tests/audio-preprocess-cases.json

[
 {
  "file": "audio/speech_en_train_16k.wav",
  "sample_rate": 16000,
  "channels": 1,
  "n_samples16k": 61749,
  "n_frames": 385,
  "n_tokens": 97,
  "bucket": 1024,
  "wave_file": "audio/ref/speech_en_train_16k.wave16k.f32",
  "features_file": "audio/ref/speech_en_train_16k.features.f32",
  "wave_tolerance": 0.0,
  "feature_tolerance": 0.02,
  "python_decoder": "libsndfile",
  "note": null
 },
 {
  "file": "audio/speech_en_parcel_8k.wav",
  "sample_rate": 8000,
  "channels": 1,
  "n_samples16k": 66558,
  "n_frames": 415,
  "n_tokens": 104,
  "bucket": 1024,
  "wave_file": "audio/ref/speech_en_parcel_8k.wave16k.f32",
  "features_file": "audio/ref/speech_en_parcel_8k.features.f32",
  "wave_tolerance": 1e-06,
  "feature_tolerance": 0.02,
  "python_decoder": "libsndfile",
  "note": null
 },
 {
  "file": "audio/speech_it_mercato_22k.aiff",
  "sample_rate": 22050,
  "channels": 1,
  "n_samples16k": 74743,
  "n_frames": 467,
  "n_tokens": 117,
  "bucket": 1024,
  "wave_file": "audio/ref/speech_it_mercato_22k.wave16k.f32",
  "features_file": "audio/ref/speech_it_mercato_22k.features.f32",
  "wave_tolerance": 1e-06,
  "feature_tolerance": 0.02,
  "python_decoder": "libsndfile",
  "note": null
 },
 {
  "file": "audio/speech_fr_pont_44k.m4a",
  "sample_rate": 44100,
  "channels": 1,
  "n_samples16k": 48751,
  "n_frames": 304,
  "n_tokens": 76,
  "bucket": 384,
  "wave_file": "audio/ref/speech_fr_pont_44k.wave16k.f32",
  "features_file": "audio/ref/speech_fr_pont_44k.features.f32",
  "wave_tolerance": 1e-06,
  "feature_tolerance": 0.02,
  "python_decoder": "afconvert",
  "note": null
 },
 {
  "file": "audio/speech_de_garten_48k_stereo.wav",
  "sample_rate": 48000,
  "channels": 2,
  "n_samples16k": 50307,
  "n_frames": 314,
  "n_tokens": 79,
  "bucket": 384,
  "wave_file": "audio/ref/speech_de_garten_48k_stereo.wave16k.f32",
  "features_file": "audio/ref/speech_de_garten_48k_stereo.features.f32",
  "wave_tolerance": 1e-06,
  "feature_tolerance": 0.02,
  "python_decoder": "libsndfile",
  "note": null
 },
 {
  "file": "audio/speech_pt_barco_44k.caf",
  "sample_rate": 44100,
  "channels": 1,
  "n_samples16k": 73981,
  "n_frames": 462,
  "n_tokens": 116,
  "bucket": 1024,
  "wave_file": "audio/ref/speech_pt_barco_44k.wave16k.f32",
  "features_file": "audio/ref/speech_pt_barco_44k.features.f32",
  "wave_tolerance": 1e-06,
  "feature_tolerance": 0.02,
  "python_decoder": "libsndfile",
  "note": null
 },
 {
  "file": "audio/speech_es_tienda_16k.flac",
  "sample_rate": 16000,
  "channels": 1,
  "n_samples16k": 58569,
  "n_frames": 366,
  "n_tokens": 92,
  "bucket": 384,
  "wave_file": "audio/ref/speech_es_tienda_16k.wave16k.f32",
  "features_file": "audio/ref/speech_es_tienda_16k.features.f32",
  "wave_tolerance": 0.0,
  "feature_tolerance": 0.02,
  "python_decoder": "libsndfile",
  "note": null
 },
 {
  "file": "audio/speech_sv_buss_22k.mp3",
  "sample_rate": 22050,
  "channels": 1,
  "n_samples16k": 68097,
  "n_frames": 425,
  "n_tokens": 107,
  "bucket": 1024,
  "wave_file": "audio/ref/speech_sv_buss_22k.wave16k.f32",
  "features_file": "audio/ref/speech_sv_buss_22k.features.f32",
  "wave_tolerance": 0.0001,
  "feature_tolerance": 1.0,
  "python_decoder": "libsndfile",
  "note": "mp3: libsndfile (Python) vs Apple (Swift) decoders differ by ~3e-5 per sample"
 },
 {
  "file": "audio/sound_chord_piano_44k_stereo.wav",
  "sample_rate": 44100,
  "channels": 2,
  "n_samples16k": 48000,
  "n_frames": 299,
  "n_tokens": 75,
  "bucket": 384,
  "wave_file": "audio/ref/sound_chord_piano_44k_stereo.wave16k.f32",
  "features_file": "audio/ref/sound_chord_piano_44k_stereo.features.f32",
  "wave_tolerance": 1e-06,
  "feature_tolerance": 0.02,
  "python_decoder": "libsndfile",
  "note": null
 },
 {
  "file": "audio/sound_clicks_noise_22k.wav",
  "sample_rate": 22050,
  "channels": 1,
  "n_samples16k": 64000,
  "n_frames": 399,
  "n_tokens": 100,
  "bucket": 1024,
  "wave_file": "audio/ref/sound_clicks_noise_22k.wave16k.f32",
  "features_file": "audio/ref/sound_clicks_noise_22k.features.f32",
  "wave_tolerance": 1e-06,
  "feature_tolerance": 0.02,
  "python_decoder": "libsndfile",
  "note": null
 },
 {
  "file": "audio/sound_chirp_siren_48k.flac",
  "sample_rate": 48000,
  "channels": 1,
  "n_samples16k": 80000,
  "n_frames": 499,
  "n_tokens": 125,
  "bucket": 1024,
  "wave_file": "audio/ref/sound_chirp_siren_48k.wave16k.f32",
  "features_file": "audio/ref/sound_chirp_siren_48k.features.f32",
  "wave_tolerance": 1e-06,
  "feature_tolerance": 0.02,
  "python_decoder": "libsndfile",
  "note": null
 },
 {
  "file": "audio/sound_birds_rain_16k_stereo.m4a",
  "sample_rate": 16000,
  "channels": 2,
  "n_samples16k": 80000,
  "n_frames": 499,
  "n_tokens": 125,
  "bucket": 1024,
  "wave_file": "audio/ref/sound_birds_rain_16k_stereo.wave16k.f32",
  "features_file": "audio/ref/sound_birds_rain_16k_stereo.features.f32",
  "wave_tolerance": 1e-06,
  "feature_tolerance": 0.02,
  "python_decoder": "afconvert",
  "note": null
 },
 {
  "file": "audio/edge_silence_1s_16k.wav",
  "sample_rate": 16000,
  "channels": 1,
  "n_samples16k": 16000,
  "n_frames": 99,
  "n_tokens": 25,
  "bucket": 192,
  "wave_file": "audio/ref/edge_silence_1s_16k.wave16k.f32",
  "features_file": "audio/ref/edge_silence_1s_16k.features.f32",
  "wave_tolerance": 0.0,
  "feature_tolerance": 0.02,
  "python_decoder": "libsndfile",
  "note": null
 },
 {
  "file": "audio/edge_tiny_100smp_16k.wav",
  "sample_rate": 16000,
  "channels": 1,
  "n_samples16k": 100,
  "n_frames": 0,
  "n_tokens": 0,
  "bucket": 192,
  "wave_file": "audio/ref/edge_tiny_100smp_16k.wave16k.f32",
  "features_file": "audio/ref/edge_tiny_100smp_16k.features.f32",
  "wave_tolerance": 0.0,
  "feature_tolerance": 0.02,
  "python_decoder": "libsndfile",
  "note": null
 },
 {
  "file": "audio/edge_tone_10khz_96k.wav",
  "sample_rate": 96000,
  "channels": 1,
  "n_samples16k": 32000,
  "n_frames": 199,
  "n_tokens": 50,
  "bucket": 384,
  "wave_file": "audio/ref/edge_tone_10khz_96k.wave16k.f32",
  "features_file": "audio/ref/edge_tone_10khz_96k.features.f32",
  "wave_tolerance": 1e-06,
  "feature_tolerance": 0.02,
  "python_decoder": "libsndfile",
  "note": null
 },
 {
  "file": "audio/long_melody_35s_16k.mp3",
  "sample_rate": 16000,
  "channels": 1,
  "n_samples16k": 560000,
  "n_frames": 2999,
  "n_tokens": 750,
  "bucket": 3000,
  "wave_file": "audio/ref/long_melody_35s_16k.wave16k.f32",
  "features_file": "audio/ref/long_melody_35s_16k.features.f32",
  "wave_tolerance": 0.0001,
  "feature_tolerance": 1.0,
  "python_decoder": "libsndfile",
  "note": "mp3: libsndfile (Python) vs Apple (Swift) decoders differ by ~3e-5 per sample"
 }
]