reverb_snr10": { "si_sdr_db": -12.78483479809291, "pesq_wb": 1.1102552652359008, "stoi": 0.5134229363311036, "vad_f1": 0.7793212008443005, "vad_false_alarm": 0.09484392770931992, "wer": 0.8210526315789474, "insertions": 16, "substitutions": 79, "deletions": 61, "ref_words": 190 }, "telephony": { "si_sdr_db": 11.514546326678804, "pesq_wb": 2.2686702489852903, "stoi": 0.9065281570785431, "vad_f1": 0.9174853543173865, "vad_false_alarm": 0.14605093386603654, "wer": 0.15263157894736842, "insertions": 0, "substitutions": 25, "deletions": 4, "ref_words": 190 }, "telephony_snr10": { "si_sdr_db": 10.013812820414266, "pesq_wb": 1.7431542515754699, "stoi": 0.8642276707368157, "vad_f1": 0.9176491008034311, "vad_false_alarm": 0.11788862606853864, "wer": 0.2, "insertions": 4, "substitutions": 29, "deletions": 5, "ref_words": 190 }, "telephony_alaw": { "si_sdr_db": 8.776289558270799, "pesq_wb": 1.5609478950500488, "stoi": 0.8355851339636103, "vad_f1": 0.9129433069684186, "vad_false_alarm": 0.11202009316880508, "wer": 0.18421052631578946, "insertions": 2, "substitutions": 29, "deletions": 4, "ref_words": 190 }, "telephony_g722": { "si_sdr_db": 15.124523260048473, "pesq_wb": 2.427763032913208, "stoi": 0.9348248158054208, "vad_f1": 0.9322514680591638, "vad_false_alarm": 0.16357829079187505, "wer": 0.06315789473684211, "insertions": 3, "substitutions": 9, "deletions": 0, "ref_words": 190 }, "telephony_opus12k": { "si_sdr_db": 4.897918755962683, "pesq_wb": 2.19540559053421, "stoi": 0.8898843663913694, "vad_f1": 0.932103062108672, "vad_false_alarm": 0.13752384261369555, "wer": 0.11052631578947368, "insertions": 3, "substitutions": 17, "deletions": 1, "ref_words": 190 }, "telephony_loss3": { "si_sdr_db": 7.246136469118445, "pesq_wb": 1.5176615118980408, "stoi": 0.8161518946084213, "vad_f1": 0.9080583202503478, "vad_false_alarm": 0.10358555776836417, "wer": 0.24736842105263157, "insertions": 4, "substitutions": 33, "deletions": 10, "ref_words": 190 }, "telephony_loss10": { "si_sdr_db": 4.743712202740573, "pesq_wb": 1.3357622146606445, "stoi": 0.7654556200010274, "vad_f1": 0.8975623614062421, "vad_false_alarm": 0.10112391934181346, "wer": 0.30526315789473685, "insertions": 3, "substitutions": 36, "deletions": 19, "ref_words": 190 } }, "nativeSampleRate": 48000, "resampled": true }, { "id": "clearline", "label": "Clearline (8 kHz, resampled)", "license": "internal", "shippable": false, "rtf": 0.04810167003325479, "blockMsP99": 4.082023273804261, "realtimeHeadroomP99": 3.9196248837377947, "latencyMs": 16.0, "siSdrDb": 0.6915653633410099, "siSdrDeltaDb": -10.22916686145057, "pesqWb": 1.6845682044823964, "stoi": 0.7438748199572827, "estoi": 0.6603211022006628, "wer": 0.39649122807017545, "werSubstitutions": 776, "werInsertions": 410, "werDeletions": 170, "werRefWords": 3420, "vadF1": 0.8352046826888598, "vadFalseAlarm": 0.1513078868219344, "conditions": { "clean": { "si_sdr_db": 16.174847672139514, "pesq_wb": 3.540698838233948, "stoi": 0.9941585473752562, "vad_f1": 0.9636592006440843, "vad_false_alarm": 0.05628302376274158, "wer": 0.05789473684210526, "insertions": 1, "substitutions": 9, "deletions": 1, "ref_words": 190 }, "noise_snr-5": { "si_sdr_db": -11.398264191677125, "pesq_wb": 1.152290940284729, "stoi": 0.5384032589719093, "vad_f1": 0.5584285632892494, "vad_false_alarm": 0.07300209318571076, "wer": 0.6368421052631579, "insertions": 9, "substitutions": 82, "deletions": 30, "ref_words": 190 }, "noise_snr0": { "si_sdr_db": -2.257317254879139, "pesq_wb": 1.270313274860382, "stoi": 0.7364384084071928, "vad_f1": 0.8839959290566209, "vad_false_alarm": 0.06500871047714056, "wer": 0.2894736842105263, "insertions": 8, "substitutions": 44, "deletions": 3, "ref_words": 190 }, "noise_snr5": { "si_sdr_db": 0.19667999312854023, "pesq_wb": 1.5918202638626098, "stoi": 0.6942321295266751, "vad_f1": 0.7515542213937226, "vad_false_alarm": 0.03119914383391593, "wer": 0.33157894736842103, "insertions": 14, "substitutions": 33, "deletions": 16, "ref_words": 190 }, "noise_snr10": { "si_sdr_db": 7.897187803836827, "pesq_wb": 1.7518604040145873, "stoi": 0.8842469140053184, "vad_f1": 0.9443150125262626, "vad_false_alarm": 0.06703351482728162, "wer": 0.10526315789473684, "insertions": 2, "substitutions": 16, "deletions": 2, "ref_words": 190 }, "noise_snr20": { "si_sdr_db": 13.926220744101238, "pesq_wb": 2.312462604045868, "stoi": 0.9748244876524206, "vad_f1": 0.9614350998059799, "vad_false_alarm": 0.058599127332328325, "wer": 0.06842105263157895, "insertions": 1, "substitutions": 12, "deletions": 0, "ref_words": 190 }, "babble_snr5": { "si_sdr_db": 1.020970438015269, "pesq_wb": 1.2842444062232972, "stoi": 0.6591956495754716, "vad_f1": 0.8412910206701305, "vad_false_alarm": 0.8245655397508633, "wer": 1.3842105263157896, "insertions": 169, "substitutions": 67, "deletions": 27, "ref_words": 190 }, "competing_snr0": { "si_sdr_db": -12.567017454429052, "pesq_wb": 1.1190677762031556, "stoi": 0.4114872672734804, "vad_f1": 0.8340497505651937, "vad_false_alarm": 0.5470619226644609, "wer": 1.068421052631579, "insertions": 36, "substitutions": 144, "deletions": 23, "ref_words": 190 }, "competing_snr5": { "si_sdr_db": -7.8192606235432836, "pesq_wb": 1.1977710843086242, "stoi": 0.4953814542471215, "vad_f1": 0.8456747668106559, "vad_false_alarm": 0.5064379447044965, "wer": 0.8736842105263158, "insertions": 29, "substitutions": 120, "deletions": 17, "ref_words": 190 }, "reverb": { "si_sdr_db": -16.347363836157847, "pesq_wb": 1.3191485285758973, "stoi": 0.6697790107649584, "vad_f1": 0.9499291448779305, "vad_false_alarm": 0.0775710381248718, "wer": 0.12105263157894737, "insertions": 2, "substitutions": 18, "deletions": 3, "ref_words": 190 }, "reverb_snr10": { "si_sdr_db": -22.37184413469708, "pesq_wb": 1.1343780040740967, "stoi": 0.4687588574214157, "vad_f1": 0.40941952940389925, "vad_false_alarm": 0.03473534263007947, "wer": 1.1631578947368422, "insertions": 117, "substitutions": 75, "deletions": 29, "ref_words": 190 }, "telephony": { "si_sdr_db": 16.008687476162002, "pesq_wb": 3.3375594854354858, "stoi": 0.9900126113706813, "vad_f1": 0.9650165632066645, "vad_false_alarm": 0.04909296296443336, "wer": 0.05789473684210526, "insertions": 1, "substitutions": 9, "deletions": 1, "ref_words": 190 }, "telephony_snr10": { "si_sdr_db": 6.644479985358816, "pesq_wb": 1.5577993988990784, "stoi": 0.811030498959824, "vad_f1": 0.9077547314304052, "vad_false_alarm": 0.05754845231043152, "wer": 0.14736842105263157, "insertions": 4, "substitutions": 21, "deletions": 3, "ref_words": 190 }, "telephony_alaw": { "si_sdr_db": 6.925573429558627, "pesq_wb": 1.6052649974823, "stoi": 0.8581832326854182, "vad_f1": 0.8641218393092366, "vad_false_alarm": 0.04977519763543309, "wer": 0.1368421052631579, "insertions": 4, "substitutions": 22, "deletions": 0, "ref_words": 190 }, "telephony_g722": { "si_sdr_db": 6.205060985312767, "pesq_wb": 1.655719769001007, "stoi": 0.8396955167140469, "vad_f1": 0.8669201896983616, "vad_false_alarm": 0.06384447644207934, "wer": 0.09473684210526316, "insertions": 2, "substitutions": 13, "deletions": 3, "ref_words": 190 }, "telephony_opus12k": { "si_sdr_db": 0.5436683302396027, "pesq_wb": 1.5959327220916748, "stoi": 0.7512323692105696, "vad_f1": 0.7223385358402183, "vad_false_alarm": 0.036662546660987026, "wer": 0.2736842105263158, "insertions": 8, "substitutions": 39, "deletions": 5, "ref_words": 190 }, "telephony_loss3": { "si_sdr_db": 5.9521113330086735, "pesq_wb": 1.5184477686882019, "stoi": 0.8265683651173914, "vad_f1": 0.8872113747105181, "vad_false_alarm": 0.054581839293268396, "wer": 0.17894736842105263, "insertions": 1, "substitutions": 30, "deletions": 3, "ref_words": 190 }, "telephony_loss10": { "si_sdr_db": 3.713755844659828, "pesq_wb": 1.3774474143981934, "stoi": 0.7861181799519361, "vad_f1": 0.8765688151603429, "vad_false_alarm": 0.07053908619429598, "wer": 0.14736842105263157, "insertions": 2, "substitutions": 22, "deletions": 4, "ref_words": 190 } }, "nativeSampleRate": 8000, "resampled": true } ], "controls": [ { "id": "resample8k", "label": "Resample-only 16\u21928\u219216 (control)", "license": "\u2014", "shippable": true, "rtf": 4.921931203743445e-05, "blockMsP99": 0.00633677409496158, "realtimeHeadroomP99": 2524.9440425407824, "latencyMs": 0.0, "siSdrDb": 2.8910241024588577, "siSdrDeltaDb": -8.029708122332723, "pesqWb": 1.6827420009507073, "stoi": 0.8399078346600883, "estoi": 0.7303222259005707, "wer": 0.18625730994152045, "werSubstitutions": 413, "werInsertions": 173, "werDeletions": 51, "werRefWords": 3420, "vadF1": 0.947758705466079, "vadFalseAlarm": 0.15743141907060987, "conditions": { "clean": { "si_sdr_db": 16.181072759957154, "pesq_wb": 3.5533412218093874, "stoi": 0.9941955580772441, "vad_f1": 0.9712635637963618, "vad_false_alarm": 0.05200245311745607, "wer": 0.05789473684210526, "insertions": 1, "substitutions": 9, "deletions": 1, "ref_words": 190 }, "noise_snr-5": { "si_sdr_db": -4.440188248246602, "pesq_wb": 1.1188627481460571, "stoi": 0.7164119054336003, "vad_f1": 0.9377278146663655, "vad_false_alarm": 0.06968870677046633, "wer": 0.22631578947368422, "insertions": 2, "substitutions": 25, "deletions": 16, "ref_words": 190 }, "noise_snr0": { "si_sdr_db": 0.04849369219904885, "pesq_wb": 1.2397638082504272, "stoi": 0.816136061114556, "vad_f1": 0.9545007401758253, "vad_false_alarm": 0.0595259879130645, "wer": 0.11052631578947368, "insertions": 2, "substitutions": 17, "deletions": 2, "ref_words": 190 }, "noise_snr5": { "si_sdr_db": 4.726079777541773, "pesq_wb": 1.546132779121399, "stoi": 0.8549254440394009, "vad_f1": 0.9582578022161936, "vad_false_alarm": 0.044553263199557255, "wer": 0.15263157894736842, "insertions": 3, "substitutions": 22, "deletions": 4, "ref_words": 190 }, "noise_snr10": { "si_sdr_db": 8.94335304688487, "pesq_wb": 1.6530146718025207, "stoi": 0.9100971104837668, "vad_f1": 0.9615197503069066, "vad_false_alarm": 0.0481076787123384, "wer": 0.07894736842105263, "insertions": 2, "substitutions": 12, "deletions": 1, "ref_words": 190 }, "noise_snr20": { "si_sdr_db": 13.963746022894602, "pesq_wb": 2.2431580066680907, "stoi": 0.9743394182501575, "vad_f1": 0.9663056465859977, "vad_false_alarm": 0.05101900866322123, "wer": 0.06315789473684211, "insertions": 1, "substitutions": 11, "deletions": 0, "ref_words": 190 }, "babble_snr5": { "si_sdr_db": 4.1275561654747746, "pesq_wb": 1.2039869785308839, "stoi": 0.7713887972245035, "vad_f1": 0.8457910902708192, "vad_false_alarm": 0.9675959456163833, "wer": 0.39473684210526316, "insertions": 19, "substitutions": 50, "deletions": 6, "ref_words": 190 }, "competing_snr0": { "si_sdr_db": -0.45506378851821355, "pesq_wb": 1.1940839529037475, "stoi": 0.7295058628065306, "vad_f1": 0.8999935792990227, "vad_false_alarm": 0.5502636515372893, "wer": 0.8263157894736842, "insertions": 78, "substitutions": 78, "deletions": 1, "ref_words": 190 }, "competing_snr5": { "si_sdr_db": 4.16349047900726, "pesq_wb": 1.3125433087348939, "stoi": 0.8163737176940199, "vad_f1": 0.9014767841893823, "vad_false_alarm": 0.5514102682937138, "wer": 0.47368421052631576, "insertions": 46, "substitutions": 40, "deletions": 4, "ref_words": 190 }, "reverb": { "si_sdr_db": -27.134763535168094, "pesq_wb": 1.328452968597412, "stoi": 0.6274501427804513, "vad_f1": 0.9583328025791953, "vad_false_alarm": 0.05963797890143611, "wer": 0.12105263157894737, "insertions": 2, "substitutions": 19, "deletions": 2, "ref_words": 190 }, "reverb_snr10": { "si_sdr_db": -27.359511441429383, "pesq_wb": 1.198651623725891, "stoi": 0.5831639824613584, "vad_f1": 0.9554963044744713, "vad_false_alarm": 0.06122488482456919, "wer": 0.23157894736842105, "insertions": 6, "substitutions": 33, "deletions": 5, "ref_words": 190 }, "telephony": { "si_sdr_db": 16.015048923813037, "pesq_wb": 3.3414860486984255, "stoi": 0.9900642691359955, "vad_f1": 0.9714686487029821, "vad_false_alarm": 0.04815629927130223, "wer": 0.042105263157894736, "insertions": 1, "substitutions": 6, "deletions": 1, "ref_words": 190 }, "telephony_snr10": { "si_sdr_db": 8.887789327264034, "pesq_wb": 1.519613540172577, "stoi": 0.8976657510049035, "vad_f1": 0.9677159386790878, "vad_false_alarm": 0.036804818945029405, "wer": 0.07894736842105263, "insertions": 2, "substitutions": 12, "deletions": 1, "ref_words": 190 }, "telephony_alaw": { "si_sdr_db": 8.719387345134914, "pesq_wb": 1.5844810128211975, "stoi": 0.9064668702092649, "vad_f1": 0.9655068779155312, "vad_false_alarm": 0.039532012162843834, "wer": 0.08947368421052632, "insertions": 3, "substitutions": 13, "deletions": 1, "ref_words": 190 }, "telephony_g722": { "si_sdr_db": 8.737019027067662, "pesq_wb": 1.6834400177001954, "stoi": 0.9179976963635257, "vad_f1": 0.964512414320537, "vad_false_alarm": 0.0585150530804621, "wer": 0.07368421052631578, "insertions": 1, "substitutions": 12, "deletions": 1, "ref_words": 190 }, "telephony_opus12k": { "si_sdr_db": 4.307973824737365, "pesq_wb": 1.5951621532440186, "stoi": 0.8689518174275583, "vad_f1": 0.964156603322062, "vad_false_alarm": 0.03710981279825136, "wer": 0.11052631578947368, "insertions": 1, "substitutions": 18, "deletions": 2, "ref_words": 190 }, "telephony_loss3": { "si_sdr_db": 7.535812367085562, "pesq_wb": 1.5400220274925231, "stoi": 0.8816538276747782, "vad_f1": 0.9588668108994429, "vad_false_alarm": 0.045096973984416894, "wer": 0.11578947368421053, "insertions": 2, "substitutions": 19, "deletions": 1, "ref_words": 190 }, "telephony_loss10": { "si_sdr_db": 5.071138098559676, "pesq_wb": 1.4331591486930848, "stoi": 0.8615527916999751, "vad_f1": 0.9567635259892378, "vad_false_alarm": 0.05352074547917619, "wer": 0.10526315789473684, "insertions": 1, "substitutions": 17, "deletions": 2, "ref_words": 190 } }, "nativeSampleRate": 8000, "resampled": true } ], "splitPipeline": { "passthrough": { "label": "Passthrough (control)", "mean_wer_enhanced": 0.1476608187134503, "mean_wer_raw": 0.1476608187134503, "wer_delta_pp": 0.0, "conditions_where_enhancement_helps_stt": 0, "conditions_scored": 18, "mean_vad_f1_enhanced": 0.9503607044937831, "mean_vad_f1_raw": 0.9503607044937831, "vad_f1_delta": 0.0, "insertions_enhanced": 101, "insertions_raw": 101, "per_condition": { "clean": { "vad_f1_enhanced": 0.9756248909923848, "vad_f1_raw": 0.9756248909923848, "vad_fa_enhanced": 0.07023743685253026, "vad_fa_raw": 0.07023743685253026, "wer_enhanced": 0.04736842105263158, "wer_raw": 0.04736842105263158, "ins_enhanced": 2, "ins_raw": 2 }, "noise_snr-5": { "vad_f1_enhanced": 0.9284106039361377, "vad_f1_raw": 0.9284106039361377, "vad_fa_enhanced": 0.06810271419067943, "vad_fa_raw": 0.06810271419067943, "wer_enhanced": 0.18421052631578946, "wer_raw": 0.18421052631578946, "ins_enhanced": 4, "ins_raw": 4 }, "noise_snr0": { "vad_f1_enhanced": 0.9650457996816915, "vad_f1_raw": 0.9650457996816915, "vad_fa_enhanced": 0.07585414368555106, "vad_fa_raw": 0.07585414368555106, "wer_enhanced": 0.11052631578947368, "wer_raw": 0.11052631578947368, "ins_enhanced": 2, "ins_raw": 2 }, "noise_snr5": { "vad_f1_enhanced": 0.9655544660175688, "vad_f1_raw": 0.9655544660175688, "vad_fa_enhanced": 0.06337917413073132, "vad_fa_raw": 0.06337917413073132, "wer_enhanced": 0.08947368421052632, "wer_raw": 0.08947368421052632, "ins_enhanced": 2, "ins_raw": 2 }, "noise_snr10": { "vad_f1_enhanced": 0.9715691872453027, "vad_f1_raw": 0.9715691872453027, "vad_fa_enhanced": 0.0521201836224206, "vad_fa_raw": 0.0521201836224206, "wer_enhanced": 0.06315789473684211, "wer_raw": 0.06315789473684211, "ins_enhanced": 2, "ins_raw": 2 }, "noise_snr20": { "vad_f1_enhanced": 0.9722639322945248, "vad_f1_raw": 0.9722639322945248, "vad_fa_enhanced": 0.0769268153848264, "vad_fa_raw": 0.0769268153848264, "wer_enhanced": 0.042105263157894736, "wer_raw": 0.042105263157894736, "ins_enhanced": 2, "ins_raw": 2 }, "babble_snr5": { "vad_f1_enhanced": 0.8446580505770814, "vad_f1_raw": 0.8446580505770814, "vad_fa_enhanced": 0.9770890756151038, "vad_fa_raw": 0.9770890756151038, "wer_enhanced": 0.24736842105263157, "wer_raw": 0.24736842105263157, "ins_enhanced": 10, "ins_raw": 10 }, "competing_snr0": { "vad_f1_enhanced": 0.8978559481992789, "vad_f1_raw": 0.8978559481992789, "vad_fa_enhanced": 0.5817642991147487, "vad_fa_raw": 0.5817642991147487, "wer_enhanced": 0.8473684210526315, "wer_raw": 0.8473684210526315, "ins_enhanced": 45, "ins_raw": 45 }, "competing_snr5": { "vad_f1_enhanced": 0.8993866413420953, "vad_f1_raw": 0.8993866413420953, "vad_fa_enhanced": 0.5812582262402548, "vad_fa_raw": 0.5812582262402548, "wer_enhanced": 0.2631578947368421, "wer_raw": 0.2631578947368421, "ins_enhanced": 19, "ins_raw": 19 }, "reverb": { "vad_f1_enhanced": 0.9623023734900757, "vad_f1_raw": 0.9623023734900757, "vad_fa_enhanced": 0.07937271734601117, "vad_fa_raw": 0.07937271734601117, "wer_enhanced": 0.05789473684210526, "wer_raw": 0.05789473684210526, "ins_enhanced": 1, "ins_raw": 1 }, "reverb_snr10": { "vad_f1_enhanced": 0.9565243987100389, "vad_f1_raw": 0.9565243987100389, "vad_fa_enhanced": 0.07830285083543445, "vad_fa_raw": 0.07830285083543445, "wer_enhanced": 0.1368421052631579, "wer_raw": 0.1368421052631579, "ins_enhanced": 0, "ins_raw": 0 }, "telephony": { "vad_f1_enhanced": 0.9761642766503543, "vad_f1_raw": 0.9761642766503543, "vad_fa_enhanced": 0.05308086433332433, "vad_fa_raw": 0.05308086433332433, "wer_enhanced": 0.05789473684210526, "wer_raw": 0.05789473684210526, "ins_enhanced": 1, "ins_raw": 1 }, "telephony_snr10": { "vad_f1_enhanced": 0.9688306828446971, "vad_f1_raw": 0.9688306828446971, "vad_fa_enhanced": 0.05104087749954463, "vad_fa_raw": 0.05104087749954463, "wer_enhanced": 0.10526315789473684, "wer_raw": 0.10526315789473684, "ins_enhanced": 3, "ins_raw": 3 }, "telephony_alaw": { "vad_f1_enhanced": 0.9649250943221878, "vad_f1_raw": 0.9649250943221878, "vad_fa_enhanced": 0.04056293999789538, "vad_fa_raw": 0.04056293999789538, "wer_enhanced": 0.07368421052631578, "wer_raw": 0.07368421052631578, "ins_enhanced": 2, "ins_raw": 2 }, "telephony_g722": { "vad_f1_enhanced": 0.9718918630456168, "vad_f1_raw": 0.9718918630456168, "vad_fa_enhanced": 0.07682802264269852, "vad_fa_raw": 0.07682802264269852, "wer_enhanced": 0.05263157894736842, "wer_raw": 0.05263157894736842, "ins_enhanced": 1, "ins_raw": 1 }, "telephony_opus12k": { "vad_f1_enhanced": 0.9700698810886056, "vad_f1_raw": 0.9700698810886056, "vad_fa_enhanced": 0.06495112394829519, "vad_fa_raw": 0.06495112394829519, "wer_enhanced": 0.06842105263157895, "wer_raw": 0.06842105263157895, "ins_enhanced": 2, "ins_raw": 2 }, "telephony_loss3": { "vad_f1_enhanced": 0.9585494127803207, "vad_f1_raw": 0.9585494127803207, "vad_fa_enhanced": 0.045096973984416894, "vad_fa_raw": 0.045096973984416894, "wer_enhanced": 0.11052631578947368, "wer_raw": 0.11052631578947368, "ins_enhanced": 2, "ins_raw": 2 }, "telephony_loss10": { "vad_f1_enhanced": 0.9568651776701316, "vad_f1_raw": 0.9568651776701316, "vad_fa_enhanced": 0.052286177577941614, "vad_fa_raw": 0.052286177577941614, "wer_enhanced": 0.1, "wer_raw": 0.1, "ins_enhanced": 1, "ins_raw": 1 } } }, "gtcrn": { "label": "GTCRN", "mean_wer_enhanced": 0.2026315789473684, "mean_wer_raw": 0.1476608187134503, "wer_delta_pp": 5.497076023391811, "conditions_where_enhancement_helps_stt": 1, "conditions_scored": 18, "mean_vad_f1_enhanced": 0.9472170326540659, "mean_vad_f1_raw": 0.9503607044937831, "vad_f1_delta": -0.003143671839717199, "insertions_enhanced": 210, "insertions_raw": 101, "per_condition": { "clean": { "vad_f1_enhanced": 0.9660543915503998, "vad_f1_raw": 0.9756248909923848, "vad_fa_enhanced": 0.08697693497142549, "vad_fa_raw": 0.07023743685253026, "wer_enhanced": 0.06315789473684211, "wer_raw": 0.04736842105263158, "ins_enhanced": 2, "ins_raw": 2 }, "noise_snr-5": { "vad_f1_enhanced": 0.9552931392660063, "vad_f1_raw": 0.9284106039361377, "vad_fa_enhanced": 0.06843675237866058, "vad_fa_raw": 0.06810271419067943, "wer_enhanced": 0.19473684210526315, "wer_raw": 0.18421052631578946, "ins_enhanced": 4, "ins_raw": 4 }, "noise_snr0": { "vad_f1_enhanced": 0.9624445080427307, "vad_f1_raw": 0.9650457996816915, "vad_fa_enhanced": 0.05563103296541287, "vad_fa_raw": 0.07585414368555106, "wer_enhanced": 0.08421052631578947, "wer_raw": 0.11052631578947368, "ins_enhanced": 3, "ins_raw": 2 }, "noise_snr5": { "vad_f1_enhanced": 0.9654063506734911, "vad_f1_raw": 0.9655544660175688, "vad_fa_enhanced": 0.047388994708232095, "vad_fa_raw": 0.06337917413073132, "wer_enhanced": 0.1, "wer_raw": 0.08947368421052632, "ins_enhanced": 3, "ins_raw": 2 }, "noise_snr10": { "vad_f1_enhanced": 0.964118704379777, "vad_f1_raw": 0.9715691872453027, "vad_fa_enhanced": 0.0550873866267928, "vad_fa_raw": 0.0521201836224206, "wer_enhanced": 0.08421052631578947, "wer_raw": 0.06315789473684211, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr20": { "vad_f1_enhanced": 0.9661604272190638, "vad_f1_raw": 0.9722639322945248, "vad_fa_enhanced": 0.07713223243072398, "vad_fa_raw": 0.0769268153848264, "wer_enhanced": 0.06315789473684211, "wer_raw": 0.042105263157894736, "ins_enhanced": 2, "ins_raw": 2 }, "babble_snr5": { "vad_f1_enhanced": 0.8492274596155684, "vad_f1_raw": 0.8446580505770814, "vad_fa_enhanced": 0.9190328053687009, "vad_fa_raw": 0.9770890756151038, "wer_enhanced": 0.6105263157894737, "wer_raw": 0.24736842105263157, "ins_enhanced": 63, "ins_raw": 10 }, "competing_snr0": { "vad_f1_enhanced": 0.8983963703575533, "vad_f1_raw": 0.8978559481992789, "vad_fa_enhanced": 0.5711054910377819, "vad_fa_raw": 0.5817642991147487, "wer_enhanced": 0.8789473684210526, "wer_raw": 0.8473684210526315, "ins_enhanced": 64, "ins_raw": 45 }, "competing_snr5": { "vad_f1_enhanced": 0.8999027489965339, "vad_f1_raw": 0.8993866413420953, "vad_fa_enhanced": 0.5657796308855336, "vad_fa_raw": 0.5812582262402548, "wer_enhanced": 0.4473684210526316, "wer_raw": 0.2631578947368421, "ins_enhanced": 46, "ins_raw": 19 }, "reverb": { "vad_f1_enhanced": 0.9628998075243203, "vad_f1_raw": 0.9623023734900757, "vad_fa_enhanced": 0.09150975804156156, "vad_fa_raw": 0.07937271734601117, "wer_enhanced": 0.06842105263157895, "wer_raw": 0.05789473684210526, "ins_enhanced": 1, "ins_raw": 1 }, "reverb_snr10": { "vad_f1_enhanced": 0.961264056628323, "vad_f1_raw": 0.9565243987100389, "vad_fa_enhanced": 0.07846796133865161, "vad_fa_raw": 0.07830285083543445, "wer_enhanced": 0.14736842105263157, "wer_raw": 0.1368421052631579, "ins_enhanced": 3, "ins_raw": 0 }, "telephony": { "vad_f1_enhanced": 0.9602097278763091, "vad_f1_raw": 0.9761642766503543, "vad_fa_enhanced": 0.056051811379656355, "vad_fa_raw": 0.05308086433332433, "wer_enhanced": 0.08947368421052632, "wer_raw": 0.05789473684210526, "ins_enhanced": 2, "ins_raw": 1 }, "telephony_snr10": { "vad_f1_enhanced": 0.9566377472547558, "vad_f1_raw": 0.9688306828446971, "vad_fa_enhanced": 0.037106203972974955, "vad_fa_raw": 0.05104087749954463, "wer_enhanced": 0.11052631578947368, "wer_raw": 0.10526315789473684, "ins_enhanced": 3, "ins_raw": 3 }, "telephony_alaw": { "vad_f1_enhanced": 0.9555179803091851, "vad_f1_raw": 0.9649250943221878, "vad_fa_enhanced": 0.04011145552226702, "vad_fa_raw": 0.04056293999789538, "wer_enhanced": 0.1368421052631579, "wer_raw": 0.07368421052631578, "ins_enhanced": 2, "ins_raw": 2 }, "telephony_g722": { "vad_f1_enhanced": 0.9656159755941424, "vad_f1_raw": 0.9718918630456168, "vad_fa_enhanced": 0.06969920953477984, "vad_fa_raw": 0.07682802264269852, "wer_enhanced": 0.05789473684210526, "wer_raw": 0.05263157894736842, "ins_enhanced": 2, "ins_raw": 1 }, "telephony_opus12k": { "vad_f1_enhanced": 0.9666243797228073, "vad_f1_raw": 0.9700698810886056, "vad_fa_enhanced": 0.05336690244052579, "vad_fa_raw": 0.06495112394829519, "wer_enhanced": 0.12105263157894737, "wer_raw": 0.06842105263157895, "ins_enhanced": 4, "ins_raw": 2 }, "telephony_loss3": { "vad_f1_enhanced": 0.9520409864804803, "vad_f1_raw": 0.9585494127803207, "vad_fa_enhanced": 0.04274530171733586, "vad_fa_raw": 0.045096973984416894, "wer_enhanced": 0.14210526315789473, "wer_raw": 0.11052631578947368, "ins_enhanced": 1, "ins_raw": 2 }, "telephony_loss10": { "vad_f1_enhanced": 0.9420918262817392, "vad_f1_raw": 0.9568651776701316, "vad_fa_enhanced": 0.03219891098687841, "vad_fa_raw": 0.052286177577941614, "wer_enhanced": 0.24736842105263157, "wer_raw": 0.1, "ins_enhanced": 4, "ins_raw": 1 } } }, "fastenhancer-t": { "label": "FastEnhancer-T", "mean_wer_enhanced": 0.22456140350877193, "mean_wer_raw": 0.1476608187134503, "wer_delta_pp": 7.690058479532164, "conditions_where_enhancement_helps_stt": 3, "conditions_scored": 18, "mean_vad_f1_enhanced": 0.9411334979174379, "mean_vad_f1_raw": 0.9503607044937831, "vad_f1_delta": -0.009227206576345126, "insertions_enhanced": 145, "insertions_raw": 101, "per_condition": { "clean": { "vad_f1_enhanced": 0.9666838181089294, "vad_f1_raw": 0.9756248909923848, "vad_fa_enhanced": 0.08724581971269307, "vad_fa_raw": 0.07023743685253026, "wer_enhanced": 0.06315789473684211, "wer_raw": 0.04736842105263158, "ins_enhanced": 2, "ins_raw": 2 }, "noise_snr-5": { "vad_f1_enhanced": 0.9433551696343677, "vad_f1_raw": 0.9284106039361377, "vad_fa_enhanced": 0.05365350409425416, "vad_fa_raw": 0.06810271419067943, "wer_enhanced": 0.24210526315789474, "wer_raw": 0.18421052631578946, "ins_enhanced": 3, "ins_raw": 4 }, "noise_snr0": { "vad_f1_enhanced": 0.9577605170422323, "vad_f1_raw": 0.9650457996816915, "vad_fa_enhanced": 0.05724875066252398, "vad_fa_raw": 0.07585414368555106, "wer_enhanced": 0.10526315789473684, "wer_raw": 0.11052631578947368, "ins_enhanced": 2, "ins_raw": 2 }, "noise_snr5": { "vad_f1_enhanced": 0.9648037870822685, "vad_f1_raw": 0.9655544660175688, "vad_fa_enhanced": 0.04392770615922425, "vad_fa_raw": 0.06337917413073132, "wer_enhanced": 0.06842105263157895, "wer_raw": 0.08947368421052632, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr10": { "vad_f1_enhanced": 0.9633761399270855, "vad_f1_raw": 0.9715691872453027, "vad_fa_enhanced": 0.05267319293479119, "vad_fa_raw": 0.0521201836224206, "wer_enhanced": 0.05263157894736842, "wer_raw": 0.06315789473684211, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr20": { "vad_f1_enhanced": 0.9666896621410606, "vad_f1_raw": 0.9722639322945248, "vad_fa_enhanced": 0.07356986477149877, "vad_fa_raw": 0.0769268153848264, "wer_enhanced": 0.042105263157894736, "wer_raw": 0.042105263157894736, "ins_enhanced": 2, "ins_raw": 2 }, "babble_snr5": { "vad_f1_enhanced": 0.8559662388718697, "vad_f1_raw": 0.8446580505770814, "vad_fa_enhanced": 0.885740892003619, "vad_fa_raw": 0.9770890756151038, "wer_enhanced": 0.3894736842105263, "wer_raw": 0.24736842105263157, "ins_enhanced": 32, "ins_raw": 10 }, "competing_snr0": { "vad_f1_enhanced": 0.899602895214499, "vad_f1_raw": 0.8978559481992789, "vad_fa_enhanced": 0.5668208226389345, "vad_fa_raw": 0.5817642991147487, "wer_enhanced": 0.868421052631579, "wer_raw": 0.8473684210526315, "ins_enhanced": 45, "ins_raw": 45 }, "competing_snr5": { "vad_f1_enhanced": 0.9001131911973104, "vad_f1_raw": 0.8993866413420953, "vad_fa_enhanced": 0.5652997988267542, "vad_fa_raw": 0.5812582262402548, "wer_enhanced": 0.4631578947368421, "wer_raw": 0.2631578947368421, "ins_enhanced": 33, "ins_raw": 19 }, "reverb": { "vad_f1_enhanced": 0.9527805509568713, "vad_f1_raw": 0.9623023734900757, "vad_fa_enhanced": 0.09492847996937877, "vad_fa_raw": 0.07937271734601117, "wer_enhanced": 0.21052631578947367, "wer_raw": 0.05789473684210526, "ins_enhanced": 3, "ins_raw": 1 }, "reverb_snr10": { "vad_f1_enhanced": 0.9468638736513982, "vad_f1_raw": 0.9565243987100389, "vad_fa_enhanced": 0.083511584208375, "vad_fa_raw": 0.07830285083543445, "wer_enhanced": 0.2894736842105263, "wer_raw": 0.1368421052631579, "ins_enhanced": 5, "ins_raw": 0 }, "telephony": { "vad_f1_enhanced": 0.9545402835712841, "vad_f1_raw": 0.9761642766503543, "vad_fa_enhanced": 0.05618961682560466, "vad_fa_raw": 0.05308086433332433, "wer_enhanced": 0.12631578947368421, "wer_raw": 0.05789473684210526, "ins_enhanced": 1, "ins_raw": 1 }, "telephony_snr10": { "vad_f1_enhanced": 0.9412494555526203, "vad_f1_raw": 0.9688306828446971, "vad_fa_enhanced": 0.023367166032038828, "vad_fa_raw": 0.05104087749954463, "wer_enhanced": 0.1736842105263158, "wer_raw": 0.10526315789473684, "ins_enhanced": 3, "ins_raw": 3 }, "telephony_alaw": { "vad_f1_enhanced": 0.9316559988106936, "vad_f1_raw": 0.9649250943221878, "vad_fa_enhanced": 0.024854343466531315, "vad_fa_raw": 0.04056293999789538, "wer_enhanced": 0.21052631578947367, "wer_raw": 0.07368421052631578, "ins_enhanced": 2, "ins_raw": 2 }, "telephony_g722": { "vad_f1_enhanced": 0.9664999455237862, "vad_f1_raw": 0.9718918630456168, "vad_fa_enhanced": 0.06248539147997112, "vad_fa_raw": 0.07682802264269852, "wer_enhanced": 0.05789473684210526, "wer_raw": 0.05263157894736842, "ins_enhanced": 1, "ins_raw": 1 }, "telephony_opus12k": { "vad_f1_enhanced": 0.9653357140525506, "vad_f1_raw": 0.9700698810886056, "vad_fa_enhanced": 0.04857607922864995, "vad_fa_raw": 0.06495112394829519, "wer_enhanced": 0.08947368421052632, "wer_raw": 0.06842105263157895, "ins_enhanced": 1, "ins_raw": 2 }, "telephony_loss3": { "vad_f1_enhanced": 0.9318630451322856, "vad_f1_raw": 0.9585494127803207, "vad_fa_enhanced": 0.02819005151475672, "vad_fa_raw": 0.045096973984416894, "wer_enhanced": 0.23684210526315788, "wer_raw": 0.11052631578947368, "ins_enhanced": 2, "ins_raw": 2 }, "telephony_loss10": { "vad_f1_enhanced": 0.9312626760427681, "vad_f1_raw": 0.9568651776701316, "vad_fa_enhanced": 0.029769343031466045, "vad_fa_raw": 0.052286177577941614, "wer_enhanced": 0.3526315789473684, "wer_raw": 0.1, "ins_enhanced": 6, "ins_raw": 1 } } }, "fastenhancer-b": { "label": "FastEnhancer-B", "mean_wer_enhanced": 0.21140350877192982, "mean_wer_raw": 0.1476608187134503, "wer_delta_pp": 6.374269005847951, "conditions_where_enhancement_helps_stt": 3, "conditions_scored": 18, "mean_vad_f1_enhanced": 0.9442287866040439, "mean_vad_f1_raw": 0.9503607044937831, "vad_f1_delta": -0.0061319178897392135, "insertions_enhanced": 159, "insertions_raw": 101, "per_condition": { "clean": { "vad_f1_enhanced": 0.9665544167255531, "vad_f1_raw": 0.9756248909923848, "vad_fa_enhanced": 0.0848434576737768, "vad_fa_raw": 0.07023743685253026, "wer_enhanced": 0.05789473684210526, "wer_raw": 0.04736842105263158, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr-5": { "vad_f1_enhanced": 0.9486474458583054, "vad_f1_raw": 0.9284106039361377, "vad_fa_enhanced": 0.04696256044091943, "vad_fa_raw": 0.06810271419067943, "wer_enhanced": 0.18421052631578946, "wer_raw": 0.18421052631578946, "ins_enhanced": 3, "ins_raw": 4 }, "noise_snr0": { "vad_f1_enhanced": 0.9631194721873815, "vad_f1_raw": 0.9650457996816915, "vad_fa_enhanced": 0.06304611532347819, "vad_fa_raw": 0.07585414368555106, "wer_enhanced": 0.07368421052631578, "wer_raw": 0.11052631578947368, "ins_enhanced": 2, "ins_raw": 2 }, "noise_snr5": { "vad_f1_enhanced": 0.9677787893171903, "vad_f1_raw": 0.9655544660175688, "vad_fa_enhanced": 0.04610694342618081, "vad_fa_raw": 0.06337917413073132, "wer_enhanced": 0.06842105263157895, "wer_raw": 0.08947368421052632, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr10": { "vad_f1_enhanced": 0.9664037535861152, "vad_f1_raw": 0.9715691872453027, "vad_fa_enhanced": 0.060846861774346575, "vad_fa_raw": 0.0521201836224206, "wer_enhanced": 0.05789473684210526, "wer_raw": 0.06315789473684211, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr20": { "vad_f1_enhanced": 0.968372309057336, "vad_f1_raw": 0.9722639322945248, "vad_fa_enhanced": 0.06914094507860882, "vad_fa_raw": 0.0769268153848264, "wer_enhanced": 0.05263157894736842, "wer_raw": 0.042105263157894736, "ins_enhanced": 1, "ins_raw": 2 }, "babble_snr5": { "vad_f1_enhanced": 0.8539578737993498, "vad_f1_raw": 0.8446580505770814, "vad_fa_enhanced": 0.8737395622775651, "vad_fa_raw": 0.9770890756151038, "wer_enhanced": 0.3684210526315789, "wer_raw": 0.24736842105263157, "ins_enhanced": 25, "ins_raw": 10 }, "competing_snr0": { "vad_f1_enhanced": 0.8984464534886838, "vad_f1_raw": 0.8978559481992789, "vad_fa_enhanced": 0.5720027589612217, "vad_fa_raw": 0.5817642991147487, "wer_enhanced": 0.9263157894736842, "wer_raw": 0.8473684210526315, "ins_enhanced": 57, "ins_raw": 45 }, "competing_snr5": { "vad_f1_enhanced": 0.9000794486343897, "vad_f1_raw": 0.8993866413420953, "vad_fa_enhanced": 0.5631822984264705, "vad_fa_raw": 0.5812582262402548, "wer_enhanced": 0.5052631578947369, "wer_raw": 0.2631578947368421, "ins_enhanced": 41, "ins_raw": 19 }, "reverb": { "vad_f1_enhanced": 0.9386444498238232, "vad_f1_raw": 0.9623023734900757, "vad_fa_enhanced": 0.0975312019452212, "vad_fa_raw": 0.07937271734601117, "wer_enhanced": 0.2578947368421053, "wer_raw": 0.05789473684210526, "ins_enhanced": 3, "ins_raw": 1 }, "reverb_snr10": { "vad_f1_enhanced": 0.9434659771081435, "vad_f1_raw": 0.9565243987100389, "vad_fa_enhanced": 0.08586169665095951, "vad_fa_raw": 0.07830285083543445, "wer_enhanced": 0.34210526315789475, "wer_raw": 0.1368421052631579, "ins_enhanced": 9, "ins_raw": 0 }, "telephony": { "vad_f1_enhanced": 0.9617762772896791, "vad_f1_raw": 0.9761642766503543, "vad_fa_enhanced": 0.06606518580019238, "vad_fa_raw": 0.05308086433332433, "wer_enhanced": 0.06842105263157895, "wer_raw": 0.05789473684210526, "ins_enhanced": 2, "ins_raw": 1 }, "telephony_snr10": { "vad_f1_enhanced": 0.9511253080366208, "vad_f1_raw": 0.9688306828446971, "vad_fa_enhanced": 0.03909554636293228, "vad_fa_raw": 0.05104087749954463, "wer_enhanced": 0.1368421052631579, "wer_raw": 0.10526315789473684, "ins_enhanced": 2, "ins_raw": 3 }, "telephony_alaw": { "vad_f1_enhanced": 0.9506472775000845, "vad_f1_raw": 0.9649250943221878, "vad_fa_enhanced": 0.04273317783109369, "vad_fa_raw": 0.04056293999789538, "wer_enhanced": 0.16842105263157894, "wer_raw": 0.07368421052631578, "ins_enhanced": 3, "ins_raw": 2 }, "telephony_g722": { "vad_f1_enhanced": 0.9668796232189438, "vad_f1_raw": 0.9718918630456168, "vad_fa_enhanced": 0.0695962794586116, "vad_fa_raw": 0.07682802264269852, "wer_enhanced": 0.05263157894736842, "wer_raw": 0.05263157894736842, "ins_enhanced": 1, "ins_raw": 1 }, "telephony_opus12k": { "vad_f1_enhanced": 0.9682160061569374, "vad_f1_raw": 0.9700698810886056, "vad_fa_enhanced": 0.049905613891517944, "vad_fa_raw": 0.06495112394829519, "wer_enhanced": 0.08421052631578947, "wer_raw": 0.06842105263157895, "ins_enhanced": 2, "ins_raw": 2 }, "telephony_loss3": { "vad_f1_enhanced": 0.9457321066331812, "vad_f1_raw": 0.9585494127803207, "vad_fa_enhanced": 0.03829010256246379, "vad_fa_raw": 0.045096973984416894, "wer_enhanced": 0.16842105263157894, "wer_raw": 0.11052631578947368, "ins_enhanced": 2, "ins_raw": 2 }, "telephony_loss10": { "vad_f1_enhanced": 0.9362711704510704, "vad_f1_raw": 0.9568651776701316, "vad_fa_enhanced": 0.04163185945491459, "vad_fa_raw": 0.052286177577941614, "wer_enhanced": 0.23157894736842105, "wer_raw": 0.1, "ins_enhanced": 3, "ins_raw": 1 } } }, "fastenhancer-s": { "label": "FastEnhancer-S", "mean_wer_enhanced": 0.2064327485380117, "mean_wer_raw": 0.1476608187134503, "wer_delta_pp": 5.8771929824561395, "conditions_where_enhancement_helps_stt": 3, "conditions_scored": 18, "mean_vad_f1_enhanced": 0.9452405508174968, "mean_vad_f1_raw": 0.9503607044937831, "vad_f1_delta": -0.005120153676286265, "insertions_enhanced": 159, "insertions_raw": 101, "per_condition": { "clean": { "vad_f1_enhanced": 0.9666243625184097, "vad_f1_raw": 0.9756248909923848, "vad_fa_enhanced": 0.0872933030935098, "vad_fa_raw": 0.07023743685253026, "wer_enhanced": 0.05789473684210526, "wer_raw": 0.04736842105263158, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr-5": { "vad_f1_enhanced": 0.9543291432211548, "vad_f1_raw": 0.9284106039361377, "vad_fa_enhanced": 0.050790979530382696, "vad_fa_raw": 0.06810271419067943, "wer_enhanced": 0.22631578947368422, "wer_raw": 0.18421052631578946, "ins_enhanced": 9, "ins_raw": 4 }, "noise_snr0": { "vad_f1_enhanced": 0.9627685527161255, "vad_f1_raw": 0.9650457996816915, "vad_fa_enhanced": 0.06696549623758156, "vad_fa_raw": 0.07585414368555106, "wer_enhanced": 0.08947368421052632, "wer_raw": 0.11052631578947368, "ins_enhanced": 3, "ins_raw": 2 }, "noise_snr5": { "vad_f1_enhanced": 0.9693534128979128, "vad_f1_raw": 0.9655544660175688, "vad_fa_enhanced": 0.04713787126123235, "vad_fa_raw": 0.06337917413073132, "wer_enhanced": 0.08421052631578947, "wer_raw": 0.08947368421052632, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr10": { "vad_f1_enhanced": 0.9657387351921551, "vad_f1_raw": 0.9715691872453027, "vad_fa_enhanced": 0.05911004697911254, "vad_fa_raw": 0.0521201836224206, "wer_enhanced": 0.07368421052631578, "wer_raw": 0.06315789473684211, "ins_enhanced": 2, "ins_raw": 2 }, "noise_snr20": { "vad_f1_enhanced": 0.9664454793934116, "vad_f1_raw": 0.9722639322945248, "vad_fa_enhanced": 0.07605903222268466, "vad_fa_raw": 0.0769268153848264, "wer_enhanced": 0.05263157894736842, "wer_raw": 0.042105263157894736, "ins_enhanced": 1, "ins_raw": 2 }, "babble_snr5": { "vad_f1_enhanced": 0.8577665455520223, "vad_f1_raw": 0.8446580505770814, "vad_fa_enhanced": 0.8505172318248697, "vad_fa_raw": 0.9770890756151038, "wer_enhanced": 0.3736842105263158, "wer_raw": 0.24736842105263157, "ins_enhanced": 13, "ins_raw": 10 }, "competing_snr0": { "vad_f1_enhanced": 0.8979486609373313, "vad_f1_raw": 0.8978559481992789, "vad_fa_enhanced": 0.5681011160616604, "vad_fa_raw": 0.5817642991147487, "wer_enhanced": 0.8631578947368421, "wer_raw": 0.8473684210526315, "ins_enhanced": 63, "ins_raw": 45 }, "competing_snr5": { "vad_f1_enhanced": 0.9010474911433507, "vad_f1_raw": 0.8993866413420953, "vad_fa_enhanced": 0.5559165919852933, "vad_fa_raw": 0.5812582262402548, "wer_enhanced": 0.5105263157894737, "wer_raw": 0.2631578947368421, "ins_enhanced": 35, "ins_raw": 19 }, "reverb": { "vad_f1_enhanced": 0.9411470356453613, "vad_f1_raw": 0.9623023734900757, "vad_fa_enhanced": 0.08884428522578725, "vad_fa_raw": 0.07937271734601117, "wer_enhanced": 0.22631578947368422, "wer_raw": 0.05789473684210526, "ins_enhanced": 9, "ins_raw": 1 }, "reverb_snr10": { "vad_f1_enhanced": 0.9488516914877356, "vad_f1_raw": 0.9565243987100389, "vad_fa_enhanced": 0.08334507746767365, "vad_fa_raw": 0.07830285083543445, "wer_enhanced": 0.2578947368421053, "wer_raw": 0.1368421052631579, "ins_enhanced": 5, "ins_raw": 0 }, "telephony": { "vad_f1_enhanced": 0.9629450687541201, "vad_f1_raw": 0.9761642766503543, "vad_fa_enhanced": 0.06770453006248747, "vad_fa_raw": 0.05308086433332433, "wer_enhanced": 0.05263157894736842, "wer_raw": 0.05789473684210526, "ins_enhanced": 1, "ins_raw": 1 }, "telephony_snr10": { "vad_f1_enhanced": 0.952131310382198, "vad_f1_raw": 0.9688306828446971, "vad_fa_enhanced": 0.04433800863550697, "vad_fa_raw": 0.05104087749954463, "wer_enhanced": 0.14210526315789473, "wer_raw": 0.10526315789473684, "ins_enhanced": 3, "ins_raw": 3 }, "telephony_alaw": { "vad_f1_enhanced": 0.9457094682914681, "vad_f1_raw": 0.9649250943221878, "vad_fa_enhanced": 0.049777509449397364, "vad_fa_raw": 0.04056293999789538, "wer_enhanced": 0.1631578947368421, "wer_raw": 0.07368421052631578, "ins_enhanced": 2, "ins_raw": 2 }, "telephony_g722": { "vad_f1_enhanced": 0.966584863318619, "vad_f1_raw": 0.9718918630456168, "vad_fa_enhanced": 0.07172363649743699, "vad_fa_raw": 0.07682802264269852, "wer_enhanced": 0.05789473684210526, "wer_raw": 0.05263157894736842, "ins_enhanced": 1, "ins_raw": 1 }, "telephony_opus12k": { "vad_f1_enhanced": 0.9678500739831397, "vad_f1_raw": 0.9700698810886056, "vad_fa_enhanced": 0.050673430949711475, "vad_fa_raw": 0.06495112394829519, "wer_enhanced": 0.06842105263157895, "wer_raw": 0.06842105263157895, "ins_enhanced": 3, "ins_raw": 2 }, "telephony_loss3": { "vad_f1_enhanced": 0.9464638375279874, "vad_f1_raw": 0.9585494127803207, "vad_fa_enhanced": 0.04881442287730438, "vad_fa_raw": 0.045096973984416894, "wer_enhanced": 0.16842105263157894, "wer_raw": 0.11052631578947368, "ins_enhanced": 3, "ins_raw": 2 }, "telephony_loss10": { "vad_f1_enhanced": 0.9406241817524394, "vad_f1_raw": 0.9568651776701316, "vad_fa_enhanced": 0.045137506932255246, "vad_fa_raw": 0.052286177577941614, "wer_enhanced": 0.24736842105263157, "wer_raw": 0.1, "ins_enhanced": 4, "ins_raw": 1 } } }, "fastenhancer-m": { "label": "FastEnhancer-M", "mean_wer_enhanced": 0.2239766081871345, "mean_wer_raw": 0.1476608187134503, "wer_delta_pp": 7.6315789473684195, "conditions_where_enhancement_helps_stt": 5, "conditions_scored": 18, "mean_vad_f1_enhanced": 0.9405634506008924, "mean_vad_f1_raw": 0.9503607044937831, "vad_f1_delta": -0.009797253892890656, "insertions_enhanced": 217, "insertions_raw": 101, "per_condition": { "clean": { "vad_f1_enhanced": 0.9646538637898979, "vad_f1_raw": 0.9756248909923848, "vad_fa_enhanced": 0.0910952605066407, "vad_fa_raw": 0.07023743685253026, "wer_enhanced": 0.042105263157894736, "wer_raw": 0.04736842105263158, "ins_enhanced": 2, "ins_raw": 2 }, "noise_snr-5": { "vad_f1_enhanced": 0.9472517306837446, "vad_f1_raw": 0.9284106039361377, "vad_fa_enhanced": 0.04154623326386513, "vad_fa_raw": 0.06810271419067943, "wer_enhanced": 0.15263157894736842, "wer_raw": 0.18421052631578946, "ins_enhanced": 6, "ins_raw": 4 }, "noise_snr0": { "vad_f1_enhanced": 0.9628264939790558, "vad_f1_raw": 0.9650457996816915, "vad_fa_enhanced": 0.07001115798376083, "vad_fa_raw": 0.07585414368555106, "wer_enhanced": 0.07894736842105263, "wer_raw": 0.11052631578947368, "ins_enhanced": 2, "ins_raw": 2 }, "noise_snr5": { "vad_f1_enhanced": 0.9660458156554125, "vad_f1_raw": 0.9655544660175688, "vad_fa_enhanced": 0.04938579582784024, "vad_fa_raw": 0.06337917413073132, "wer_enhanced": 0.09473684210526316, "wer_raw": 0.08947368421052632, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr10": { "vad_f1_enhanced": 0.9665975500045096, "vad_f1_raw": 0.9715691872453027, "vad_fa_enhanced": 0.058201537637794364, "vad_fa_raw": 0.0521201836224206, "wer_enhanced": 0.05789473684210526, "wer_raw": 0.06315789473684211, "ins_enhanced": 2, "ins_raw": 2 }, "noise_snr20": { "vad_f1_enhanced": 0.9664991502166893, "vad_f1_raw": 0.9722639322945248, "vad_fa_enhanced": 0.07626595923666148, "vad_fa_raw": 0.0769268153848264, "wer_enhanced": 0.031578947368421054, "wer_raw": 0.042105263157894736, "ins_enhanced": 1, "ins_raw": 2 }, "babble_snr5": { "vad_f1_enhanced": 0.8587428616865319, "vad_f1_raw": 0.8446580505770814, "vad_fa_enhanced": 0.7461100741696446, "vad_fa_raw": 0.9770890756151038, "wer_enhanced": 0.3, "wer_raw": 0.24736842105263157, "ins_enhanced": 9, "ins_raw": 10 }, "competing_snr0": { "vad_f1_enhanced": 0.8949618104404795, "vad_f1_raw": 0.8978559481992789, "vad_fa_enhanced": 0.5473616995996899, "vad_fa_raw": 0.5817642991147487, "wer_enhanced": 0.9631578947368421, "wer_raw": 0.8473684210526315, "ins_enhanced": 71, "ins_raw": 45 }, "competing_snr5": { "vad_f1_enhanced": 0.9045856994266565, "vad_f1_raw": 0.8993866413420953, "vad_fa_enhanced": 0.4828302424810251, "vad_fa_raw": 0.5812582262402548, "wer_enhanced": 0.46842105263157896, "wer_raw": 0.2631578947368421, "ins_enhanced": 30, "ins_raw": 19 }, "reverb": { "vad_f1_enhanced": 0.9155334128372689, "vad_f1_raw": 0.9623023734900757, "vad_fa_enhanced": 0.09937079113985055, "vad_fa_raw": 0.07937271734601117, "wer_enhanced": 0.2736842105263158, "wer_raw": 0.05789473684210526, "ins_enhanced": 13, "ins_raw": 1 }, "reverb_snr10": { "vad_f1_enhanced": 0.9511791968018667, "vad_f1_raw": 0.9565243987100389, "vad_fa_enhanced": 0.08211856272974793, "vad_fa_raw": 0.07830285083543445, "wer_enhanced": 0.22105263157894736, "wer_raw": 0.1368421052631579, "ins_enhanced": 4, "ins_raw": 0 }, "telephony": { "vad_f1_enhanced": 0.956563721091612, "vad_f1_raw": 0.9761642766503543, "vad_fa_enhanced": 0.05917274923662395, "vad_fa_raw": 0.05308086433332433, "wer_enhanced": 0.05789473684210526, "wer_raw": 0.05789473684210526, "ins_enhanced": 2, "ins_raw": 1 }, "telephony_snr10": { "vad_f1_enhanced": 0.9459190574496201, "vad_f1_raw": 0.9688306828446971, "vad_fa_enhanced": 0.04643360249724299, "vad_fa_raw": 0.05104087749954463, "wer_enhanced": 0.12105263157894737, "wer_raw": 0.10526315789473684, "ins_enhanced": 2, "ins_raw": 3 }, "telephony_alaw": { "vad_f1_enhanced": 0.9362633185965639, "vad_f1_raw": 0.9649250943221878, "vad_fa_enhanced": 0.041002393693906944, "vad_fa_raw": 0.04056293999789538, "wer_enhanced": 0.48947368421052634, "wer_raw": 0.07368421052631578, "ins_enhanced": 61, "ins_raw": 2 }, "telephony_g722": { "vad_f1_enhanced": 0.9652604647870209, "vad_f1_raw": 0.9718918630456168, "vad_fa_enhanced": 0.06459809121692303, "vad_fa_raw": 0.07682802264269852, "wer_enhanced": 0.06842105263157895, "wer_raw": 0.05263157894736842, "ins_enhanced": 1, "ins_raw": 1 }, "telephony_opus12k": { "vad_f1_enhanced": 0.966958273782544, "vad_f1_raw": 0.9700698810886056, "vad_fa_enhanced": 0.04406370625512908, "vad_fa_raw": 0.06495112394829519, "wer_enhanced": 0.08421052631578947, "wer_raw": 0.06842105263157895, "ins_enhanced": 2, "ins_raw": 2 }, "telephony_loss3": { "vad_f1_enhanced": 0.9350656543242144, "vad_f1_raw": 0.9585494127803207, "vad_fa_enhanced": 0.04241791165623176, "vad_fa_raw": 0.045096973984416894, "wer_enhanced": 0.2, "wer_raw": 0.11052631578947368, "ins_enhanced": 2, "ins_raw": 2 }, "telephony_loss10": { "vad_f1_enhanced": 0.9252340352623738, "vad_f1_raw": 0.9568651776701316, "vad_fa_enhanced": 0.04228854272952001, "vad_fa_raw": 0.052286177577941614, "wer_enhanced": 0.3263157894736842, "wer_raw": 0.1, "ins_enhanced": 6, "ins_raw": 1 } } }, "fastenhancer-l": { "label": "FastEnhancer-L", "mean_wer_enhanced": 0.2043859649122807, "mean_wer_raw": 0.1476608187134503, "wer_delta_pp": 5.6725146198830405, "conditions_where_enhancement_helps_stt": 6, "conditions_scored": 18, "mean_vad_f1_enhanced": 0.9417271235541121, "mean_vad_f1_raw": 0.9503607044937831, "vad_f1_delta": -0.008633580939670926, "insertions_enhanced": 159, "insertions_raw": 101, "per_condition": { "clean": { "vad_f1_enhanced": 0.9629243179178815, "vad_f1_raw": 0.9756248909923848, "vad_fa_enhanced": 0.09624337669876035, "vad_fa_raw": 0.07023743685253026, "wer_enhanced": 0.05263157894736842, "wer_raw": 0.04736842105263158, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr-5": { "vad_f1_enhanced": 0.9579620400813467, "vad_f1_raw": 0.9284106039361377, "vad_fa_enhanced": 0.04727318113617144, "vad_fa_raw": 0.06810271419067943, "wer_enhanced": 0.1368421052631579, "wer_raw": 0.18421052631578946, "ins_enhanced": 2, "ins_raw": 4 }, "noise_snr0": { "vad_f1_enhanced": 0.9627543916240529, "vad_f1_raw": 0.9650457996816915, "vad_fa_enhanced": 0.07848709604461479, "vad_fa_raw": 0.07585414368555106, "wer_enhanced": 0.07894736842105263, "wer_raw": 0.11052631578947368, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr5": { "vad_f1_enhanced": 0.9649830801016552, "vad_f1_raw": 0.9655544660175688, "vad_fa_enhanced": 0.06291504690619273, "vad_fa_raw": 0.06337917413073132, "wer_enhanced": 0.06842105263157895, "wer_raw": 0.08947368421052632, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr10": { "vad_f1_enhanced": 0.9630539185795298, "vad_f1_raw": 0.9715691872453027, "vad_fa_enhanced": 0.07026693016822186, "vad_fa_raw": 0.0521201836224206, "wer_enhanced": 0.05263157894736842, "wer_raw": 0.06315789473684211, "ins_enhanced": 3, "ins_raw": 2 }, "noise_snr20": { "vad_f1_enhanced": 0.9658739810576314, "vad_f1_raw": 0.9722639322945248, "vad_fa_enhanced": 0.08561768613224802, "vad_fa_raw": 0.0769268153848264, "wer_enhanced": 0.03684210526315789, "wer_raw": 0.042105263157894736, "ins_enhanced": 1, "ins_raw": 2 }, "babble_snr5": { "vad_f1_enhanced": 0.8511730293294211, "vad_f1_raw": 0.8446580505770814, "vad_fa_enhanced": 0.6824619387440339, "vad_fa_raw": 0.9770890756151038, "wer_enhanced": 0.3684210526315789, "wer_raw": 0.24736842105263157, "ins_enhanced": 12, "ins_raw": 10 }, "competing_snr0": { "vad_f1_enhanced": 0.8826170709857722, "vad_f1_raw": 0.8978559481992789, "vad_fa_enhanced": 0.5451808126594629, "vad_fa_raw": 0.5817642991147487, "wer_enhanced": 0.8421052631578947, "wer_raw": 0.8473684210526315, "ins_enhanced": 31, "ins_raw": 45 }, "competing_snr5": { "vad_f1_enhanced": 0.9059799600495037, "vad_f1_raw": 0.8993866413420953, "vad_fa_enhanced": 0.3904156960002697, "vad_fa_raw": 0.5812582262402548, "wer_enhanced": 0.7631578947368421, "wer_raw": 0.2631578947368421, "ins_enhanced": 85, "ins_raw": 19 }, "reverb": { "vad_f1_enhanced": 0.9127343755383647, "vad_f1_raw": 0.9623023734900757, "vad_fa_enhanced": 0.09703355738966754, "vad_fa_raw": 0.07937271734601117, "wer_enhanced": 0.1736842105263158, "wer_raw": 0.05789473684210526, "ins_enhanced": 3, "ins_raw": 1 }, "reverb_snr10": { "vad_f1_enhanced": 0.9474239824936153, "vad_f1_raw": 0.9565243987100389, "vad_fa_enhanced": 0.08594514993159044, "vad_fa_raw": 0.07830285083543445, "wer_enhanced": 0.22631578947368422, "wer_raw": 0.1368421052631579, "ins_enhanced": 1, "ins_raw": 0 }, "telephony": { "vad_f1_enhanced": 0.959613620904966, "vad_f1_raw": 0.9761642766503543, "vad_fa_enhanced": 0.07541109124065136, "vad_fa_raw": 0.05308086433332433, "wer_enhanced": 0.07368421052631578, "wer_raw": 0.05789473684210526, "ins_enhanced": 1, "ins_raw": 1 }, "telephony_snr10": { "vad_f1_enhanced": 0.9532800954240364, "vad_f1_raw": 0.9688306828446971, "vad_fa_enhanced": 0.05091177459836592, "vad_fa_raw": 0.05104087749954463, "wer_enhanced": 0.1368421052631579, "wer_raw": 0.10526315789473684, "ins_enhanced": 3, "ins_raw": 3 }, "telephony_alaw": { "vad_f1_enhanced": 0.9500018687942842, "vad_f1_raw": 0.9649250943221878, "vad_fa_enhanced": 0.05337487974278775, "vad_fa_raw": 0.04056293999789538, "wer_enhanced": 0.12105263157894737, "wer_raw": 0.07368421052631578, "ins_enhanced": 1, "ins_raw": 2 }, "telephony_g722": { "vad_f1_enhanced": 0.9651000357824729, "vad_f1_raw": 0.9718918630456168, "vad_fa_enhanced": 0.07506196752001078, "vad_fa_raw": 0.07682802264269852, "wer_enhanced": 0.05263157894736842, "wer_raw": 0.05263157894736842, "ins_enhanced": 1, "ins_raw": 1 }, "telephony_opus12k": { "vad_f1_enhanced": 0.9659973396042536, "vad_f1_raw": 0.9700698810886056, "vad_fa_enhanced": 0.05232077297435804, "vad_fa_raw": 0.06495112394829519, "wer_enhanced": 0.09473684210526316, "wer_raw": 0.06842105263157895, "ins_enhanced": 4, "ins_raw": 2 }, "telephony_loss3": { "vad_f1_enhanced": 0.9377517967544996, "vad_f1_raw": 0.9585494127803207, "vad_fa_enhanced": 0.05672220228888042, "vad_fa_raw": 0.045096973984416894, "wer_enhanced": 0.2, "wer_raw": 0.11052631578947368, "ins_enhanced": 5, "ins_raw": 2 }, "telephony_loss10": { "vad_f1_enhanced": 0.9418633189507318, "vad_f1_raw": 0.9568651776701316, "vad_fa_enhanced": 0.0484339305035929, "vad_fa_raw": 0.052286177577941614, "wer_enhanced": 0.2, "wer_raw": 0.1, "ins_enhanced": 3, "ins_raw": 1 } } }, "dfn3": { "label": "DeepFilterNet3", "mean_wer_enhanced": 0.3517543859649123, "mean_wer_raw": 0.1476608187134503, "wer_delta_pp": 20.409356725146203, "conditions_where_enhancement_helps_stt": 0, "conditions_scored": 18, "mean_vad_f1_enhanced": 0.9013407581938893, "mean_vad_f1_raw": 0.9503607044937831, "vad_f1_delta": -0.04901994629989381, "insertions_enhanced": 380, "insertions_raw": 101, "per_condition": { "clean": { "vad_f1_enhanced": 0.9344634830661823, "vad_f1_raw": 0.9756248909923848, "vad_fa_enhanced": 0.1692684998627181, "vad_fa_raw": 0.07023743685253026, "wer_enhanced": 0.04736842105263158, "wer_raw": 0.04736842105263158, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr-5": { "vad_f1_enhanced": 0.917273489908753, "vad_f1_raw": 0.9284106039361377, "vad_fa_enhanced": 0.10080445428748104, "vad_fa_raw": 0.06810271419067943, "wer_enhanced": 0.32105263157894737, "wer_raw": 0.18421052631578946, "ins_enhanced": 10, "ins_raw": 4 }, "noise_snr0": { "vad_f1_enhanced": 0.934640575037475, "vad_f1_raw": 0.9650457996816915, "vad_fa_enhanced": 0.12288608193714445, "vad_fa_raw": 0.07585414368555106, "wer_enhanced": 0.1368421052631579, "wer_raw": 0.11052631578947368, "ins_enhanced": 4, "ins_raw": 2 }, "noise_snr5": { "vad_f1_enhanced": 0.931097136592846, "vad_f1_raw": 0.9655544660175688, "vad_fa_enhanced": 0.13011499502599427, "vad_fa_raw": 0.06337917413073132, "wer_enhanced": 0.14210526315789473, "wer_raw": 0.08947368421052632, "ins_enhanced": 4, "ins_raw": 2 }, "noise_snr10": { "vad_f1_enhanced": 0.9294361150908566, "vad_f1_raw": 0.9715691872453027, "vad_fa_enhanced": 0.13400594715946113, "vad_fa_raw": 0.0521201836224206, "wer_enhanced": 0.1, "wer_raw": 0.06315789473684211, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr20": { "vad_f1_enhanced": 0.9320230637003544, "vad_f1_raw": 0.9722639322945248, "vad_fa_enhanced": 0.16500763107842736, "vad_fa_raw": 0.0769268153848264, "wer_enhanced": 0.04736842105263158, "wer_raw": 0.042105263157894736, "ins_enhanced": 1, "ins_raw": 2 }, "babble_snr5": { "vad_f1_enhanced": 0.8501436209011339, "vad_f1_raw": 0.8446580505770814, "vad_fa_enhanced": 0.9235918334215316, "vad_fa_raw": 0.9770890756151038, "wer_enhanced": 0.35789473684210527, "wer_raw": 0.24736842105263157, "ins_enhanced": 8, "ins_raw": 10 }, "competing_snr0": { "vad_f1_enhanced": 0.8930722311313269, "vad_f1_raw": 0.8978559481992789, "vad_fa_enhanced": 0.5851163493910241, "vad_fa_raw": 0.5817642991147487, "wer_enhanced": 1.9947368421052631, "wer_raw": 0.8473684210526315, "ins_enhanced": 271, "ins_raw": 45 }, "competing_snr5": { "vad_f1_enhanced": 0.8944450385946545, "vad_f1_raw": 0.8993866413420953, "vad_fa_enhanced": 0.5745996540065439, "vad_fa_raw": 0.5812582262402548, "wer_enhanced": 0.5, "wer_raw": 0.2631578947368421, "ins_enhanced": 38, "ins_raw": 19 }, "reverb": { "vad_f1_enhanced": 0.8101647187084605, "vad_f1_raw": 0.9623023734900757, "vad_fa_enhanced": 0.11145347037808477, "vad_fa_raw": 0.07937271734601117, "wer_enhanced": 0.6, "wer_raw": 0.05789473684210526, "ins_enhanced": 7, "ins_raw": 1 }, "reverb_snr10": { "vad_f1_enhanced": 0.7793212008443005, "vad_f1_raw": 0.9565243987100389, "vad_fa_enhanced": 0.09484392770931992, "vad_fa_raw": 0.07830285083543445, "wer_enhanced": 0.8210526315789474, "wer_raw": 0.1368421052631579, "ins_enhanced": 16, "ins_raw": 0 }, "telephony": { "vad_f1_enhanced": 0.9174853543173865, "vad_f1_raw": 0.9761642766503543, "vad_fa_enhanced": 0.14605093386603654, "vad_fa_raw": 0.05308086433332433, "wer_enhanced": 0.15263157894736842, "wer_raw": 0.05789473684210526, "ins_enhanced": 0, "ins_raw": 1 }, "telephony_snr10": { "vad_f1_enhanced": 0.9176491008034311, "vad_f1_raw": 0.9688306828446971, "vad_fa_enhanced": 0.11788862606853864, "vad_fa_raw": 0.05104087749954463, "wer_enhanced": 0.2, "wer_raw": 0.10526315789473684, "ins_enhanced": 4, "ins_raw": 3 }, "telephony_alaw": { "vad_f1_enhanced": 0.9129433069684186, "vad_f1_raw": 0.9649250943221878, "vad_fa_enhanced": 0.11202009316880508, "vad_fa_raw": 0.04056293999789538, "wer_enhanced": 0.18421052631578946, "wer_raw": 0.07368421052631578, "ins_enhanced": 2, "ins_raw": 2 }, "telephony_g722": { "vad_f1_enhanced": 0.9322514680591638, "vad_f1_raw": 0.9718918630456168, "vad_fa_enhanced": 0.16357829079187505, "vad_fa_raw": 0.07682802264269852, "wer_enhanced": 0.06315789473684211, "wer_raw": 0.05263157894736842, "ins_enhanced": 3, "ins_raw": 1 }, "telephony_opus12k": { "vad_f1_enhanced": 0.932103062108672, "vad_f1_raw": 0.9700698810886056, "vad_fa_enhanced": 0.13752384261369555, "vad_fa_raw": 0.06495112394829519, "wer_enhanced": 0.11052631578947368, "wer_raw": 0.06842105263157895, "ins_enhanced": 3, "ins_raw": 2 }, "telephony_loss3": { "vad_f1_enhanced": 0.9080583202503478, "vad_f1_raw": 0.9585494127803207, "vad_fa_enhanced": 0.10358555776836417, "vad_fa_raw": 0.045096973984416894, "wer_enhanced": 0.24736842105263157, "wer_raw": 0.11052631578947368, "ins_enhanced": 4, "ins_raw": 2 }, "telephony_loss10": { "vad_f1_enhanced": 0.8975623614062421, "vad_f1_raw": 0.9568651776701316, "vad_fa_enhanced": 0.10112391934181346, "vad_fa_raw": 0.052286177577941614, "wer_enhanced": 0.30526315789473685, "wer_raw": 0.1, "ins_enhanced": 3, "ins_raw": 1 } } }, "dfn3-hq": { "label": "DeepFilterNet3 (96-frame warm-up)", "mean_wer_enhanced": 0.316374269005848, "mean_wer_raw": 0.1476608187134503, "wer_delta_pp": 16.871345029239766, "conditions_where_enhancement_helps_stt": 0, "conditions_scored": 18, "mean_vad_f1_enhanced": 0.9142399530378845, "mean_vad_f1_raw": 0.9503607044937831, "vad_f1_delta": -0.03612075145589855, "insertions_enhanced": 368, "insertions_raw": 101, "per_condition": { "clean": { "vad_f1_enhanced": 0.9360986893875168, "vad_f1_raw": 0.9756248909923848, "vad_fa_enhanced": 0.1692684998627181, "vad_fa_raw": 0.07023743685253026, "wer_enhanced": 0.05263157894736842, "wer_raw": 0.04736842105263158, "ins_enhanced": 2, "ins_raw": 2 }, "noise_snr-5": { "vad_f1_enhanced": 0.9261450104900151, "vad_f1_raw": 0.9284106039361377, "vad_fa_enhanced": 0.1298462579001574, "vad_fa_raw": 0.06810271419067943, "wer_enhanced": 0.34210526315789475, "wer_raw": 0.18421052631578946, "ins_enhanced": 12, "ins_raw": 4 }, "noise_snr0": { "vad_f1_enhanced": 0.9343800093823976, "vad_f1_raw": 0.9650457996816915, "vad_fa_enhanced": 0.13884014483050114, "vad_fa_raw": 0.07585414368555106, "wer_enhanced": 0.14736842105263157, "wer_raw": 0.11052631578947368, "ins_enhanced": 5, "ins_raw": 2 }, "noise_snr5": { "vad_f1_enhanced": 0.934412249993897, "vad_f1_raw": 0.9655544660175688, "vad_fa_enhanced": 0.13414231682950556, "vad_fa_raw": 0.06337917413073132, "wer_enhanced": 0.14210526315789473, "wer_raw": 0.08947368421052632, "ins_enhanced": 5, "ins_raw": 2 }, "noise_snr10": { "vad_f1_enhanced": 0.9365538934717816, "vad_f1_raw": 0.9715691872453027, "vad_fa_enhanced": 0.12638295564827942, "vad_fa_raw": 0.0521201836224206, "wer_enhanced": 0.06842105263157895, "wer_raw": 0.06315789473684211, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr20": { "vad_f1_enhanced": 0.9355316846968706, "vad_f1_raw": 0.9722639322945248, "vad_fa_enhanced": 0.16237605213105896, "vad_fa_raw": 0.0769268153848264, "wer_enhanced": 0.042105263157894736, "wer_raw": 0.042105263157894736, "ins_enhanced": 1, "ins_raw": 2 }, "babble_snr5": { "vad_f1_enhanced": 0.8500041792905947, "vad_f1_raw": 0.8446580505770814, "vad_fa_enhanced": 0.925784551273795, "vad_fa_raw": 0.9770890756151038, "wer_enhanced": 0.4842105263157895, "wer_raw": 0.24736842105263157, "ins_enhanced": 34, "ins_raw": 10 }, "competing_snr0": { "vad_f1_enhanced": 0.8919447790287733, "vad_f1_raw": 0.8978559481992789, "vad_fa_enhanced": 0.5804903911568184, "vad_fa_raw": 0.5817642991147487, "wer_enhanced": 1.431578947368421, "wer_raw": 0.8473684210526315, "ins_enhanced": 175, "ins_raw": 45 }, "competing_snr5": { "vad_f1_enhanced": 0.8938889229328989, "vad_f1_raw": 0.8993866413420953, "vad_fa_enhanced": 0.5657146779539984, "vad_fa_raw": 0.5812582262402548, "wer_enhanced": 0.7526315789473684, "wer_raw": 0.2631578947368421, "ins_enhanced": 91, "ins_raw": 19 }, "reverb": { "vad_f1_enhanced": 0.8825847575856812, "vad_f1_raw": 0.9623023734900757, "vad_fa_enhanced": 0.14675663130372454, "vad_fa_raw": 0.07937271734601117, "wer_enhanced": 0.43157894736842106, "wer_raw": 0.05789473684210526, "ins_enhanced": 6, "ins_raw": 1 }, "reverb_snr10": { "vad_f1_enhanced": 0.9006316937393082, "vad_f1_raw": 0.9565243987100389, "vad_fa_enhanced": 0.12369974154733454, "vad_fa_raw": 0.07830285083543445, "wer_enhanced": 0.6, "wer_raw": 0.1368421052631579, "ins_enhanced": 19, "ins_raw": 0 }, "telephony": { "vad_f1_enhanced": 0.91882113721847, "vad_f1_raw": 0.9761642766503543, "vad_fa_enhanced": 0.1457169680366967, "vad_fa_raw": 0.05308086433332433, "wer_enhanced": 0.15789473684210525, "wer_raw": 0.05789473684210526, "ins_enhanced": 0, "ins_raw": 1 }, "telephony_snr10": { "vad_f1_enhanced": 0.9184329136320004, "vad_f1_raw": 0.9688306828446971, "vad_fa_enhanced": 0.11763246572465899, "vad_fa_raw": 0.05104087749954463, "wer_enhanced": 0.1631578947368421, "wer_raw": 0.10526315789473684, "ins_enhanced": 2, "ins_raw": 3 }, "telephony_alaw": { "vad_f1_enhanced": 0.9142942556059177, "vad_f1_raw": 0.9649250943221878, "vad_fa_enhanced": 0.11332284007301852, "vad_fa_raw": 0.04056293999789538, "wer_enhanced": 0.1736842105263158, "wer_raw": 0.07368421052631578, "ins_enhanced": 2, "ins_raw": 2 }, "telephony_g722": { "vad_f1_enhanced": 0.9347436832223854, "vad_f1_raw": 0.9718918630456168, "vad_fa_enhanced": 0.17487761621319514, "vad_fa_raw": 0.07682802264269852, "wer_enhanced": 0.06842105263157895, "wer_raw": 0.05263157894736842, "ins_enhanced": 3, "ins_raw": 1 }, "telephony_opus12k": { "vad_f1_enhanced": 0.9353605962179072, "vad_f1_raw": 0.9700698810886056, "vad_fa_enhanced": 0.13879215661680241, "vad_fa_raw": 0.06495112394829519, "wer_enhanced": 0.1, "wer_raw": 0.06842105263157895, "ins_enhanced": 3, "ins_raw": 2 }, "telephony_loss3": { "vad_f1_enhanced": 0.9097221457532649, "vad_f1_raw": 0.9585494127803207, "vad_fa_enhanced": 0.10226976829467996, "vad_fa_raw": 0.045096973984416894, "wer_enhanced": 0.2, "wer_raw": 0.11052631578947368, "ins_enhanced": 2, "ins_raw": 2 }, "telephony_loss10": { "vad_f1_enhanced": 0.9027685530322402, "vad_f1_raw": 0.9568651776701316, "vad_fa_enhanced": 0.09970812620961592, "vad_fa_raw": 0.052286177577941614, "wer_enhanced": 0.3368421052631579, "wer_raw": 0.1, "ins_enhanced": 5, "ins_raw": 1 } } }, "aic-quail-vf": { "label": "ai-coustics Quail VF 2.2 L", "mean_wer_enhanced": 0.18099415204678362, "mean_wer_raw": 0.1476608187134503, "wer_delta_pp": 3.3333333333333326, "conditions_where_enhancement_helps_stt": 5, "conditions_scored": 18, "mean_vad_f1_enhanced": 0.9451311774451506, "mean_vad_f1_raw": 0.9503607044937831, "vad_f1_delta": -0.005229527048632443, "insertions_enhanced": 161, "insertions_raw": 101, "per_condition": { "clean": { "vad_f1_enhanced": 0.9642642935054493, "vad_f1_raw": 0.9756248909923848, "vad_fa_enhanced": 0.09843234304981663, "vad_fa_raw": 0.07023743685253026, "wer_enhanced": 0.05789473684210526, "wer_raw": 0.04736842105263158, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr-5": { "vad_f1_enhanced": 0.9388323659781698, "vad_f1_raw": 0.9284106039361377, "vad_fa_enhanced": 0.08067543587989572, "vad_fa_raw": 0.06810271419067943, "wer_enhanced": 0.24736842105263157, "wer_raw": 0.18421052631578946, "ins_enhanced": 7, "ins_raw": 4 }, "noise_snr0": { "vad_f1_enhanced": 0.9541062681659322, "vad_f1_raw": 0.9650457996816915, "vad_fa_enhanced": 0.08695552061901465, "vad_fa_raw": 0.07585414368555106, "wer_enhanced": 0.1, "wer_raw": 0.11052631578947368, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr5": { "vad_f1_enhanced": 0.9585920604028434, "vad_f1_raw": 0.9655544660175688, "vad_fa_enhanced": 0.08098640311755946, "vad_fa_raw": 0.06337917413073132, "wer_enhanced": 0.09473684210526316, "wer_raw": 0.08947368421052632, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr10": { "vad_f1_enhanced": 0.9575286185992142, "vad_f1_raw": 0.9715691872453027, "vad_fa_enhanced": 0.08155190805646076, "vad_fa_raw": 0.0521201836224206, "wer_enhanced": 0.05789473684210526, "wer_raw": 0.06315789473684211, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr20": { "vad_f1_enhanced": 0.9608867606416368, "vad_f1_raw": 0.9722639322945248, "vad_fa_enhanced": 0.09897604266658552, "vad_fa_raw": 0.0769268153848264, "wer_enhanced": 0.05789473684210526, "wer_raw": 0.042105263157894736, "ins_enhanced": 1, "ins_raw": 2 }, "babble_snr5": { "vad_f1_enhanced": 0.8690845654840278, "vad_f1_raw": 0.8446580505770814, "vad_fa_enhanced": 0.7463232636699273, "vad_fa_raw": 0.9770890756151038, "wer_enhanced": 0.20526315789473684, "wer_raw": 0.24736842105263157, "ins_enhanced": 7, "ins_raw": 10 }, "competing_snr0": { "vad_f1_enhanced": 0.8955673227016481, "vad_f1_raw": 0.8978559481992789, "vad_fa_enhanced": 0.4431741328075541, "vad_fa_raw": 0.5817642991147487, "wer_enhanced": 1.168421052631579, "wer_raw": 0.8473684210526315, "ins_enhanced": 109, "ins_raw": 45 }, "competing_snr5": { "vad_f1_enhanced": 0.9204030792116544, "vad_f1_raw": 0.8993866413420953, "vad_fa_enhanced": 0.37171796917161315, "vad_fa_raw": 0.5812582262402548, "wer_enhanced": 0.20526315789473684, "wer_raw": 0.2631578947368421, "ins_enhanced": 10, "ins_raw": 19 }, "reverb": { "vad_f1_enhanced": 0.9536769965561019, "vad_f1_raw": 0.9623023734900757, "vad_fa_enhanced": 0.08770041193161836, "vad_fa_raw": 0.07937271734601117, "wer_enhanced": 0.15263157894736842, "wer_raw": 0.05789473684210526, "ins_enhanced": 5, "ins_raw": 1 }, "reverb_snr10": { "vad_f1_enhanced": 0.9425579524995179, "vad_f1_raw": 0.9565243987100389, "vad_fa_enhanced": 0.07576065480987951, "vad_fa_raw": 0.07830285083543445, "wer_enhanced": 0.23157894736842105, "wer_raw": 0.1368421052631579, "ins_enhanced": 6, "ins_raw": 0 }, "telephony": { "vad_f1_enhanced": 0.9626392421208759, "vad_f1_raw": 0.9761642766503543, "vad_fa_enhanced": 0.08560959061088316, "vad_fa_raw": 0.05308086433332433, "wer_enhanced": 0.08421052631578947, "wer_raw": 0.05789473684210526, "ins_enhanced": 3, "ins_raw": 1 }, "telephony_snr10": { "vad_f1_enhanced": 0.9571519498721031, "vad_f1_raw": 0.9688306828446971, "vad_fa_enhanced": 0.057651468730805844, "vad_fa_raw": 0.05104087749954463, "wer_enhanced": 0.09473684210526316, "wer_raw": 0.10526315789473684, "ins_enhanced": 3, "ins_raw": 3 }, "telephony_alaw": { "vad_f1_enhanced": 0.9584714101325235, "vad_f1_raw": 0.9649250943221878, "vad_fa_enhanced": 0.0535979419445999, "vad_fa_raw": 0.04056293999789538, "wer_enhanced": 0.10526315789473684, "wer_raw": 0.07368421052631578, "ins_enhanced": 3, "ins_raw": 2 }, "telephony_g722": { "vad_f1_enhanced": 0.961230532167449, "vad_f1_raw": 0.9718918630456168, "vad_fa_enhanced": 0.08843995543495607, "vad_fa_raw": 0.07682802264269852, "wer_enhanced": 0.06842105263157895, "wer_raw": 0.05263157894736842, "ins_enhanced": 1, "ins_raw": 1 }, "telephony_opus12k": { "vad_f1_enhanced": 0.9605288807028897, "vad_f1_raw": 0.9700698810886056, "vad_fa_enhanced": 0.08400093121227807, "vad_fa_raw": 0.06495112394829519, "wer_enhanced": 0.09473684210526316, "wer_raw": 0.06842105263157895, "ins_enhanced": 1, "ins_raw": 2 }, "telephony_loss3": { "vad_f1_enhanced": 0.9489408846529939, "vad_f1_raw": 0.9585494127803207, "vad_fa_enhanced": 0.06050294787293735, "vad_fa_raw": 0.045096973984416894, "wer_enhanced": 0.11052631578947368, "wer_raw": 0.11052631578947368, "ins_enhanced": 0, "ins_raw": 2 }, "telephony_loss10": { "vad_f1_enhanced": 0.9478980106176815, "vad_f1_raw": 0.9568651776701316, "vad_fa_enhanced": 0.07004849684243278, "vad_fa_raw": 0.052286177577941614, "wer_enhanced": 0.12105263157894737, "wer_raw": 0.1, "ins_enhanced": 1, "ins_raw": 1 } } }, "aic-quail": { "label": "ai-coustics Quail L", "mean_wer_enhanced": 0.15555555555555556, "mean_wer_raw": 0.1476608187134503, "wer_delta_pp": 0.789473684210526, "conditions_where_enhancement_helps_stt": 3, "conditions_scored": 18, "mean_vad_f1_enhanced": 0.9449486118098753, "mean_vad_f1_raw": 0.9503607044937831, "vad_f1_delta": -0.005412092683907743, "insertions_enhanced": 126, "insertions_raw": 101, "per_condition": { "clean": { "vad_f1_enhanced": 0.9639893357243585, "vad_f1_raw": 0.9756248909923848, "vad_fa_enhanced": 0.098763985449553, "vad_fa_raw": 0.07023743685253026, "wer_enhanced": 0.06315789473684211, "wer_raw": 0.04736842105263158, "ins_enhanced": 2, "ins_raw": 2 }, "noise_snr-5": { "vad_f1_enhanced": 0.9477738242853431, "vad_f1_raw": 0.9284106039361377, "vad_fa_enhanced": 0.0845464844086952, "vad_fa_raw": 0.06810271419067943, "wer_enhanced": 0.14210526315789473, "wer_raw": 0.18421052631578946, "ins_enhanced": 4, "ins_raw": 4 }, "noise_snr0": { "vad_f1_enhanced": 0.9604395874001461, "vad_f1_raw": 0.9650457996816915, "vad_fa_enhanced": 0.09786018692279255, "vad_fa_raw": 0.07585414368555106, "wer_enhanced": 0.06315789473684211, "wer_raw": 0.11052631578947368, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr5": { "vad_f1_enhanced": 0.9607153562156581, "vad_f1_raw": 0.9655544660175688, "vad_fa_enhanced": 0.08036451327419945, "vad_fa_raw": 0.06337917413073132, "wer_enhanced": 0.11578947368421053, "wer_raw": 0.08947368421052632, "ins_enhanced": 3, "ins_raw": 2 }, "noise_snr10": { "vad_f1_enhanced": 0.9607995627662849, "vad_f1_raw": 0.9715691872453027, "vad_fa_enhanced": 0.07517488945140421, "vad_fa_raw": 0.0521201836224206, "wer_enhanced": 0.06315789473684211, "wer_raw": 0.06315789473684211, "ins_enhanced": 2, "ins_raw": 2 }, "noise_snr20": { "vad_f1_enhanced": 0.9634842371764554, "vad_f1_raw": 0.9722639322945248, "vad_fa_enhanced": 0.0989882280312728, "vad_fa_raw": 0.0769268153848264, "wer_enhanced": 0.05789473684210526, "wer_raw": 0.042105263157894736, "ins_enhanced": 2, "ins_raw": 2 }, "babble_snr5": { "vad_f1_enhanced": 0.847271888651999, "vad_f1_raw": 0.8446580505770814, "vad_fa_enhanced": 0.9596433700403029, "vad_fa_raw": 0.9770890756151038, "wer_enhanced": 0.2736842105263158, "wer_raw": 0.24736842105263157, "ins_enhanced": 6, "ins_raw": 10 }, "competing_snr0": { "vad_f1_enhanced": 0.8976787826711222, "vad_f1_raw": 0.8978559481992789, "vad_fa_enhanced": 0.5766324244238711, "vad_fa_raw": 0.5817642991147487, "wer_enhanced": 0.7947368421052632, "wer_raw": 0.8473684210526315, "ins_enhanced": 57, "ins_raw": 45 }, "competing_snr5": { "vad_f1_enhanced": 0.8986293123244862, "vad_f1_raw": 0.8993866413420953, "vad_fa_enhanced": 0.5756152417780034, "vad_fa_raw": 0.5812582262402548, "wer_enhanced": 0.3684210526315789, "wer_raw": 0.2631578947368421, "ins_enhanced": 30, "ins_raw": 19 }, "reverb": { "vad_f1_enhanced": 0.9579868395928663, "vad_f1_raw": 0.9623023734900757, "vad_fa_enhanced": 0.10152439665663182, "vad_fa_raw": 0.07937271734601117, "wer_enhanced": 0.06842105263157895, "wer_raw": 0.05789473684210526, "ins_enhanced": 1, "ins_raw": 1 }, "reverb_snr10": { "vad_f1_enhanced": 0.9523859457055025, "vad_f1_raw": 0.9565243987100389, "vad_fa_enhanced": 0.0979205227484904, "vad_fa_raw": 0.07830285083543445, "wer_enhanced": 0.15789473684210525, "wer_raw": 0.1368421052631579, "ins_enhanced": 3, "ins_raw": 0 }, "telephony": { "vad_f1_enhanced": 0.9614507398427781, "vad_f1_raw": 0.9761642766503543, "vad_fa_enhanced": 0.08403608096882278, "vad_fa_raw": 0.05308086433332433, "wer_enhanced": 0.07368421052631578, "wer_raw": 0.05789473684210526, "ins_enhanced": 2, "ins_raw": 1 }, "telephony_snr10": { "vad_f1_enhanced": 0.9574210139093473, "vad_f1_raw": 0.9688306828446971, "vad_fa_enhanced": 0.060592939216135976, "vad_fa_raw": 0.05104087749954463, "wer_enhanced": 0.12105263157894737, "wer_raw": 0.10526315789473684, "ins_enhanced": 4, "ins_raw": 3 }, "telephony_alaw": { "vad_f1_enhanced": 0.955208853759897, "vad_f1_raw": 0.9649250943221878, "vad_fa_enhanced": 0.05951805488700949, "vad_fa_raw": 0.04056293999789538, "wer_enhanced": 0.07368421052631578, "wer_raw": 0.07368421052631578, "ins_enhanced": 2, "ins_raw": 2 }, "telephony_g722": { "vad_f1_enhanced": 0.9633286137928344, "vad_f1_raw": 0.9718918630456168, "vad_fa_enhanced": 0.09278590360631025, "vad_fa_raw": 0.07682802264269852, "wer_enhanced": 0.06842105263157895, "wer_raw": 0.05263157894736842, "ins_enhanced": 3, "ins_raw": 1 }, "telephony_opus12k": { "vad_f1_enhanced": 0.9605505231284232, "vad_f1_raw": 0.9700698810886056, "vad_fa_enhanced": 0.09029396961269336, "vad_fa_raw": 0.06495112394829519, "wer_enhanced": 0.07368421052631578, "wer_raw": 0.06842105263157895, "ins_enhanced": 2, "ins_raw": 2 }, "telephony_loss3": { "vad_f1_enhanced": 0.950505258663141, "vad_f1_raw": 0.9585494127803207, "vad_fa_enhanced": 0.06229282988880808, "vad_fa_raw": 0.045096973984416894, "wer_enhanced": 0.11578947368421053, "wer_raw": 0.11052631578947368, "ins_enhanced": 1, "ins_raw": 2 }, "telephony_loss10": { "vad_f1_enhanced": 0.9494553369671153, "vad_f1_raw": 0.9568651776701316, "vad_fa_enhanced": 0.0639371384014472, "vad_fa_raw": 0.052286177577941614, "wer_enhanced": 0.10526315789473684, "wer_raw": 0.1, "ins_enhanced": 1, "ins_raw": 1 } } }, "resample8k": { "label": "Resample-only 16\u21928\u219216 (control)", "mean_wer_enhanced": 0.18625730994152045, "mean_wer_raw": 0.1476608187134503, "wer_delta_pp": 3.8596491228070158, "conditions_where_enhancement_helps_stt": 3, "conditions_scored": 18, "mean_vad_f1_enhanced": 0.947758705466079, "mean_vad_f1_raw": 0.9503607044937831, "vad_f1_delta": -0.0026019990277040828, "insertions_enhanced": 173, "insertions_raw": 101, "per_condition": { "clean": { "vad_f1_enhanced": 0.9712635637963618, "vad_f1_raw": 0.9756248909923848, "vad_fa_enhanced": 0.05200245311745607, "vad_fa_raw": 0.07023743685253026, "wer_enhanced": 0.05789473684210526, "wer_raw": 0.04736842105263158, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr-5": { "vad_f1_enhanced": 0.9377278146663655, "vad_f1_raw": 0.9284106039361377, "vad_fa_enhanced": 0.06968870677046633, "vad_fa_raw": 0.06810271419067943, "wer_enhanced": 0.22631578947368422, "wer_raw": 0.18421052631578946, "ins_enhanced": 2, "ins_raw": 4 }, "noise_snr0": { "vad_f1_enhanced": 0.9545007401758253, "vad_f1_raw": 0.9650457996816915, "vad_fa_enhanced": 0.0595259879130645, "vad_fa_raw": 0.07585414368555106, "wer_enhanced": 0.11052631578947368, "wer_raw": 0.11052631578947368, "ins_enhanced": 2, "ins_raw": 2 }, "noise_snr5": { "vad_f1_enhanced": 0.9582578022161936, "vad_f1_raw": 0.9655544660175688, "vad_fa_enhanced": 0.044553263199557255, "vad_fa_raw": 0.06337917413073132, "wer_enhanced": 0.15263157894736842, "wer_raw": 0.08947368421052632, "ins_enhanced": 3, "ins_raw": 2 }, "noise_snr10": { "vad_f1_enhanced": 0.9615197503069066, "vad_f1_raw": 0.9715691872453027, "vad_fa_enhanced": 0.0481076787123384, "vad_fa_raw": 0.0521201836224206, "wer_enhanced": 0.07894736842105263, "wer_raw": 0.06315789473684211, "ins_enhanced": 2, "ins_raw": 2 }, "noise_snr20": { "vad_f1_enhanced": 0.9663056465859977, "vad_f1_raw": 0.9722639322945248, "vad_fa_enhanced": 0.05101900866322123, "vad_fa_raw": 0.0769268153848264, "wer_enhanced": 0.06315789473684211, "wer_raw": 0.042105263157894736, "ins_enhanced": 1, "ins_raw": 2 }, "babble_snr5": { "vad_f1_enhanced": 0.8457910902708192, "vad_f1_raw": 0.8446580505770814, "vad_fa_enhanced": 0.9675959456163833, "vad_fa_raw": 0.9770890756151038, "wer_enhanced": 0.39473684210526316, "wer_raw": 0.24736842105263157, "ins_enhanced": 19, "ins_raw": 10 }, "competing_snr0": { "vad_f1_enhanced": 0.8999935792990227, "vad_f1_raw": 0.8978559481992789, "vad_fa_enhanced": 0.5502636515372893, "vad_fa_raw": 0.5817642991147487, "wer_enhanced": 0.8263157894736842, "wer_raw": 0.8473684210526315, "ins_enhanced": 78, "ins_raw": 45 }, "competing_snr5": { "vad_f1_enhanced": 0.9014767841893823, "vad_f1_raw": 0.8993866413420953, "vad_fa_enhanced": 0.5514102682937138, "vad_fa_raw": 0.5812582262402548, "wer_enhanced": 0.47368421052631576, "wer_raw": 0.2631578947368421, "ins_enhanced": 46, "ins_raw": 19 }, "reverb": { "vad_f1_enhanced": 0.9583328025791953, "vad_f1_raw": 0.9623023734900757, "vad_fa_enhanced": 0.05963797890143611, "vad_fa_raw": 0.07937271734601117, "wer_enhanced": 0.12105263157894737, "wer_raw": 0.05789473684210526, "ins_enhanced": 2, "ins_raw": 1 }, "reverb_snr10": { "vad_f1_enhanced": 0.9554963044744713, "vad_f1_raw": 0.9565243987100389, "vad_fa_enhanced": 0.06122488482456919, "vad_fa_raw": 0.07830285083543445, "wer_enhanced": 0.23157894736842105, "wer_raw": 0.1368421052631579, "ins_enhanced": 6, "ins_raw": 0 }, "telephony": { "vad_f1_enhanced": 0.9714686487029821, "vad_f1_raw": 0.9761642766503543, "vad_fa_enhanced": 0.04815629927130223, "vad_fa_raw": 0.05308086433332433, "wer_enhanced": 0.042105263157894736, "wer_raw": 0.05789473684210526, "ins_enhanced": 1, "ins_raw": 1 }, "telephony_snr10": { "vad_f1_enhanced": 0.9677159386790878, "vad_f1_raw": 0.9688306828446971, "vad_fa_enhanced": 0.036804818945029405, "vad_fa_raw": 0.05104087749954463, "wer_enhanced": 0.07894736842105263, "wer_raw": 0.10526315789473684, "ins_enhanced": 2, "ins_raw": 3 }, "telephony_alaw": { "vad_f1_enhanced": 0.9655068779155312, "vad_f1_raw": 0.9649250943221878, "vad_fa_enhanced": 0.039532012162843834, "vad_fa_raw": 0.04056293999789538, "wer_enhanced": 0.08947368421052632, "wer_raw": 0.07368421052631578, "ins_enhanced": 3, "ins_raw": 2 }, "telephony_g722": { "vad_f1_enhanced": 0.964512414320537, "vad_f1_raw": 0.9718918630456168, "vad_fa_enhanced": 0.0585150530804621, "vad_fa_raw": 0.07682802264269852, "wer_enhanced": 0.07368421052631578, "wer_raw": 0.05263157894736842, "ins_enhanced": 1, "ins_raw": 1 }, "telephony_opus12k": { "vad_f1_enhanced": 0.964156603322062, "vad_f1_raw": 0.9700698810886056, "vad_fa_enhanced": 0.03710981279825136, "vad_fa_raw": 0.06495112394829519, "wer_enhanced": 0.11052631578947368, "wer_raw": 0.06842105263157895, "ins_enhanced": 1, "ins_raw": 2 }, "telephony_loss3": { "vad_f1_enhanced": 0.9588668108994429, "vad_f1_raw": 0.9585494127803207, "vad_fa_enhanced": 0.045096973984416894, "vad_fa_raw": 0.045096973984416894, "wer_enhanced": 0.11578947368421053, "wer_raw": 0.11052631578947368, "ins_enhanced": 2, "ins_raw": 2 }, "telephony_loss10": { "vad_f1_enhanced": 0.9567635259892378, "vad_f1_raw": 0.9568651776701316, "vad_fa_enhanced": 0.05352074547917619, "vad_fa_raw": 0.052286177577941614, "wer_enhanced": 0.10526315789473684, "wer_raw": 0.1, "ins_enhanced": 1, "ins_raw": 1 } } }, "clearline": { "label": "Clearline (8 kHz, resampled)", "mean_wer_enhanced": 0.39649122807017545, "mean_wer_raw": 0.1476608187134503, "wer_delta_pp": 24.883040935672515, "conditions_where_enhancement_helps_stt": 0, "conditions_scored": 18, "mean_vad_f1_enhanced": 0.8352046826888598, "mean_vad_f1_raw": 0.9503607044937831, "vad_f1_delta": -0.11515602180492324, "insertions_enhanced": 410, "insertions_raw": 101, "per_condition": { "clean": { "vad_f1_enhanced": 0.9636592006440843, "vad_f1_raw": 0.9756248909923848, "vad_fa_enhanced": 0.05628302376274158, "vad_fa_raw": 0.07023743685253026, "wer_enhanced": 0.05789473684210526, "wer_raw": 0.04736842105263158, "ins_enhanced": 1, "ins_raw": 2 }, "noise_snr-5": { "vad_f1_enhanced": 0.5584285632892494, "vad_f1_raw": 0.9284106039361377, "vad_fa_enhanced": 0.07300209318571076, "vad_fa_raw": 0.06810271419067943, "wer_enhanced": 0.6368421052631579, "wer_raw": 0.18421052631578946, "ins_enhanced": 9, "ins_raw": 4 }, "noise_snr0": { "vad_f1_enhanced": 0.8839959290566209, "vad_f1_raw": 0.9650457996816915, "vad_fa_enhanced": 0.06500871047714056, "vad_fa_raw": 0.07585414368555106, "wer_enhanced": 0.2894736842105263, "wer_raw": 0.11052631578947368, "ins_enhanced": 8, "ins_raw": 2 }, "noise_snr5": { "vad_f1_enhanced": 0.7515542213937226, "vad_f1_raw": 0.9655544660175688, "vad_fa_enhanced": 0.03119914383391593, "vad_fa_raw": 0.06337917413073132, "wer_enhanced": 0.33157894736842103, "wer_raw": 0.08947368421052632, "ins_enhanced": 14, "ins_raw": 2 }, "noise_snr10": { "vad_f1_enhanced": 0.9443150125262626, "vad_f1_raw": 0.9715691872453027, "vad_fa_enhanced": 0.06703351482728162, "vad_fa_raw": 0.0521201836224206, "wer_enhanced": 0.10526315789473684, "wer_raw": 0.06315789473684211, "ins_enhanced": 2, "ins_raw": 2 }, "noise_snr20": { "vad_f1_enhanced": 0.9614350998059799, "vad_f1_raw": 0.9722639322945248, "vad_fa_enhanced": 0.058599127332328325, "vad_fa_raw": 0.0769268153848264, "wer_enhanced": 0.06842105263157895, "wer_raw": 0.042105263157894736, "ins_enhanced": 1, "ins_raw": 2 }, "babble_snr5": { "vad_f1_enhanced": 0.8412910206701305, "vad_f1_raw": 0.8446580505770814, "vad_fa_enhanced": 0.8245655397508633, "vad_fa_raw": 0.9770890756151038, "wer_enhanced": 1.3842105263157896, "wer_raw": 0.24736842105263157, "ins_enhanced": 169, "ins_raw": 10 }, "competing_snr0": { "vad_f1_enhanced": 0.8340497505651937, "vad_f1_raw": 0.8978559481992789, "vad_fa_enhanced": 0.5470619226644609, "vad_fa_raw": 0.5817642991147487, "wer_enhanced": 1.068421052631579, "wer_raw": 0.8473684210526315, "ins_enhanced": 36, "ins_raw": 45 }, "competing_snr5": { "vad_f1_enhanced": 0.8456747668106559, "vad_f1_raw": 0.8993866413420953, "vad_fa_enhanced": 0.5064379447044965, "vad_fa_raw": 0.5812582262402548, "wer_enhanced": 0.8736842105263158, "wer_raw": 0.2631578947368421, "ins_enhanced": 29, "ins_raw": 19 }, "reverb": { "vad_f1_enhanced": 0.9499291448779305, "vad_f1_raw": 0.9623023734900757, "vad_fa_enhanced": 0.0775710381248718, "vad_fa_raw": 0.07937271734601117, "wer_enhanced": 0.12105263157894737, "wer_raw": 0.05789473684210526, "ins_enhanced": 2, "ins_raw": 1 }, "reverb_snr10": { "vad_f1_enhanced": 0.40941952940389925, "vad_f1_raw": 0.9565243987100389, "vad_fa_enhanced": 0.03473534263007947, "vad_fa_raw": 0.07830285083543445, "wer_enhanced": 1.1631578947368422, "wer_raw": 0.1368421052631579, "ins_enhanced": 117, "ins_raw": 0 }, "telephony": { "vad_f1_enhanced": 0.9650165632066645, "vad_f1_raw": 0.9761642766503543, "vad_fa_enhanced": 0.04909296296443336, "vad_fa_raw": 0.05308086433332433, "wer_enhanced": 0.05789473684210526, "wer_raw": 0.05789473684210526, "ins_enhanced": 1, "ins_raw": 1 }, "telephony_snr10": { "vad_f1_enhanced": 0.9077547314304052, "vad_f1_raw": 0.9688306828446971, "vad_fa_enhanced": 0.05754845231043152, "vad_fa_raw": 0.05104087749954463, "wer_enhanced": 0.14736842105263157, "wer_raw": 0.10526315789473684, "ins_enhanced": 4, "ins_raw": 3 }, "telephony_alaw": { "vad_f1_enhanced": 0.8641218393092366, "vad_f1_raw": 0.9649250943221878, "vad_fa_enhanced": 0.04977519763543309, "vad_fa_raw": 0.04056293999789538, "wer_enhanced": 0.1368421052631579, "wer_raw": 0.07368421052631578, "ins_enhanced": 4, "ins_raw": 2 }, "telephony_g722": { "vad_f1_enhanced": 0.8669201896983616, "vad_f1_raw": 0.9718918630456168, "vad_fa_enhanced": 0.06384447644207934, "vad_fa_raw": 0.07682802264269852, "wer_enhanced": 0.09473684210526316, "wer_raw": 0.05263157894736842, "ins_enhanced": 2, "ins_raw": 1 }, "telephony_opus12k": { "vad_f1_enhanced": 0.7223385358402183, "vad_f1_raw": 0.9700698810886056, "vad_fa_enhanced": 0.036662546660987026, "vad_fa_raw": 0.06495112394829519, "wer_enhanced": 0.2736842105263158, "wer_raw": 0.06842105263157895, "ins_enhanced": 8, "ins_raw": 2 }, "telephony_loss3": { "vad_f1_enhanced": 0.8872113747105181, "vad_f1_raw": 0.9585494127803207, "vad_fa_enhanced": 0.054581839293268396, "vad_fa_raw": 0.045096973984416894, "wer_enhanced": 0.17894736842105263, "wer_raw": 0.11052631578947368, "ins_enhanced": 1, "ins_raw": 2 }, "telephony_loss10": { "vad_f1_enhanced": 0.8765688151603429, "vad_f1_raw": 0.9568651776701316, "vad_fa_enhanced": 0.07053908619429598, "vad_fa_raw": 0.052286177577941614, "wer_enhanced": 0.14736842105263157, "wer_raw": 0.1, "ins_enhanced": 2, "ins_raw": 1 } } } }, "caveats": [ "DeepFilterNet3 is measured through a **block-online** adapter: every public ONNX export of it lacks recurrent state tensors, so per-frame streaming is impossible with the published weights. Its RTF here is an upper bound and its latency (100 ms) is far above the ~40 ms a native (Rust/tract) runtime achieves. Read the DFN3 rows as a lower bound on quality and an upper bound on cost.", "ai-coustics rows are a **proprietary baseline** run through the licensed aic-sdk. They are a reference point, never a shipping code path.", "WER is measured with faster-whisper:base.en (CTranslate2 int8 CPU, greedy, model=base.en). A different recogniser will give different absolute numbers; the harness supports Deepgram and AssemblyAI so the *ordering* can be checked on another engine.", "**Clearline is an 8 kHz model and these conditions are stored at 16 kHz**, so its row is measured through the harness's resample path \u2014 16 k -> 8 k -> model -> 8 k -> 16 k (soxr), the same path the native-8 kHz ai-coustics baselines take. On `clean`, `noise_*`, `babble_snr5`, `competing_*` and `reverb*` that is not its operating point: the 4-8 kHz band is discarded before the model ever sees it. The `resample8k` control row is that round trip **with no model in it**, so the band limit and the model can be told apart; and the telephony sub-aggregate below reports the regime the model was actually built for.", "**The Clearline checkpoint benchmarked here is not shippable, and this row is not a claim that it is.** Two gates it does not pass, both measured on its own eval (`runs/t4_300h_200ep_solo`, 1,800-clip test_v2) and neither of them visible in this table: (1) a **solo regression** \u2014 on clips with no competing speaker the model makes the audio worse, -0.874 dB SI-SDR with 37.1% of solo clips degraded (-0.595 dB / 33.0% on val_v2); (2) **speaker conditioning contributes approximately nothing** \u2014 the matched ablation scores FiLM +2.815 dB against no-conditioning +2.690 dB, a 0.125 dB difference against a 0.3 dB noise floor, so what is being measured here is a plain denoiser and not the speaker isolation the architecture is for. The row exists so Clearline can be *placed* next to the competition on identical audio, not to argue that it is ready.", "Clearline's own eval (`clearline/evaluate.py`, 1,800 clips, difficulty buckets) and this leaderboard are **not comparable and were never meant to be**: different clips, different conditions, different pooling. Only the recogniser is shared. This row exists precisely so that a comparable number exists.", "DeepFilterNet3 runs at 48 kHz on 16 kHz source material upsampled to 48 kHz \u2014 there is no real content above 8 kHz for it to work with. That is the honest telephony/VoIP situation, but it is not the condition DFN3 was designed for.", "The test set is 10 speakers / ~190 reference words per condition. That is enough to separate large effects and not enough to resolve differences of a point or two of WER.", "**Run-to-run repeatability of the recogniser was measured, not assumed \u2014 and the standing warning did not reproduce here.** Elsewhere in this project faster-whisper on CPU has been seen to return different transcripts for byte-identical input (three runs of one aggregate gave 121.65 / 122.15 / 121.40%), and the standing guidance from that is to treat differences below roughly 6 pp on a small aggregate as decoder noise. On this host and this harness it did not happen: 25/25 scopes were bit-identical across repeated decodes (raw on 18 condition(s), 180 clips x 3 decodes; clearline on 5 condition(s), 50 clips x 3 decodes), including the arms where the recogniser is already hallucinating past 100% WER. Independently, re-measuring `raw` and `gtcrn` in a fresh process two days after the published run reproduced all 54 rows exactly (max \u0394WER 0.0000 pp). Three repeats cannot prove determinism and a different host, venv or thread count may well behave differently, so the conservative 6 pp guidance still governs anything quoted from another machine \u2014 but on **this** table the resolution limit is set by the size of the test set, not by the decoder: one reference word is 0.53 pp in a single condition (190 words), and about 0.03 pp pooled across all 18. Differences of a point or two between adjacent rows are still not results; differences of tens of points are.", "onnxruntime is pinned to 1 intra-op thread with sequential execution: the measured cost is one stream on one core, not a fan-out across the whole machine.", "The `telephony*` conditions in this table are stored at 16 kHz after a full 8 kHz carrier leg (band-limit, codec, packet loss), so every backend sees them at its own native rate. The genuinely narrowband experiment \u2014 where nothing is resampled \u2014 is a separate artefact (`bench/out/telephony.json`).", "Packet loss is modelled as bursty (2-state Gilbert-Elliott) erasure of 20 ms frames with attenuated-repeat concealment, not i.i.d. loss with silence fill. i.i.d. loss flatters every decoder and silence fill overstates the damage." ], "skipped": {}, "telephony": { "schema": 1, "generatedAt": "2026-08-14T01:28:47Z", "narrowSampleRate": 8000, "sttIngestResampler": "soxr_vhq", "conditions": [ "nb_clean", "nb_snr10", "nb_snr0", "nb_alaw_snr10", "nb_loss3_snr10", "nb_loss10_snr10" ], "resamplers": { "soxr_vhq": "polyphase/sinc, very high quality \u2014 the correct implementation", "linear": "linear interpolation, no anti-alias filter", "naive": "sample-repeat up / sample-drop down, no filtering at all" }, "clips": 60, "rows": [ { "path": "raw8k", "backend": null, "resampler": null, "wer": 0.11140350877192982, "insertions": 12, "refWords": 1140, "siSdrDb": 12.525213474088531, "pesqNb": 2.458134686946869, "stoi": 0.8844167191926248, "rtf": null, "vsRaw8kPp": null, "vsCorrectResamplerPp": null, "conditions": { "nb_clean": 0.05789473684210526, "nb_snr10": 0.1, "nb_snr0": 0.1631578947368421, "nb_alaw_snr10": 0.09473684210526316, "nb_loss3_snr10": 0.11578947368421053, "nb_loss10_snr10": 0.1368421052631579 } }, { "path": "resample_only", "backend": null, "resampler": "linear", "wer": 0.11140350877192982, "insertions": 12, "refWords": 1140, "siSdrDb": 12.525213474088531, "pesqNb": 2.458134686946869, "stoi": 0.8844167191926248, "rtf": null, "vsRaw8kPp": 0.0, "vsCorrectResamplerPp": 0.17543859649122862, "conditions": { "nb_clean": 0.05789473684210526, "nb_snr10": 0.1, "nb_snr0": 0.1631578947368421, "nb_alaw_snr10": 0.09473684210526316, "nb_loss3_snr10": 0.11578947368421053, "nb_loss10_snr10": 0.1368421052631579 } }, { "path": "resample_only", "backend": null, "resampler": "naive", "wer": 0.11140350877192982, "insertions": 12, "refWords": 1140, "siSdrDb": 12.525213474088531, "pesqNb": 2.458134686946869, "stoi": 0.8844167191926248, "rtf": null, "vsRaw8kPp": 0.0, "vsCorrectResamplerPp": 0.17543859649122862, "conditions": { "nb_clean": 0.05789473684210526, "nb_snr10": 0.1, "nb_snr0": 0.1631578947368421, "nb_alaw_snr10": 0.09473684210526316, "nb_loss3_snr10": 0.11578947368421053, "nb_loss10_snr10": 0.1368421052631579 } }, { "path": "resample_only", "backend": null, "resampler": "soxr_vhq", "wer": 0.10964912280701754, "insertions": 11, "refWords": 1140, "siSdrDb": 12.162170890416405, "pesqNb": 2.45813464919726, "stoi": 0.8845007792037541, "rtf": null, "vsRaw8kPp": -0.17543859649122862, "vsCorrectResamplerPp": 0.0, "conditions": { "nb_clean": 0.05789473684210526, "nb_snr10": 0.1, "nb_snr0": 0.1631578947368421, "nb_alaw_snr10": 0.09473684210526316, "nb_loss3_snr10": 0.11578947368421053, "nb_loss10_snr10": 0.12631578947368421 } }, { "path": "native8k", "backend": "aic-quail-l-8khz", "resampler": null, "wer": 0.11052631578947368, "insertions": 9, "refWords": 1140, "siSdrDb": 13.773610728612715, "pesqNb": 2.6548771460851035, "stoi": 0.9011927464204977, "rtf": 0.09945269982619676, "vsRaw8kPp": -0.08771929824561431, "vsCorrectResamplerPp": null, "conditions": { "nb_clean": 0.05789473684210526, "nb_snr10": 0.10526315789473684, "nb_snr0": 0.1368421052631579, "nb_alaw_snr10": 0.09473684210526316, "nb_loss3_snr10": 0.12105263157894737, "nb_loss10_snr10": 0.14736842105263157 } }, { "path": "native8k", "backend": "aic-quail-s-8khz", "resampler": null, "wer": 0.11315789473684211, "insertions": 11, "refWords": 1140, "siSdrDb": 13.076862054972684, "pesqNb": 2.5477500796318053, "stoi": 0.8931104679822167, "rtf": 0.03222398439002159, "vsRaw8kPp": 0.17543859649122862, "vsCorrectResamplerPp": null, "conditions": { "nb_clean": 0.05263157894736842, "nb_snr10": 0.10526315789473684, "nb_snr0": 0.1631578947368421, "nb_alaw_snr10": 0.10526315789473684, "nb_loss3_snr10": 0.12105263157894737, "nb_loss10_snr10": 0.13157894736842105 } }, { "path": "native8k", "backend": "aic-rook-l-8khz", "resampler": null, "wer": 0.12456140350877193, "insertions": 11, "refWords": 1140, "siSdrDb": 15.257852693419673, "pesqNb": 3.118241818745931, "stoi": 0.9234311934504228, "rtf": 0.10019061789966811, "vsRaw8kPp": 1.3157894736842104, "vsCorrectResamplerPp": null, "conditions": { "nb_clean": 0.05789473684210526, "nb_snr10": 0.10526315789473684, "nb_snr0": 0.18947368421052632, "nb_alaw_snr10": 0.1, "nb_loss3_snr10": 0.13157894736842105, "nb_loss10_snr10": 0.1631578947368421 } }, { "path": "upsampled", "backend": "aic-quail", "resampler": "linear", "wer": 0.11140350877192982, "insertions": 8, "refWords": 1140, "siSdrDb": 14.226466423885753, "pesqNb": 2.7789712250232697, "stoi": 0.9067417806570496, "rtf": 0.11800341615549469, "vsRaw8kPp": 0.0, "vsCorrectResamplerPp": -0.6140350877192988, "conditions": { "nb_clean": 0.05789473684210526, "nb_snr10": 0.09473684210526316, "nb_snr0": 0.15263157894736842, "nb_alaw_snr10": 0.10526315789473684, "nb_loss3_snr10": 0.11578947368421053, "nb_loss10_snr10": 0.14210526315789473 } }, { "path": "upsampled", "backend": "aic-quail", "resampler": "naive", "wer": 0.11315789473684211, "insertions": 9, "refWords": 1140, "siSdrDb": 14.236659840480682, "pesqNb": 2.7824217538038885, "stoi": 0.9074872304365146, "rtf": 0.11992096472242043, "vsRaw8kPp": 0.17543859649122862, "vsCorrectResamplerPp": -0.43859649122807015, "conditions": { "nb_clean": 0.05789473684210526, "nb_snr10": 0.11052631578947368, "nb_snr0": 0.14736842105263157, "nb_alaw_snr10": 0.11052631578947368, "nb_loss3_snr10": 0.12105263157894737, "nb_loss10_snr10": 0.13157894736842105 } }, { "path": "upsampled", "backend": "aic-quail", "resampler": "soxr_vhq", "wer": 0.11754385964912281, "insertions": 17, "refWords": 1140, "siSdrDb": 13.114288793412335, "pesqNb": 2.7524392704168954, "stoi": 0.9019117188424168, "rtf": 0.11808472569661514, "vsRaw8kPp": 0.6140350877192988, "vsCorrectResamplerPp": 0.0, "conditions": { "nb_clean": 0.05789473684210526, "nb_snr10": 0.11578947368421053, "nb_snr0": 0.14736842105263157, "nb_alaw_snr10": 0.11578947368421053, "nb_loss3_snr10": 0.14736842105263157, "nb_loss10_snr10": 0.12105263157894737 } }, { "path": "upsampled", "backend": "aic-quail-vf", "resampler": "soxr_vhq", "wer": 0.12982456140350876, "insertions": 19, "refWords": 1140, "siSdrDb": 7.968834377985416, "pesqNb": 2.673751833041509, "stoi": 0.8951284096810702, "rtf": 0.07728767551356262, "vsRaw8kPp": 1.8421052631578936, "vsCorrectResamplerPp": 0.0, "conditions": { "nb_clean": 0.06842105263157895, "nb_snr10": 0.1, "nb_snr0": 0.22631578947368422, "nb_alaw_snr10": 0.1, "nb_loss3_snr10": 0.13157894736842105, "nb_loss10_snr10": 0.15263157894736842 } }, { "path": "upsampled", "backend": "fastenhancer-b", "resampler": "linear", "wer": 0.12017543859649123, "insertions": 10, "refWords": 1140, "siSdrDb": 15.328257608916642, "pesqNb": 2.7815293451150254, "stoi": 0.8917028585787793, "rtf": 0.02197351266748906, "vsRaw8kPp": 0.8771929824561403, "vsCorrectResamplerPp": -5.614035087719298, "conditions": { "nb_clean": 0.07368421052631578, "nb_snr10": 0.08947368421052632, "nb_snr0": 0.1736842105263158, "nb_alaw_snr10": 0.09473684210526316, "nb_loss3_snr10": 0.12631578947368421, "nb_loss10_snr10": 0.1631578947368421 } }, { "path": "upsampled", "backend": "fastenhancer-b", "resampler": "naive", "wer": 0.12982456140350876, "insertions": 18, "refWords": 1140, "siSdrDb": 15.468859028767662, "pesqNb": 2.890658458073934, "stoi": 0.8956259925793467, "rtf": 0.02149461005252426, "vsRaw8kPp": 1.8421052631578936, "vsCorrectResamplerPp": -4.6491228070175445, "conditions": { "nb_clean": 0.08421052631578947, "nb_snr10": 0.08947368421052632, "nb_snr0": 0.19473684210526315, "nb_alaw_snr10": 0.11052631578947368, "nb_loss3_snr10": 0.1368421052631579, "nb_loss10_snr10": 0.1631578947368421 } }, { "path": "upsampled", "backend": "fastenhancer-b", "resampler": "soxr_vhq", "wer": 0.1763157894736842, "insertions": 13, "refWords": 1140, "siSdrDb": 12.283759857506311, "pesqNb": 2.378990751504898, "stoi": 0.8559385093106865, "rtf": 0.022410925492694984, "vsRaw8kPp": 6.491228070175438, "vsCorrectResamplerPp": 0.0, "conditions": { "nb_clean": 0.1, "nb_snr10": 0.1368421052631579, "nb_snr0": 0.2789473684210526, "nb_alaw_snr10": 0.13157894736842105, "nb_loss3_snr10": 0.17894736842105263, "nb_loss10_snr10": 0.23157894736842105 } }, { "path": "upsampled", "backend": "fastenhancer-m", "resampler": "soxr_vhq", "wer": 0.21140350877192982, "insertions": 12, "refWords": 1140, "siSdrDb": 12.930181684345479, "pesqNb": 2.354076820611954, "stoi": 0.8676525489969052, "rtf": 0.10996475338227442, "vsRaw8kPp": 10.0, "vsCorrectResamplerPp": 0.0, "conditions": { "nb_clean": 0.25263157894736843, "nb_snr10": 0.14736842105263157, "nb_snr0": 0.24736842105263157, "nb_alaw_snr10": 0.16842105263157894, "nb_loss3_snr10": 0.21052631578947367, "nb_loss10_snr10": 0.24210526315789474 } }, { "path": "upsampled", "backend": "fastenhancer-t", "resampler": "soxr_vhq", "wer": 0.2482456140350877, "insertions": 21, "refWords": 1140, "siSdrDb": 11.363794248957232, "pesqNb": 2.19638086160024, "stoi": 0.8003671162252292, "rtf": 0.012468538393757842, "vsRaw8kPp": 13.684210526315788, "vsCorrectResamplerPp": 0.0, "conditions": { "nb_clean": 0.14736842105263157, "nb_snr10": 0.2, "nb_snr0": 0.3736842105263158, "nb_alaw_snr10": 0.19473684210526315, "nb_loss3_snr10": 0.22631578947368422, "nb_loss10_snr10": 0.3473684210526316 } }, { "path": "upsampled", "backend": "gtcrn", "resampler": "linear", "wer": 0.1368421052631579, "insertions": 20, "refWords": 1140, "siSdrDb": 15.833773176177793, "pesqNb": 3.0203232904275255, "stoi": 0.9069205221177506, "rtf": 0.04735899624559134, "vsRaw8kPp": 2.543859649122808, "vsCorrectResamplerPp": -2.894736842105261, "conditions": { "nb_clean": 0.05789473684210526, "nb_snr10": 0.11578947368421053, "nb_snr0": 0.17894736842105263, "nb_alaw_snr10": 0.11052631578947368, "nb_loss3_snr10": 0.13157894736842105, "nb_loss10_snr10": 0.22631578947368422 } }, { "path": "upsampled", "backend": "gtcrn", "resampler": "naive", "wer": 0.12894736842105264, "insertions": 17, "refWords": 1140, "siSdrDb": 14.69574660064911, "pesqNb": 2.979688175519307, "stoi": 0.905108856130799, "rtf": 0.046923498565533295, "vsRaw8kPp": 1.754385964912282, "vsCorrectResamplerPp": -3.684210526315787, "conditions": { "nb_clean": 0.07368421052631578, "nb_snr10": 0.09473684210526316, "nb_snr0": 0.15789473684210525, "nb_alaw_snr10": 0.11052631578947368, "nb_loss3_snr10": 0.13157894736842105, "nb_loss10_snr10": 0.20526315789473684 } }, { "path": "upsampled", "backend": "gtcrn", "resampler": "soxr_vhq", "wer": 0.16578947368421051, "insertions": 16, "refWords": 1140, "siSdrDb": 13.856523068781831, "pesqNb": 2.7123501221338913, "stoi": 0.8768258575298433, "rtf": 0.04581640576161222, "vsRaw8kPp": 5.438596491228069, "vsCorrectResamplerPp": 0.0, "conditions": { "nb_clean": 0.06842105263157895, "nb_snr10": 0.14736842105263157, "nb_snr0": 0.2578947368421053, "nb_alaw_snr10": 0.14210526315789473, "nb_loss3_snr10": 0.1631578947368421, "nb_loss10_snr10": 0.21578947368421053 } } ], "roundtripControl": [ { "via_hz": 16000, "resampler": "soxr_vhq", "note": "polyphase/sinc, very high quality \u2014 the correct implementation", "si_sdr_vs_input_db": 38.27942749939827, "clips": 60 }, { "via_hz": 16000, "resampler": "soxr_hq", "note": "polyphase/sinc at the library default quality", "si_sdr_vs_input_db": 38.451110814678685, "clips": 60 }, { "via_hz": 16000, "resampler": "linear", "note": "linear interpolation, no anti-alias filter", "si_sdr_vs_input_db": 150.0, "clips": 60 }, { "via_hz": 16000, "resampler": "naive", "note": "sample-repeat up / sample-drop down, no filtering at all", "si_sdr_vs_input_db": 150.0, "clips": 60 }, { "via_hz": 24000, "resampler": "soxr_vhq", "note": "polyphase/sinc, very high quality \u2014 the correct implementation", "si_sdr_vs_input_db": 38.2701620741559, "clips": 60 }, { "via_hz": 24000, "resampler": "soxr_hq", "note": "polyphase/sinc at the library default quality", "si_sdr_vs_input_db": 38.43527783205894, "clips": 60 }, { "via_hz": 24000, "resampler": "linear", "note": "linear interpolation, no anti-alias filter", "si_sdr_vs_input_db": 150.0, "clips": 60 }, { "via_hz": 24000, "resampler": "naive", "note": "sample-repeat up / sample-drop down, no filtering at all", "si_sdr_vs_input_db": 150.0, "clips": 60 }, { "via_hz": 44100, "resampler": "soxr_vhq", "note": "polyphase/sinc, very high quality \u2014 the correct implementation", "si_sdr_vs_input_db": 38.359401444832194, "clips": 60 }, { "via_hz": 44100, "resampler": "soxr_hq", "note": "polyphase/sinc at the library default quality", "si_sdr_vs_input_db": 38.59377953095991, "clips": 60 }, { "via_hz": 44100, "resampler": "linear", "note": "linear interpolation, no anti-alias filter", "si_sdr_vs_input_db": 32.493585013196075, "clips": 60 }, { "via_hz": 44100, "resampler": "naive", "note": "sample-repeat up / sample-drop down, no filtering at all", "si_sdr_vs_input_db": 150.0, "clips": 60 }, { "via_hz": 48000, "resampler": "soxr_vhq", "note": "polyphase/sinc, very high quality \u2014 the correct implementation", "si_sdr_vs_input_db": 38.31835918379795, "clips": 60 }, { "via_hz": 48000, "resampler": "soxr_hq", "note": "polyphase/sinc at the library default quality", "si_sdr_vs_input_db": 38.51518775445072, "clips": 60 }, { "via_hz": 48000, "resampler": "linear", "note": "linear interpolation, no anti-alias filter", "si_sdr_vs_input_db": 150.0, "clips": 60 }, { "via_hz": 48000, "resampler": "naive", "note": "sample-repeat up / sample-drop down, no filtering at all", "si_sdr_vs_input_db": 150.0, "clips": 60 } ], "nativeVsUpsampled": { "native_backend": "aic-quail-l-8khz", "upsampled_backend": "aic-quail", "native_wer": 0.11052631578947368, "upsampled_wer": 0.11754385964912281, "delta_pp": -0.7017543859649131, "note": "same vendor, same family, 8 kHz weights vs 16 kHz weights on a resampled path" }, "findings": [ "**Nobody ships speaker isolation at 8 kHz, and almost nobody ships 8 kHz at all.** Every ai-coustics Voice Focus and VAD-VF model in their catalogue is 16 kHz. ai-coustics is the *only* vendor here with any genuine narrowband enhancement weights (`aic-quail-l-8khz`, `aic-quail-s-8khz`, `aic-rook-l-8khz`); GTCRN, FastEnhancer and DeepFilterNet3 have no 8 kHz checkpoint at all, so on a phone call they can only be run through a resample wrapper. That gap is structural, and it is the wedge \u2014 the opportunity is scarcity, not damage.", "**Native 8 kHz beats the resampled wideband path by 0.7 pp of WER** for the same vendor (`aic-quail-l-8khz` 11.1% vs `aic-quail` 11.8%). Real, but modest \u2014 within a small multiple of the 0.09 pp resolution limit, so it is a reason to build native, not a headline number to market.", "**Nothing beat the untouched caller audio.** The best path (aic-quail-l-8khz) ties the raw8k control \u2014 11.05% vs 11.14%, a 0.09 pp difference, which is 1 word out of 1,140. Every other path is worse than doing nothing. On narrowband telephony the null hypothesis is still undefeated, which is precisely the space Clearline has to win.", "**The resample round trip on its own does nothing \u2014 the 'resample tax' premise did not survive its control.** 8k->16k->8k with a naive sample-repeat/sample-drop resampler scores 150 dB SI-SDR against its input, because at an exact 2:1 ratio it is algebraically the identity. A correct polyphase resampler scores 38 dB \u2014 *lower*, because it actually filters. Any claim that resampling per se destroys a fixed fraction of caller audio is not supported here, and we should not make one.", "**With no model in the path, resampler quality is worth 0.18 pp of WER** \u2014 `soxr_vhq` 10.96% against `linear` 11.14%, versus a raw8k control of 11.14%. That is 2 words; it is a tie.", "**With a wideband model in the path, resampler choice moved WER by -3.0 pp on average \u2014 and it moved the wrong way.** Sign convention: positive = the naive/linear resampler produced more errors than `soxr_vhq`. Across 6 (model, resampler) pairs the mean is -3.0 pp, so on balance the *naive* resampler transcribed better. Every single pair is negative \u2014 the naive resampler helped *every* model, not just on average. Least affected: `aic-quail`/`naive` at -0.4 pp. Most affected: `fastenhancer-b`/`linear` at -5.6 pp. The plausible mechanism is that a polyphase resampler is an extra filtering stage the model was not trained on, while sample-repeat/sample-drop at 2:1 is an identity \u2014 so the naive path hands the model a *less* altered signal. Either way the data does not support a resample-quality claim." ], "skipped": {} }, "regimes": { "raw": { "label": "Raw (no processing)", "native_sample_rate": 16000, "resampled": false, "telephony": { "conditions": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "conditions_scored": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "wer": 0.081203007518797, "ref_words": 1330, "substitutions": 90, "insertions": 12, "deletions": 6, "si_sdr_db": 8.655543798473788, "pesq_wb": 1.8996521898678371, "stoi": 0.9051823174004429 }, "wideband": { "conditions": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "conditions_scored": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "wer": 0.18995215311004784, "ref_words": 2090, "substitutions": 254, "insertions": 89, "deletions": 54, "si_sdr_db": 12.362215768811991, "pesq_wb": 1.7653018582950937, "stoi": 0.8045638168316361 } }, "passthrough": { "label": "Passthrough (control)", "native_sample_rate": 16000, "resampled": false, "telephony": { "conditions": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "conditions_scored": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "wer": 0.081203007518797, "ref_words": 1330, "substitutions": 90, "insertions": 12, "deletions": 6, "si_sdr_db": 8.655543798473788, "pesq_wb": 1.8996521898678371, "stoi": 0.9051823174004429 }, "wideband": { "conditions": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "conditions_scored": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "wer": 0.18995215311004784, "ref_words": 2090, "substitutions": 254, "insertions": 89, "deletions": 54, "si_sdr_db": 12.362215768811991, "pesq_wb": 1.7653018582950937, "stoi": 0.8045638168316361 } }, "aic-quail": { "label": "ai-coustics Quail L", "native_sample_rate": 16000, "resampled": false, "telephony": { "conditions": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "conditions_scored": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "wer": 0.09022556390977443, "ref_words": 1330, "substitutions": 96, "insertions": 15, "deletions": 9, "si_sdr_db": 10.674216522954808, "pesq_wb": 2.1670914922441757, "stoi": 0.9208374512050874 }, "wideband": { "conditions": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "conditions_scored": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "wer": 0.19712918660287082, "ref_words": 2090, "substitutions": 253, "insertions": 111, "deletions": 48, "si_sdr_db": 7.196309628640543, "pesq_wb": 1.9488284609534523, "stoi": 0.8503597635990818 } }, "aic-quail-vf": { "label": "ai-coustics Quail VF 2.2 L", "native_sample_rate": 16000, "resampled": false, "telephony": { "conditions": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "conditions_scored": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "wer": 0.09699248120300752, "ref_words": 1330, "substitutions": 101, "insertions": 12, "deletions": 16, "si_sdr_db": 8.399567157638993, "pesq_wb": 1.9862947515078953, "stoi": 0.9112984380547939 }, "wideband": { "conditions": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "conditions_scored": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "wer": 0.23444976076555024, "ref_words": 2090, "substitutions": 282, "insertions": 149, "deletions": 59, "si_sdr_db": 2.8696570260763044, "pesq_wb": 1.8282234278592198, "stoi": 0.8392710030747256 } }, "resample8k": { "label": "Resample-only 16\u21928\u219216 (control)", "native_sample_rate": 8000, "resampled": true, "telephony": { "conditions": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "conditions_scored": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "wer": 0.08796992481203007, "ref_words": 1330, "substitutions": 97, "insertions": 11, "deletions": 9, "si_sdr_db": 8.467738416237465, "pesq_wb": 1.813909135546003, "stoi": 0.9034790033594288 }, "wideband": { "conditions": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "conditions_scored": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "wer": 0.24880382775119617, "ref_words": 2090, "substitutions": 316, "insertions": 162, "deletions": 42, "si_sdr_db": -0.6577940972184374, "pesq_wb": 1.5992720062082464, "stoi": 0.7994534545786899 } }, "gtcrn": { "label": "GTCRN", "native_sample_rate": 16000, "resampled": false, "telephony": { "conditions": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "conditions_scored": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "wer": 0.1293233082706767, "ref_words": 1330, "substitutions": 137, "insertions": 18, "deletions": 17, "si_sdr_db": 11.014662725212181, "pesq_wb": 2.141714496271951, "stoi": 0.8978227266987687 }, "wideband": { "conditions": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "conditions_scored": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "wer": 0.24928229665071772, "ref_words": 2090, "substitutions": 306, "insertions": 192, "deletions": 23, "si_sdr_db": 7.366158324771667, "pesq_wb": 2.0586626161228527, "stoi": 0.8310238975448133 } }, "fastenhancer-l": { "label": "FastEnhancer-L", "native_sample_rate": 16000, "resampled": false, "telephony": { "conditions": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "conditions_scored": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "wer": 0.1255639097744361, "ref_words": 1330, "substitutions": 127, "insertions": 18, "deletions": 22, "si_sdr_db": 11.634377993246797, "pesq_wb": 2.1950592381613596, "stoi": 0.9117813328344465 }, "wideband": { "conditions": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "conditions_scored": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "wer": 0.2545454545454545, "ref_words": 2090, "substitutions": 306, "insertions": 141, "deletions": 85, "si_sdr_db": 9.939965821742252, "pesq_wb": 2.200491759993813, "stoi": 0.8114073529903998 } }, "fastenhancer-s": { "label": "FastEnhancer-S", "native_sample_rate": 16000, "resampled": false, "telephony": { "conditions": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "conditions_scored": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "wer": 0.12857142857142856, "ref_words": 1330, "substitutions": 133, "insertions": 17, "deletions": 21, "si_sdr_db": 11.13346733676073, "pesq_wb": 2.2201851487159727, "stoi": 0.903468812199643 }, "wideband": { "conditions": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "conditions_scored": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "wer": 0.25598086124401914, "ref_words": 2090, "substitutions": 342, "insertions": 142, "deletions": 51, "si_sdr_db": 9.661294024005429, "pesq_wb": 2.1961383028463883, "stoi": 0.8223451218997422 } }, "fastenhancer-b": { "label": "FastEnhancer-B", "native_sample_rate": 16000, "resampled": false, "telephony": { "conditions": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "conditions_scored": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "wer": 0.13007518796992482, "ref_words": 1330, "substitutions": 139, "insertions": 15, "deletions": 19, "si_sdr_db": 10.93899863835989, "pesq_wb": 2.168459335395268, "stoi": 0.8935970562689898 }, "wideband": { "conditions": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "conditions_scored": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "wer": 0.2631578947368421, "ref_words": 2090, "substitutions": 346, "insertions": 144, "deletions": 60, "si_sdr_db": 9.141173431297462, "pesq_wb": 2.160575257648121, "stoi": 0.8187666176626394 } }, "fastenhancer-m": { "label": "FastEnhancer-M", "native_sample_rate": 16000, "resampled": false, "telephony": { "conditions": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "conditions_scored": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "wer": 0.1924812030075188, "ref_words": 1330, "substitutions": 149, "insertions": 76, "deletions": 31, "si_sdr_db": 11.026325934588792, "pesq_wb": 2.0760015572820394, "stoi": 0.8946311231771548 }, "wideband": { "conditions": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "conditions_scored": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "wer": 0.24401913875598086, "ref_words": 2090, "substitutions": 317, "insertions": 141, "deletions": 52, "si_sdr_db": 10.382569898032829, "pesq_wb": 2.1492373607375406, "stoi": 0.8195938099296995 } }, "fastenhancer-t": { "label": "FastEnhancer-T", "native_sample_rate": 16000, "resampled": false, "telephony": { "conditions": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "conditions_scored": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "wer": 0.17819548872180452, "ref_words": 1330, "substitutions": 195, "insertions": 16, "deletions": 26, "si_sdr_db": 10.425036876583304, "pesq_wb": 1.963403846536364, "stoi": 0.8575114744837032 }, "wideband": { "conditions": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "conditions_scored": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "wer": 0.254066985645933, "ref_words": 2090, "substitutions": 319, "insertions": 129, "deletions": 83, "si_sdr_db": 8.359544554109577, "pesq_wb": 2.0836882905526597, "stoi": 0.8203849471121786 } }, "dfn3-hq": { "label": "DeepFilterNet3 (96-frame warm-up)", "native_sample_rate": 48000, "resampled": true, "telephony": { "conditions": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "conditions_scored": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "wer": 0.17142857142857143, "ref_words": 1330, "substitutions": 161, "insertions": 17, "deletions": 50, "si_sdr_db": 9.413331726351458, "pesq_wb": 1.9404954705919537, "stoi": 0.8693253247152785 }, "wideband": { "conditions": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "conditions_scored": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "wer": 0.40861244019138754, "ref_words": 2090, "substitutions": 424, "insertions": 351, "deletions": 79, "si_sdr_db": 7.620646343063509, "pesq_wb": 1.9720037547024811, "stoi": 0.8124197443112036 } }, "dfn3": { "label": "DeepFilterNet3", "native_sample_rate": 48000, "resampled": true, "telephony": { "conditions": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "conditions_scored": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "wer": 0.18045112781954886, "ref_words": 1330, "substitutions": 178, "insertions": 19, "deletions": 43, "si_sdr_db": 8.902419913319148, "pesq_wb": 1.864194963659559, "stoi": 0.858951094083601 }, "wideband": { "conditions": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "conditions_scored": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "wer": 0.46076555023923443, "ref_words": 2090, "substitutions": 448, "insertions": 361, "deletions": 154, "si_sdr_db": 6.741030967428517, "pesq_wb": 1.8725676525722852, "stoi": 0.7896259560280428 } }, "clearline": { "label": "Clearline (8 kHz, resampled)", "native_sample_rate": 8000, "resampled": true, "telephony": { "conditions": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "conditions_scored": [ "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "wer": 0.1481203007518797, "ref_words": 1330, "substitutions": 156, "insertions": 22, "deletions": 19, "si_sdr_db": 6.57047676918576, "pesq_wb": 1.8068816508565628, "stoi": 0.8375486820014096 }, "wideband": { "conditions": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "conditions_scored": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10" ], "wer": 0.5545454545454546, "ref_words": 2090, "substitutions": 620, "insertions": 388, "deletions": 151, "si_sdr_db": -3.049560076742013, "pesq_wb": 1.6067323749715632, "stoi": 0.6842641804746564 } } }, "backendProvenance": { "clearline": { "run": "t4_300h_200ep_solo", "checkpoint": "/Users/aivanouski/home/hear.ai/engine/runs/t4_300h_200ep_solo/last.pt", "checkpoint_sha256": "a564bedd3dc4fcb43c52570652ac09e2d2422d1b9e204ec3e856400084e179a8", "model_py_sha256": "8dfcce203c4c831211357eeeadca8474343817c66a3425901563ba085b9b9541", "epoch": 199, "params": 968627, "config": { "sample_rate": 8000, "n_fft": 256, "hop": 128, "win": 256, "synth_win": 0, "channels": 48, "freq_stride": 4, "gru_hidden": 96, "gru_groups": 4, "n_gru_layers": 2, "n_blocks": 3, "embed_dim": 128, "embed_channels": 48, "warmup_frames": 46, "n_speakers": 0, "conditioning": "film", "separator": "gru", "n_bands": 16, "band_dim": 64, "n_bs_blocks": 4, "interference_gate": false, "gate_floor": 0.0, "mask_type": "real", "df_order": 3, "df_bins": 64 }, "streaming_parity": { "compared_samples": 41600, "max_abs_err": 2.5331974029541016e-07, "si_sdr_db": 132.42, "ref_peak": 0.4510730504989624, "excluded": "the warm-up window (streamed dry) and the final frame (offline istft truncates the tail; the streamer does not)" }, "front_end": "StopgapConfig.off() \u2014 no level normalisation, no warm-up gate", "warmup_ms": 736.0, "resample_path": "16 kHz -> 8 kHz -> model -> 8 kHz -> 16 kHz (soxr, harness-side)" } }, "sttDeterminism": { "schema": 1, "generated_at": "2026-08-16T17:37:24Z", "elapsed_s": 64.4, "host": { "platform": "macOS-26.5.2-arm64-arm-64bit", "machine": "arm64" }, "stt": { "id": "faster-whisper:base.en", "detail": "CTranslate2 int8 CPU, greedy, model=base.en" }, "arms": [ { "backend": "raw", "conditions": [ "clean", "noise_snr-5", "noise_snr0", "noise_snr5", "noise_snr10", "noise_snr20", "babble_snr5", "competing_snr0", "competing_snr5", "reverb", "reverb_snr10", "telephony", "telephony_snr10", "telephony_alaw", "telephony_g722", "telephony_opus12k", "telephony_loss3", "telephony_loss10" ], "clips": 180, "runs": 3 }, { "backend": "clearline", "conditions": [ "babble_snr5", "competing_snr0", "competing_snr5", "reverb_snr10", "noise_snr-5" ], "clips": 50, "runs": 3 } ], "backend": "clearline", "runs": 3, "conditions": [ "babble_snr5", "competing_snr0", "competing_snr5", "reverb_snr10", "noise_snr-5" ], "clips": 50, "scopes_identical": 25, "scopes_total": 25, "worst": { "scope": "raw:clean", "clips": 10, "ref_words": 190, "wers": [ 0.04736842105263158, 0.04736842105263158, 0.04736842105263158 ], "range": 0.0, "mean": 0.04736842105263158 }, "resolution_pp": 0.53, "note": "Same process, same audio arrays, decoded repeatedly. Any spread is the decoder, not the audio and not the enhancer (which, when used, ran once before the repeats).", "per_scope": { "raw:clean": { "scope": "raw:clean", "audio": "raw", "clips": 10, "ref_words": 190, "wers": [ 0.04736842105263158, 0.04736842105263158, 0.04736842105263158 ], "range": 0.0, "mean": 0.04736842105263158, "identical": true, "per_run": [ { "wer": 0.04736842105263158, "substitutions": 6, "insertions": 2, "deletions": 1, "hits": 183, "ref_words": 190, "hyp_words": 191, "sub_rate": 0.031578947368421054, "ins_rate": 0.010526315789473684, "del_rate": 0.005263157894736842 }, { "wer": 0.04736842105263158, "substitutions": 6, "insertions": 2, "deletions": 1, "hits": 183, "ref_words": 190, "hyp_words": 191, "sub_rate": 0.031578947368421054, "ins_rate": 0.010526315789473684, "del_rate": 0.005263157894736842 }, { "wer": 0.04736842105263158, "substitutions": 6, "insertions": 2, "deletions": 1, "hits": 183, "ref_words": 190, "hyp_words": 191, "sub_rate": 0.031578947368421054, "ins_rate": 0.010526315789473684, "del_rate": 0.005263157894736842 } ] }, "raw:noise_snr-5": { "scope": "raw:noise_snr-5", "audio": "raw", "clips": 10, "ref_words": 190, "wers": [ 0.18421052631578946, 0.18421052631578946, 0.18421052631578946 ], "range": 0.0, "mean": 0.18421052631578946, "identical": true, "per_run": [ { "wer": 0.18421052631578946, "substitutions": 29, "insertions": 4, "deletions": 2, "hits": 159, "ref_words": 190, "hyp_words": 192, "sub_rate": 0.15263157894736842, "ins_rate": 0.021052631578947368, "del_rate": 0.010526315789473684 }, { "wer": 0.18421052631578946, "substitutions": 29, "insertions": 4, "deletions": 2, "hits": 159, "ref_words": 190, "hyp_words": 192, "sub_rate": 0.15263157894736842, "ins_rate": 0.021052631578947368, "del_rate": 0.010526315789473684 }, { "wer": 0.18421052631578946, "substitutions": 29, "insertions": 4, "deletions": 2, "hits": 159, "ref_words": 190, "hyp_words": 192, "sub_rate": 0.15263157894736842, "ins_rate": 0.021052631578947368, "del_rate": 0.010526315789473684 } ] }, "raw:noise_snr0": { "scope": "raw:noise_snr0", "audio": "raw", "clips": 10, "ref_words": 190, "wers": [ 0.11052631578947368, 0.11052631578947368, 0.11052631578947368 ], "range": 0.0, "mean": 0.11052631578947368, "identical": true, "per_run": [ { "wer": 0.11052631578947368, "substitutions": 18, "insertions": 2, "deletions": 1, "hits": 171, "ref_words": 190, "hyp_words": 191, "sub_rate": 0.09473684210526316, "ins_rate": 0.010526315789473684, "del_rate": 0.005263157894736842 }, { "wer": 0.11052631578947368, "substitutions": 18, "insertions": 2, "deletions": 1, "hits": 171, "ref_words": 190, "hyp_words": 191, "sub_rate": 0.09473684210526316, "ins_rate": 0.010526315789473684, "del_rate": 0.005263157894736842 }, { "wer": 0.11052631578947368, "substitutions": 18, "insertions": 2, "deletions": 1, "hits": 171, "ref_words": 190, "hyp_words": 191, "sub_rate": 0.09473684210526316, "ins_rate": 0.010526315789473684, "del_rate": 0.005263157894736842 } ] }, "raw:noise_snr5": { "scope": "raw:noise_snr5", "audio": "raw", "clips": 10, "ref_words": 190, "wers": [ 0.08947368421052632, 0.08947368421052632, 0.08947368421052632 ], "range": 0.0, "mean": 0.08947368421052632, "identical": true, "per_run": [ { "wer": 0.08947368421052632, "substitutions": 15, "insertions": 2, "deletions": 0, "hits": 175, "ref_words": 190, "hyp_words": 192, "sub_rate": 0.07894736842105263, "ins_rate": 0.010526315789473684, "del_rate": 0.0 }, { "wer": 0.08947368421052632, "substitutions": 15, "insertions": 2, "deletions": 0, "hits": 175, "ref_words": 190, "hyp_words": 192, "sub_rate": 0.07894736842105263, "ins_rate": 0.010526315789473684, "del_rate": 0.0 }, { "wer": 0.08947368421052632, "substitutions": 15, "insertions": 2, "deletions": 0, "hits": 175, "ref_words": 190, "hyp_words": 192, "sub_rate": 0.07894736842105263, "ins_rate": 0.010526315789473684, "del_rate": 0.0 } ] }, "raw:noise_snr10": { "scope": "raw:noise_snr10", "audio": "raw", "clips": 10, "ref_words": 190, "wers": [ 0.06315789473684211, 0.06315789473684211, 0.06315789473684211 ], "range": 0.0, "mean": 0.06315789473684211, "identical": true, "per_run": [ { "wer": 0.06315789473684211, "substitutions": 10, "insertions": 2, "deletions": 0, "hits": 180, "ref_words": 190, "hyp_words": 192, "sub_rate": 0.05263157894736842, "ins_rate": 0.010526315789473684, "del_rate": 0.0 }, { "wer": 0.06315789473684211, "substitutions": 10, "insertions": 2, "deletions": 0, "hits": 180, "ref_words": 190, "hyp_words": 192, "sub_rate": 0.05263157894736842, "ins_rate": 0.010526315789473684, "del_rate": 0.0 }, { "wer": 0.06315789473684211, "substitutions": 10, "insertions": 2, "deletions": 0, "hits": 180, "ref_words": 190, "hyp_words": 192, "sub_rate": 0.05263157894736842, "ins_rate": 0.010526315789473684, "del_rate": 0.0 } ] }, "raw:noise_snr20": { "scope": "raw:noise_snr20", "audio": "raw", "clips": 10, "ref_words": 190, "wers": [ 0.042105263157894736, 0.042105263157894736, 0.042105263157894736 ], "range": 0.0, "mean": 0.042105263157894736, "identical": true, "per_run": [ { "wer": 0.042105263157894736, "substitutions": 6, "insertions": 2, "deletions": 0, "hits": 184, "ref_words": 190, "hyp_words": 192, "sub_rate": 0.031578947368421054, "ins_rate": 0.010526315789473684, "del_rate": 0.0 }, { "wer": 0.042105263157894736, "substitutions": 6, "insertions": 2, "deletions": 0, "hits": 184, "ref_words": 190, "hyp_words": 192, "sub_rate": 0.031578947368421054, "ins_rate": 0.010526315789473684, "del_rate": 0.0 }, { "wer": 0.042105263157894736, "substitutions": 6, "insertions": 2, "deletions": 0, "hits": 184, "ref_words": 190, "hyp_words": 192, "sub_rate": 0.031578947368421054, "ins_rate": 0.010526315789473684, "del_rate": 0.0 } ] }, "raw:babble_snr5": { "scope": "raw:babble_snr5", "audio": "raw", "clips": 10, "ref_words": 190, "wers": [ 0.24736842105263157, 0.24736842105263157, 0.24736842105263157 ], "range": 0.0, "mean": 0.24736842105263157, "identical": true, "per_run": [ { "wer": 0.24736842105263157, "substitutions": 27, "insertions": 10, "deletions": 10, "hits": 153, "ref_words": 190, "hyp_words": 190, "sub_rate": 0.14210526315789473, "ins_rate": 0.05263157894736842, "del_rate": 0.05263157894736842 }, { "wer": 0.24736842105263157, "substitutions": 27, "insertions": 10, "deletions": 10, "hits": 153, "ref_words": 190, "hyp_words": 190, "sub_rate": 0.14210526315789473, "ins_rate": 0.05263157894736842, "del_rate": 0.05263157894736842 }, { "wer": 0.24736842105263157, "substitutions": 27, "insertions": 10, "deletions": 10, "hits": 153, "ref_words": 190, "hyp_words": 190, "sub_rate": 0.14210526315789473, "ins_rate": 0.05263157894736842, "del_rate": 0.05263157894736842 } ] }, "raw:competing_snr0": { "scope": "raw:competing_snr0", "audio": "raw", "clips": 10, "ref_words": 190, "wers": [ 0.8473684210526315, 0.8473684210526315, 0.8473684210526315 ], "range": 0.0, "mean": 0.8473684210526317, "identical": true, "per_run": [ { "wer": 0.8473684210526315, "substitutions": 81, "insertions": 45, "deletions": 35, "hits": 74, "ref_words": 190, "hyp_words": 200, "sub_rate": 0.4263157894736842, "ins_rate": 0.23684210526315788, "del_rate": 0.18421052631578946 }, { "wer": 0.8473684210526315, "substitutions": 81, "insertions": 45, "deletions": 35, "hits": 74, "ref_words": 190, "hyp_words": 200, "sub_rate": 0.4263157894736842, "ins_rate": 0.23684210526315788, "del_rate": 0.18421052631578946 }, { "wer": 0.8473684210526315, "substitutions": 81, "insertions": 45, "deletions": 35, "hits": 74, "ref_words": 190, "hyp_words": 200, "sub_rate": 0.4263157894736842, "ins_rate": 0.23684210526315788, "del_rate": 0.18421052631578946 } ] }, "raw:competing_snr5": { "scope": "raw:competing_snr5", "audio": "raw", "clips": 10, "ref_words": 190, "wers": [ 0.2631578947368421, 0.2631578947368421, 0.2631578947368421 ], "range": 0.0, "mean": 0.2631578947368421, "identical": true, "per_run": [ { "wer": 0.2631578947368421, "substitutions": 28, "insertions": 19, "deletions": 3, "hits": 159, "ref_words": 190, "hyp_words": 206, "sub_rate": 0.14736842105263157, "ins_rate": 0.1, "del_rate": 0.015789473684210527 }, { "wer": 0.2631578947368421, "substitutions": 28, "insertions": 19, "deletions": 3, "hits": 159, "ref_words": 190, "hyp_words": 206, "sub_rate": 0.14736842105263157, "ins_rate": 0.1, "del_rate": 0.015789473684210527 }, { "wer": 0.2631578947368421, "substitutions": 28, "insertions": 19, "deletions": 3, "hits": 159, "ref_words": 190, "hyp_words": 206, "sub_rate": 0.14736842105263157, "ins_rate": 0.1, "del_rate": 0.015789473684210527 } ] }, "raw:reverb": { "scope": "raw:reverb", "audio": "raw", "clips": 10, "ref_words": 190, "wers": [ 0.05789473684210526, 0.05789473684210526, 0.05789473684210526 ], "range": 0.0, "mean": 0.057894736842105256, "identical": true, "per_run": [ { "wer": 0.05789473684210526, "substitutions": 10, "insertions": 1, "deletions": 0, "hits": 180, "ref_words": 190, "hyp_words": 191, "sub_rate": 0.05263157894736842, "ins_rate": 0.005263157894736842, "del_rate": 0.0 }, { "wer": 0.05789473684210526, "substitutions": 10, "insertions": 1, "deletions": 0, "hits": 180, "ref_words": 190, "hyp_words": 191, "sub_rate": 0.05263157894736842, "ins_rate": 0.005263157894736842, "del_rate": 0.0 }, { "wer": 0.05789473684210526, "substitutions": 10, "insertions": 1, "deletions": 0, "hits": 180, "ref_words": 190, "hyp_words": 191, "sub_rate": 0.05263157894736842, "ins_rate": 0.005263157894736842, "del_rate": 0.0 } ] }, "raw:reverb_snr10": { "scope": "raw:reverb_snr10", "audio": "raw", "clips": 10, "ref_words": 190, "wers": [ 0.1368421052631579, 0.1368421052631579, 0.1368421052631579 ], "range": 0.0, "mean": 0.1368421052631579, "identical": true, "per_run": [ { "wer": 0.1368421052631579, "substitutions": 24, "insertions": 0, "deletions": 2, "hits": 164, "ref_words": 190, "hyp_words": 188, "sub_rate": 0.12631578947368421, "ins_rate": 0.0, "del_rate": 0.010526315789473684 }, { "wer": 0.1368421052631579, "substitutions": 24, "insertions": 0, "deletions": 2, "hits": 164, "ref_words": 190, "hyp_words": 188, "sub_rate": 0.12631578947368421, "ins_rate": 0.0, "del_rate": 0.010526315789473684 }, { "wer": 0.1368421052631579, "substitutions": 24, "insertions": 0, "deletions": 2, "hits": 164, "ref_words": 190, "hyp_words": 188, "sub_rate": 0.12631578947368421, "ins_rate": 0.0, "del_rate": 0.010526315789473684 } ] }, "raw:telephony": { "scope": "raw:telephony", "audio": "raw", "clips": 10, "ref_words": 190, "wers": [ 0.05789473684210526, 0.05789473684210526, 0.05789473684210526 ], "range": 0.0, "mean": 0.057894736842105256, "identical": true, "per_run": [ { "wer": 0.05789473684210526, "substitutions": 9, "insertions": 1, "deletions": 1, "hits": 180, "ref_words": 190, "hyp_words": 190, "sub_rate": 0.04736842105263158, "ins_rate": 0.005263157894736842, "del_rate": 0.005263157894736842 }, { "wer": 0.05789473684210526, "substitutions": 9, "insertions": 1, "deletions": 1, "hits": 180, "ref_words": 190, "hyp_words": 190, "sub_rate": 0.04736842105263158, "ins_rate": 0.005263157894736842, "del_rate": 0.005263157894736842 }, { "wer": 0.05789473684210526, "substitutions": 9, "insertions": 1, "deletions": 1, "hits": 180, "ref_words": 190, "hyp_words": 190, "sub_rate": 0.04736842105263158, "ins_rate": 0.005263157894736842, "del_rate": 0.005263157894736842 } ] }, "raw:telephony_snr10": { "scope": "raw:telephony_snr10", "audio": "raw", "clips": 10, "ref_words": 190, "wers": [ 0.10526315789473684, 0.10526315789473684, 0.10526315789473684 ], "range": 0.0, "mean": 0.10526315789473684, "identical": true, "per_run": [ { "wer": 0.10526315789473684, "substitutions": 16, "insertions": 3, "deletions": 1, "hits": 173, "ref_words": 190, "hyp_words": 192, "sub_rate": 0.08421052631578947, "ins_rate": 0.015789473684210527, "del_rate": 0.005263157894736842 }, { "wer": 0.10526315789473684, "substitutions": 16, "insertions": 3, "deletions": 1, "hits": 173, "ref_words": 190, "hyp_words": 192, "sub_rate": 0.08421052631578947, "ins_rate": 0.015789473684210527, "del_rate": 0.005263157894736842 }, { "wer": 0.10526315789473684, "substitutions": 16, "insertions": 3, "deletions": 1, "hits": 173, "ref_words": 190, "hyp_words": 192, "sub_rate": 0.08421052631578947, "ins_rate": 0.015789473684210527, "del_rate": 0.005263157894736842 } ] }, "raw:telephony_alaw": { "scope": "raw:telephony_alaw", "audio": "raw", "clips": 10, "ref_words": 190, "wers": [ 0.07368421052631578, 0.07368421052631578, 0.07368421052631578 ], "range": 0.0, "mean": 0.07368421052631578, "identical": true, "per_run": [ { "wer": 0.07368421052631578, "substitutions": 12, "insertions": 2, "deletions": 0, "hits": 178, "ref_words": 190, "hyp_words": 192, "sub_rate": 0.06315789473684211, "ins_rate": 0.010526315789473684, "del_rate": 0.0 }, { "wer": 0.07368421052631578, "substitutions": 12, "insertions": 2, "deletions": 0, "hits": 178, "ref_words": 190, "hyp_words": 192, "sub_rate": 0.06315789473684211, "ins_rate": 0.010526315789473684, "del_rate": 0.0 }, { "wer": 0.07368421052631578, "substitutions": 12, "insertions": 2, "deletions": 0, "hits": 178, "ref_words": 190, "hyp_words": 192, "sub_rate": 0.06315789473684211, "ins_rate": 0.010526315789473684, "del_rate": 0.0 } ] }, "raw:telephony_g722": { "scope": "raw:telephony_g722", "audio": "raw", "clips": 10, "ref_words": 190, "wers": [ 0.05263157894736842, 0.05263157894736842, 0.05263157894736842 ], "range": 0.0, "mean": 0.05263157894736842, "identical": true, "per_run": [ { "wer": 0.05263157894736842, "substitutions": 9, "insertions": 1, "deletions": 0, "hits": 181, "ref_words": 190, "hyp_words": 191, "sub_rate": 0.04736842105263158, "ins_rate": 0.005263157894736842, "del_rate": 0.0 }, { "wer": 0.05263157894736842, "substitutions": 9, "insertions": 1, "deletions": 0, "hits": 181, "ref_words": 190, "hyp_words": 191, "sub_rate": 0.04736842105263158, "ins_rate": 0.005263157894736842, "del_rate": 0.0 }, { "wer": 0.05263157894736842, "substitutions": 9, "insertions": 1, "deletions": 0, "hits": 181, "ref_words": 190, "hyp_words": 191, "sub_rate": 0.04736842105263158, "ins_rate": 0.005263157894736842, "del_rate": 0.0 } ] }, "raw:telephony_opus12k": { "scope": "raw:telephony_opus12k", "audio": "raw", "clips": 10, "ref_words": 190, "wers": [ 0.06842105263157895, 0.06842105263157895, 0.06842105263157895 ], "range": 0.0, "mean": 0.06842105263157895, "identical": true, "per_run": [ { "wer": 0.06842105263157895, "substitutions": 9, "insertions": 2, "deletions": 2, "hits": 179, "ref_words": 190, "hyp_words": 190, "sub_rate": 0.04736842105263158, "ins_rate": 0.010526315789473684, "del_rate": 0.010526315789473684 }, { "wer": 0.06842105263157895, "substitutions": 9, "insertions": 2, "deletions": 2, "hits": 179, "ref_words": 190, "hyp_words": 190, "sub_rate": 0.04736842105263158, "ins_rate": 0.010526315789473684, "del_rate": 0.010526315789473684 }, { "wer": 0.06842105263157895, "substitutions": 9, "insertions": 2, "deletions": 2, "hits": 179, "ref_words": 190, "hyp_words": 190, "sub_rate": 0.04736842105263158, "ins_rate": 0.010526315789473684, "del_rate": 0.010526315789473684 } ] }, "raw:telephony_loss3": { "scope": "raw:telephony_loss3", "audio": "raw", "clips": 10, "ref_words": 190, "wers": [ 0.11052631578947368, 0.11052631578947368, 0.11052631578947368 ], "range": 0.0, "mean": 0.11052631578947368, "identical": true, "per_run": [ { "wer": 0.11052631578947368, "substitutions": 18, "insertions": 2, "deletions": 1, "hits": 171, "ref_words": 190, "hyp_words": 191, "sub_rate": 0.09473684210526316, "ins_rate": 0.010526315789473684, "del_rate": 0.005263157894736842 }, { "wer": 0.11052631578947368, "substitutions": 18, "insertions": 2, "deletions": 1, "hits": 171, "ref_words": 190, "hyp_words": 191, "sub_rate": 0.09473684210526316, "ins_rate": 0.010526315789473684, "del_rate": 0.005263157894736842 }, { "wer": 0.11052631578947368, "substitutions": 18, "insertions": 2, "deletions": 1, "hits": 171, "ref_words": 190, "hyp_words": 191, "sub_rate": 0.09473684210526316, "ins_rate": 0.010526315789473684, "del_rate": 0.005263157894736842 } ] }, "raw:telephony_loss10": { "scope": "raw:telephony_loss10", "audio": "raw", "clips": 10, "ref_words": 190, "wers": [ 0.1, 0.1, 0.1 ], "range": 0.0, "mean": 0.10000000000000002, "identical": true, "per_run": [ { "wer": 0.1, "substitutions": 17, "insertions": 1, "deletions": 1, "hits": 172, "ref_words": 190, "hyp_words": 190, "sub_rate": 0.08947368421052632, "ins_rate": 0.005263157894736842, "del_rate": 0.005263157894736842 }, { "wer": 0.1, "substitutions": 17, "insertions": 1, "deletions": 1, "hits": 172, "ref_words": 190, "hyp_words": 190, "sub_rate": 0.08947368421052632, "ins_rate": 0.005263157894736842, "del_rate": 0.005263157894736842 }, { "wer": 0.1, "substitutions": 17, "insertions": 1, "deletions": 1, "hits": 172, "ref_words": 190, "hyp_words": 190, "sub_rate": 0.08947368421052632, "ins_rate": 0.005263157894736842, "del_rate": 0.005263157894736842 } ] }, "raw:pooled (18 conditions)": { "scope": "raw:pooled (18 conditions)", "audio": "raw", "clips": 180, "ref_words": 3420, "wers": [ 0.1476608187134503, 0.1476608187134503, 0.1476608187134503 ], "range": 0.0, "mean": 0.1476608187134503, "identical": true, "per_run": [ { "wer": 0.1476608187134503, "substitutions": 344, "insertions": 101, "deletions": 60, "hits": 3016, "ref_words": 3420, "hyp_words": 3461, "sub_rate": 0.10058479532163743, "ins_rate": 0.02953216374269006, "del_rate": 0.017543859649122806 }, { "wer": 0.1476608187134503, "substitutions": 344, "insertions": 101, "deletions": 60, "hits": 3016, "ref_words": 3420, "hyp_words": 3461, "sub_rate": 0.10058479532163743, "ins_rate": 0.02953216374269006, "del_rate": 0.017543859649122806 }, { "wer": 0.1476608187134503, "substitutions": 344, "insertions": 101, "deletions": 60, "hits": 3016, "ref_words": 3420, "hyp_words": 3461, "sub_rate": 0.10058479532163743, "ins_rate": 0.02953216374269006, "del_rate": 0.017543859649122806 } ] }, "clearline:babble_snr5": { "scope": "clearline:babble_snr5", "audio": "clearline", "clips": 10, "ref_words": 190, "wers": [ 1.3842105263157896, 1.3842105263157896, 1.3842105263157896 ], "range": 0.0, "mean": 1.3842105263157896, "identical": true, "per_run": [ { "wer": 1.3842105263157896, "substitutions": 67, "insertions": 169, "deletions": 27, "hits": 96, "ref_words": 190, "hyp_words": 332, "sub_rate": 0.3526315789473684, "ins_rate": 0.8894736842105263, "del_rate": 0.14210526315789473 }, { "wer": 1.3842105263157896, "substitutions": 67, "insertions": 169, "deletions": 27, "hits": 96, "ref_words": 190, "hyp_words": 332, "sub_rate": 0.3526315789473684, "ins_rate": 0.8894736842105263, "del_rate": 0.14210526315789473 }, { "wer": 1.3842105263157896, "substitutions": 67, "insertions": 169, "deletions": 27, "hits": 96, "ref_words": 190, "hyp_words": 332, "sub_rate": 0.3526315789473684, "ins_rate": 0.8894736842105263, "del_rate": 0.14210526315789473 } ] }, "clearline:competing_snr0": { "scope": "clearline:competing_snr0", "audio": "clearline", "clips": 10, "ref_words": 190, "wers": [ 1.068421052631579, 1.068421052631579, 1.068421052631579 ], "range": 0.0, "mean": 1.068421052631579, "identical": true, "per_run": [ { "wer": 1.068421052631579, "substitutions": 144, "insertions": 36, "deletions": 23, "hits": 23, "ref_words": 190, "hyp_words": 203, "sub_rate": 0.7578947368421053, "ins_rate": 0.18947368421052632, "del_rate": 0.12105263157894737 }, { "wer": 1.068421052631579, "substitutions": 144, "insertions": 36, "deletions": 23, "hits": 23, "ref_words": 190, "hyp_words": 203, "sub_rate": 0.7578947368421053, "ins_rate": 0.18947368421052632, "del_rate": 0.12105263157894737 }, { "wer": 1.068421052631579, "substitutions": 144, "insertions": 36, "deletions": 23, "hits": 23, "ref_words": 190, "hyp_words": 203, "sub_rate": 0.7578947368421053, "ins_rate": 0.18947368421052632, "del_rate": 0.12105263157894737 } ] }, "clearline:competing_snr5": { "scope": "clearline:competing_snr5", "audio": "clearline", "clips": 10, "ref_words": 190, "wers": [ 0.8736842105263158, 0.8736842105263158, 0.8736842105263158 ], "range": 0.0, "mean": 0.8736842105263158, "identical": true, "per_run": [ { "wer": 0.8736842105263158, "substitutions": 120, "insertions": 29, "deletions": 17, "hits": 53, "ref_words": 190, "hyp_words": 202, "sub_rate": 0.631578947368421, "ins_rate": 0.15263157894736842, "del_rate": 0.08947368421052632 }, { "wer": 0.8736842105263158, "substitutions": 120, "insertions": 29, "deletions": 17, "hits": 53, "ref_words": 190, "hyp_words": 202, "sub_rate": 0.631578947368421, "ins_rate": 0.15263157894736842, "del_rate": 0.08947368421052632 }, { "wer": 0.8736842105263158, "substitutions": 120, "insertions": 29, "deletions": 17, "hits": 53, "ref_words": 190, "hyp_words": 202, "sub_rate": 0.631578947368421, "ins_rate": 0.15263157894736842, "del_rate": 0.08947368421052632 } ] }, "clearline:reverb_snr10": { "scope": "clearline:reverb_snr10", "audio": "clearline", "clips": 10, "ref_words": 190, "wers": [ 1.1631578947368422, 1.1631578947368422, 1.1631578947368422 ], "range": 0.0, "mean": 1.1631578947368422, "identical": true, "per_run": [ { "wer": 1.1631578947368422, "substitutions": 75, "insertions": 117, "deletions": 29, "hits": 86, "ref_words": 190, "hyp_words": 278, "sub_rate": 0.39473684210526316, "ins_rate": 0.6157894736842106, "del_rate": 0.15263157894736842 }, { "wer": 1.1631578947368422, "substitutions": 75, "insertions": 117, "deletions": 29, "hits": 86, "ref_words": 190, "hyp_words": 278, "sub_rate": 0.39473684210526316, "ins_rate": 0.6157894736842106, "del_rate": 0.15263157894736842 }, { "wer": 1.1631578947368422, "substitutions": 75, "insertions": 117, "deletions": 29, "hits": 86, "ref_words": 190, "hyp_words": 278, "sub_rate": 0.39473684210526316, "ins_rate": 0.6157894736842106, "del_rate": 0.15263157894736842 } ] }, "clearline:noise_snr-5": { "scope": "clearline:noise_snr-5", "audio": "clearline", "clips": 10, "ref_words": 190, "wers": [ 0.6368421052631579, 0.6368421052631579, 0.6368421052631579 ], "range": 0.0, "mean": 0.6368421052631579, "identical": true, "per_run": [ { "wer": 0.6368421052631579, "substitutions": 82, "insertions": 9, "deletions": 30, "hits": 78, "ref_words": 190, "hyp_words": 169, "sub_rate": 0.43157894736842106, "ins_rate": 0.04736842105263158, "del_rate": 0.15789473684210525 }, { "wer": 0.6368421052631579, "substitutions": 82, "insertions": 9, "deletions": 30, "hits": 78, "ref_words": 190, "hyp_words": 169, "sub_rate": 0.43157894736842106, "ins_rate": 0.04736842105263158, "del_rate": 0.15789473684210525 }, { "wer": 0.6368421052631579, "substitutions": 82, "insertions": 9, "deletions": 30, "hits": 78, "ref_words": 190, "hyp_words": 169, "sub_rate": 0.43157894736842106, "ins_rate": 0.04736842105263158, "del_rate": 0.15789473684210525 } ] }, "clearline:pooled (5 conditions)": { "scope": "clearline:pooled (5 conditions)", "audio": "clearline", "clips": 50, "ref_words": 950, "wers": [ 1.0252631578947369, 1.0252631578947369, 1.0252631578947369 ], "range": 0.0, "mean": 1.0252631578947369, "identical": true, "per_run": [ { "wer": 1.0252631578947369, "substitutions": 488, "insertions": 360, "deletions": 126, "hits": 336, "ref_words": 950, "hyp_words": 1184, "sub_rate": 0.5136842105263157, "ins_rate": 0.37894736842105264, "del_rate": 0.13263157894736843 }, { "wer": 1.0252631578947369, "substitutions": 488, "insertions": 360, "deletions": 126, "hits": 336, "ref_words": 950, "hyp_words": 1184, "sub_rate": 0.5136842105263157, "ins_rate": 0.37894736842105264, "del_rate": 0.13263157894736843 }, { "wer": 1.0252631578947369, "substitutions": 488, "insertions": 360, "deletions": 126, "hits": 336, "ref_words": 950, "hyp_words": 1184, "sub_rate": 0.5136842105263157, "ins_rate": 0.37894736842105264, "del_rate": 0.13263157894736843 } ] } }, "decoder_noise_pp": 0.0, "decoder_reproduced_exactly": true, "one_word_pp": { "scope": "raw:clean", "ref_words": 190, "pp": 0.526 }, "resolution_source": "one reference word" } }