-
Notifications
You must be signed in to change notification settings - Fork 258
/
Copy pathaudio_preprocessing_tutorial.html
2594 lines (2264 loc) · 260 KB
/
audio_preprocessing_tutorial.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
<!DOCTYPE html>
<!--[if IE 8]><html class="no-js lt-ie9" lang="en" > <![endif]-->
<!--[if gt IE 8]><!--> <html class="no-js" lang="en" > <!--<![endif]-->
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Audio manipulation with torchaudio — PyTorch Tutorials 1.10.2+cu102 documentation</title>
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
<!-- <link rel="stylesheet" href="../_static/pygments.css" type="text/css" /> -->
<link rel="stylesheet" href="../_static/copybutton.css" type="text/css" />
<link rel="stylesheet" href="../_static/gallery.css" type="text/css" />
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/katex@0.10.0-beta/dist/katex.min.css" type="text/css" />
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/katex@0.13.11/dist/katex.min.css" type="text/css" />
<link rel="stylesheet" href="../_static/katex-math.css" type="text/css" />
<link rel="index" title="Index" href="../genindex.html" />
<link rel="search" title="Search" href="../search.html" />
<script src="../_static/js/modernizr.min.js"></script>
<!-- Preload the theme fonts -->
<link rel="preload" href="../_static/fonts/FreightSans/freight-sans-book.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="../_static/fonts/FreightSans/freight-sans-medium.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="../_static/fonts/IBMPlexMono/IBMPlexMono-Medium.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="../_static/fonts/FreightSans/freight-sans-bold.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="../_static/fonts/FreightSans/freight-sans-medium-italic.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="../_static/fonts/IBMPlexMono/IBMPlexMono-SemiBold.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<!-- Preload the katex fonts -->
<link rel="preload" href="https://cdn.jsdelivr.net/npm/katex@0.10.0/dist/fonts/KaTeX_Math-Italic.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="https://cdn.jsdelivr.net/npm/katex@0.10.0/dist/fonts/KaTeX_Main-Regular.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="https://cdn.jsdelivr.net/npm/katex@0.10.0/dist/fonts/KaTeX_Main-Bold.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="https://cdn.jsdelivr.net/npm/katex@0.10.0/dist/fonts/KaTeX_Size1-Regular.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="https://cdn.jsdelivr.net/npm/katex@0.10.0/dist/fonts/KaTeX_Size4-Regular.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="https://cdn.jsdelivr.net/npm/katex@0.10.0/dist/fonts/KaTeX_Size2-Regular.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="https://cdn.jsdelivr.net/npm/katex@0.10.0/dist/fonts/KaTeX_Size3-Regular.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="preload" href="https://cdn.jsdelivr.net/npm/katex@0.10.0/dist/fonts/KaTeX_Caligraphic-Regular.woff2" as="font" type="font/woff2" crossorigin="anonymous">
<link rel="stylesheet" href="https://use.fontawesome.com/releases/v5.15.2/css/all.css" integrity="sha384-vSIIfh2YWi9wW0r9iZe7RJPrKwp6bG+s9QZMoITbCckVJqGCCRhc+ccxNcdpHuYu" crossorigin="anonymous">
</head>
<div class="container-fluid header-holder tutorials-header" id="header-holder">
<div class="container">
<div class="header-container">
<a class="header-logo" href="https://pytorch.kr/" aria-label="PyTorch"></a>
<div class="main-menu">
<ul>
<li>
<a href="https://pytorch.kr/get-started">시작하기</a>
</li>
<li class="active">
<a href="https://tutorials.pytorch.kr">튜토리얼</a>
</li>
<li>
<a href="https://pytorch.kr/hub">허브</a>
</li>
<li>
<a href="https://discuss.pytorch.kr">커뮤니티</a>
</li>
</ul>
</div>
<a class="main-menu-open-button" href="#" data-behavior="open-mobile-menu"></a>
</div>
</div>
</div>
<body class="pytorch-body">
<div class="table-of-contents-link-wrapper">
<span>Table of Contents</span>
<a href="#" class="toggle-table-of-contents" data-behavior="toggle-table-of-contents"></a>
</div>
<nav data-toggle="wy-nav-shift" class="pytorch-left-menu" id="pytorch-left-menu">
<div class="pytorch-side-scroll">
<div class="pytorch-menu pytorch-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
<div class="pytorch-left-menu-search">
<div class="version">
1.10.2+cu102
</div>
<div role="search">
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
<input type="text" name="q" placeholder="Search Tutorials" />
<input type="hidden" name="check_keywords" value="yes" />
<input type="hidden" name="area" value="default" />
</form>
</div>
</div>
<p class="caption"><span class="caption-text">파이토치(PyTorch) 레시피</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../recipes/recipes_index.html">모든 레시피 보기</a></li>
<li class="toctree-l1"><a class="reference internal" href="../prototype/prototype_index.html">모든 프로토타입 레시피 보기</a></li>
</ul>
<p class="caption"><span class="caption-text">파이토치(PyTorch) 시작하기</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="basics/intro.html">파이토치(PyTorch) 기본 익히기</a></li>
<li class="toctree-l1"><a class="reference internal" href="basics/quickstart_tutorial.html">빠른 시작(Quickstart)</a></li>
<li class="toctree-l1"><a class="reference internal" href="basics/tensorqs_tutorial.html">텐서(Tensor)</a></li>
<li class="toctree-l1"><a class="reference internal" href="basics/data_tutorial.html">Dataset과 DataLoader</a></li>
<li class="toctree-l1"><a class="reference internal" href="basics/transforms_tutorial.html">변형(Transform)</a></li>
<li class="toctree-l1"><a class="reference internal" href="basics/buildmodel_tutorial.html">신경망 모델 구성하기</a></li>
<li class="toctree-l1"><a class="reference internal" href="basics/autogradqs_tutorial.html"><code class="docutils literal notranslate"><span class="pre">torch.autograd</span></code>를 사용한 자동 미분</a></li>
<li class="toctree-l1"><a class="reference internal" href="basics/optimization_tutorial.html">모델 매개변수 최적화하기</a></li>
<li class="toctree-l1"><a class="reference internal" href="basics/saveloadrun_tutorial.html">모델 저장하고 불러오기</a></li>
</ul>
<p class="caption"><span class="caption-text">Introduction to PyTorch on YouTube</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="introyt.html">Introduction to PyTorch - YouTube Series</a></li>
<li class="toctree-l1"><a class="reference internal" href="introyt/introyt1_tutorial.html">Introduction to PyTorch</a></li>
<li class="toctree-l1"><a class="reference internal" href="introyt/tensors_deeper_tutorial.html">Introduction to PyTorch Tensors</a></li>
<li class="toctree-l1"><a class="reference internal" href="introyt/autogradyt_tutorial.html">The Fundamentals of Autograd</a></li>
<li class="toctree-l1"><a class="reference internal" href="introyt/modelsyt_tutorial.html">Building Models with PyTorch</a></li>
<li class="toctree-l1"><a class="reference internal" href="introyt/tensorboardyt_tutorial.html">PyTorch TensorBoard Support</a></li>
<li class="toctree-l1"><a class="reference internal" href="introyt/trainingyt.html">Training with PyTorch</a></li>
<li class="toctree-l1"><a class="reference internal" href="introyt/captumyt.html">Model Understanding with Captum</a></li>
</ul>
<p class="caption"><span class="caption-text">파이토치(PyTorch) 배우기</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="deep_learning_60min_blitz.html">PyTorch로 딥러닝하기: 60분만에 끝장내기</a></li>
<li class="toctree-l1"><a class="reference internal" href="pytorch_with_examples.html">예제로 배우는 파이토치(PyTorch)</a></li>
<li class="toctree-l1"><a class="reference internal" href="nn_tutorial.html"><cite>torch.nn</cite> 이 <em>실제로</em> 무엇인가요?</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/tensorboard_tutorial.html">TensorBoard로 모델, 데이터, 학습 시각화하기</a></li>
</ul>
<p class="caption"><span class="caption-text">이미지/비디오</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/torchvision_tutorial.html">TorchVision 객체 검출 미세조정(Finetuning) 튜토리얼</a></li>
<li class="toctree-l1"><a class="reference internal" href="transfer_learning_tutorial.html">컴퓨터 비전(Vision)을 위한 전이학습(Transfer Learning)</a></li>
<li class="toctree-l1"><a class="reference internal" href="fgsm_tutorial.html">적대적 예제 생성(Adversarial Example Generation)</a></li>
<li class="toctree-l1"><a class="reference internal" href="dcgan_faces_tutorial.html">DCGAN 튜토리얼</a></li>
<li class="toctree-l1"><a class="reference internal" href="vt_tutorial.html">배포를 위한 비전 트랜스포머(Vision Transformer) 모델 최적화하기</a></li>
</ul>
<p class="caption"><span class="caption-text">오디오</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="audio_io_tutorial.html">Audio I/O</a></li>
<li class="toctree-l1"><a class="reference internal" href="audio_resampling_tutorial.html">Audio Resampling</a></li>
<li class="toctree-l1"><a class="reference internal" href="audio_data_augmentation_tutorial.html">Audio Data Augmentation</a></li>
<li class="toctree-l1"><a class="reference internal" href="audio_feature_extractions_tutorial.html">Audio Feature Extractions</a></li>
<li class="toctree-l1"><a class="reference internal" href="audio_feature_augmentation_tutorial.html">Audio Feature Augmentation</a></li>
<li class="toctree-l1"><a class="reference internal" href="audio_datasets_tutorial.html">Audio Datasets</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/speech_recognition_pipeline_tutorial.html">Speech Recognition with Wav2Vec2</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/speech_command_classification_with_torchaudio_tutorial.html">Speech Command Classification with torchaudio</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/text_to_speech_with_torchaudio.html">Text-to-speech with torchaudio</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/forced_alignment_with_torchaudio_tutorial.html">Forced Alignment with Wav2Vec2</a></li>
</ul>
<p class="caption"><span class="caption-text">텍스트</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="transformer_tutorial.html">nn.Transformer 와 TorchText 로 시퀀스-투-시퀀스(Sequence-to-Sequence) 모델링하기</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/char_rnn_classification_tutorial.html">기초부터 시작하는 NLP: 문자-단위 RNN으로 이름 분류하기</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/char_rnn_generation_tutorial.html">기초부터 시작하는 NLP: 문자-단위 RNN으로 이름 생성하기</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/seq2seq_translation_tutorial.html">기초부터 시작하는 NLP: Sequence to Sequence 네트워크와 Attention을 이용한 번역</a></li>
<li class="toctree-l1"><a class="reference internal" href="text_sentiment_ngrams_tutorial.html">torchtext 라이브러리로 텍스트 분류하기</a></li>
<li class="toctree-l1"><a class="reference internal" href="translation_transformer.html">nn.Transformer와 torchtext로 언어 번역하기</a></li>
</ul>
<p class="caption"><span class="caption-text">강화학습</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/reinforcement_q_learning.html">강화 학습 (DQN) 튜토리얼</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/mario_rl_tutorial.html">Train a Mario-playing RL Agent</a></li>
</ul>
<p class="caption"><span class="caption-text">PyTorch 모델을 프로덕션 환경에 배포하기</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/flask_rest_api_tutorial.html">Flask를 사용하여 Python에서 PyTorch를 REST API로 배포하기</a></li>
<li class="toctree-l1"><a class="reference internal" href="Intro_to_TorchScript_tutorial.html">TorchScript 소개</a></li>
<li class="toctree-l1"><a class="reference internal" href="../advanced/cpp_export.html">C++에서 TorchScript 모델 로딩하기</a></li>
<li class="toctree-l1"><a class="reference internal" href="../advanced/super_resolution_with_onnxruntime.html">(선택) PyTorch 모델을 ONNX으로 변환하고 ONNX 런타임에서 실행하기</a></li>
</ul>
<p class="caption"><span class="caption-text">Code Transforms with FX</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/fx_conv_bn_fuser.html">(베타) FX에서 합성곱/배치 정규화(Convolution/Batch Norm) 결합기(Fuser) 만들기</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/fx_profiling_tutorial.html">(beta) Building a Simple CPU Performance Profiler with FX</a></li>
</ul>
<p class="caption"><span class="caption-text">프론트엔드 API</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/memory_format_tutorial.html">(베타) PyTorch를 사용한 Channels Last 메모리 형식</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/forward_ad_usage.html">Forward-mode Automatic Differentiation (Beta)</a></li>
<li class="toctree-l1"><a class="reference internal" href="../advanced/cpp_frontend.html">PyTorch C++ 프론트엔드 사용하기</a></li>
<li class="toctree-l1"><a class="reference internal" href="../advanced/torch-script-parallelism.html">TorchScript의 동적 병렬 처리(Dynamic Parallelism)</a></li>
<li class="toctree-l1"><a class="reference internal" href="../advanced/cpp_autograd.html">C++ 프론트엔드의 자동 미분 (autograd)</a></li>
</ul>
<p class="caption"><span class="caption-text">PyTorch 확장하기</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/custom_function_double_backward_tutorial.html">Double Backward with Custom Functions</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/custom_function_conv_bn_tutorial.html">Fusing Convolution and Batch Norm using Custom Function</a></li>
<li class="toctree-l1"><a class="reference internal" href="../advanced/cpp_extension.html">Custom C++ and CUDA Extensions</a></li>
<li class="toctree-l1"><a class="reference internal" href="../advanced/torch_script_custom_ops.html">Extending TorchScript with Custom C++ Operators</a></li>
<li class="toctree-l1"><a class="reference internal" href="../advanced/torch_script_custom_classes.html">커스텀 C++ 클래스로 TorchScript 확장하기</a></li>
<li class="toctree-l1"><a class="reference internal" href="../advanced/dispatcher.html">Registering a Dispatched Operator in C++</a></li>
<li class="toctree-l1"><a class="reference internal" href="../advanced/extend_dispatcher.html">Extending dispatcher for a new backend in C++</a></li>
</ul>
<p class="caption"><span class="caption-text">모델 최적화</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="profiler.html">PyTorch 모듈 프로파일링 하기</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/tensorboard_profiler_tutorial.html">PyTorch Profiler With TensorBoard</a></li>
<li class="toctree-l1"><a class="reference internal" href="hyperparameter_tuning_tutorial.html">Hyperparameter tuning with Ray Tune</a></li>
<li class="toctree-l1"><a class="reference internal" href="vt_tutorial.html">배포를 위한 비전 트랜스포머(Vision Transformer) 모델 최적화하기</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/parametrizations.html">Parametrizations Tutorial</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/pruning_tutorial.html">가지치기 기법(Pruning) 튜토리얼</a></li>
<li class="toctree-l1"><a class="reference internal" href="../advanced/dynamic_quantization_tutorial.html">(베타) LSTM 기반 단어 단위 언어 모델의 동적 양자화</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/dynamic_quantization_bert_tutorial.html">(베타) BERT 모델 동적 양자화하기</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/quantized_transfer_learning_tutorial.html">(베타) 컴퓨터 비전 튜토리얼을 위한 양자화된 전이학습(Quantized Transfer Learning)</a></li>
<li class="toctree-l1"><a class="reference internal" href="../advanced/static_quantization_tutorial.html">(beta) Static Quantization with Eager Mode in PyTorch</a></li>
</ul>
<p class="caption"><span class="caption-text">병렬 및 분산 학습</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="dist_overview.html">PyTorch Distributed Overview</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/model_parallel_tutorial.html">단일 머신을 사용한 모델 병렬화 모범 사례</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/ddp_tutorial.html">분산 데이터 병렬 처리 시작하기</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/dist_tuto.html">PyTorch로 분산 어플리케이션 개발하기</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/rpc_tutorial.html">Getting Started with Distributed RPC Framework</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/rpc_param_server_tutorial.html">Implementing a Parameter Server Using Distributed RPC Framework</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/dist_pipeline_parallel_tutorial.html">Distributed Pipeline Parallelism Using RPC</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/rpc_async_execution.html">Implementing Batch RPC Processing Using Asynchronous Executions</a></li>
<li class="toctree-l1"><a class="reference internal" href="../advanced/rpc_ddp_tutorial.html">분산 데이터 병렬(DDP)과 분산 RPC 프레임워크 결합</a></li>
<li class="toctree-l1"><a class="reference internal" href="../intermediate/pipeline_tutorial.html">파이프라인 병렬화로 트랜스포머 모델 학습시키기</a></li>
<li class="toctree-l1"><a class="reference internal" href="../advanced/ddp_pipeline.html">분산 데이터 병렬 처리와 병렬 처리 파이프라인을 사용한 트랜스포머 모델 학습</a></li>
<li class="toctree-l1"><a class="reference internal" href="../advanced/generic_join.html">Distributed Training with Uneven Inputs Using the Join Context Manager</a></li>
</ul>
<p class="caption"><span class="caption-text">Mobile</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="deeplabv3_on_ios.html">iOS에서의 이미지 분할 DeepLapV3</a></li>
<li class="toctree-l1"><a class="reference internal" href="deeplabv3_on_android.html">안드로이드에서의 이미지 분할 DeepLapV3</a></li>
</ul>
</div>
</div>
</nav>
<div class="pytorch-container">
<div class="pytorch-page-level-bar" id="pytorch-page-level-bar">
<div class="pytorch-breadcrumbs-wrapper">
<div role="navigation" aria-label="breadcrumbs navigation">
<ul class="pytorch-breadcrumbs">
<li>
<a href="../index.html">
Tutorials
</a> >
</li>
<li>Audio manipulation with torchaudio</li>
<li class="pytorch-breadcrumbs-aside">
<a href="../_sources/beginner/audio_preprocessing_tutorial.rst.txt" rel="nofollow"><img src="../_static/images/view-page-source-icon.svg"></a>
</li>
</ul>
</div>
</div>
<div class="pytorch-shortcuts-wrapper" id="pytorch-shortcuts-wrapper">
Shortcuts
</div>
</div>
<section data-toggle="wy-nav-shift" id="pytorch-content-wrap" class="pytorch-content-wrap">
<div class="pytorch-content-left">
<div class="pytorch-call-to-action-links">
<div id="tutorial-type">beginner/audio_preprocessing_tutorial</div>
<div id="google-colab-link">
<img class="call-to-action-img" src="../_static/images/pytorch-colab.svg"/>
<div class="call-to-action-desktop-view">Run in Google Colab</div>
<div class="call-to-action-mobile-view">Colab</div>
</div>
<div id="download-notebook-link">
<img class="call-to-action-notebook-img" src="../_static/images/pytorch-download.svg"/>
<div class="call-to-action-desktop-view">Download Notebook</div>
<div class="call-to-action-mobile-view">Notebook</div>
</div>
<div id="github-view-link">
<img class="call-to-action-img" src="../_static/images/pytorch-github.svg"/>
<div class="call-to-action-desktop-view">View on GitHub</div>
<div class="call-to-action-mobile-view">GitHub</div>
</div>
</div>
<div class="rst-content">
<div role="main" class="main-content" itemscope="itemscope" itemtype="http://schema.org/Article">
<article itemprop="articleBody" id="pytorch-article" class="pytorch-article">
<div class="sphx-glr-download-link-note admonition note">
<p class="first admonition-title">Note</p>
<p class="last">Click <a class="reference internal" href="#sphx-glr-download-beginner-audio-preprocessing-tutorial-py"><span class="std std-ref">here</span></a> to download the full example code</p>
</div>
<div class="sphx-glr-example-title section" id="audio-manipulation-with-torchaudio">
<span id="sphx-glr-beginner-audio-preprocessing-tutorial-py"></span><h1>Audio manipulation with torchaudio<a class="headerlink" href="#audio-manipulation-with-torchaudio" title="Permalink to this headline">¶</a></h1>
<p><code class="docutils literal notranslate"><span class="pre">torchaudio</span></code> provides powerful audio I/O functions, preprocessing
transforms and dataset.</p>
<p>In this tutorial, we will look into how to prepare audio data and
extract features that can be fed to NN models.</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># When running this tutorial in Google Colab, install the required packages</span>
<span class="c1"># with the following.</span>
<span class="c1"># !pip install torchaudio librosa boto3</span>
<span class="kn">import</span> <span class="nn">torch</span>
<span class="kn">import</span> <span class="nn">torchaudio</span>
<span class="kn">import</span> <span class="nn">torchaudio.functional</span> <span class="k">as</span> <span class="nn">F</span>
<span class="kn">import</span> <span class="nn">torchaudio.transforms</span> <span class="k">as</span> <span class="nn">T</span>
<span class="nb">print</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">__version__</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="n">torchaudio</span><span class="o">.</span><span class="n">__version__</span><span class="p">)</span>
</pre></div>
</div>
<p class="sphx-glr-script-out">Out:</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>1.9.0+cu102
0.9.0
</pre></div>
</div>
<div class="section" id="preparing-data-and-utility-functions-skip-this-section">
<h2>Preparing data and utility functions (skip this section)<a class="headerlink" href="#preparing-data-and-utility-functions-skip-this-section" title="Permalink to this headline">¶</a></h2>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1">#@title Prepare data and utility functions. {display-mode: "form"}</span>
<span class="c1">#@markdown</span>
<span class="c1">#@markdown You do not need to look into this cell.</span>
<span class="c1">#@markdown Just execute once and you are good to go.</span>
<span class="c1">#@markdown</span>
<span class="c1">#@markdown In this tutorial, we will use a speech data from [VOiCES dataset](https://iqtlabs.github.io/voices/), which is licensed under Creative Commos BY 4.0.</span>
<span class="c1">#-------------------------------------------------------------------------------</span>
<span class="c1"># Preparation of data and helper functions.</span>
<span class="c1">#-------------------------------------------------------------------------------</span>
<span class="kn">import</span> <span class="nn">io</span>
<span class="kn">import</span> <span class="nn">os</span>
<span class="kn">import</span> <span class="nn">math</span>
<span class="kn">import</span> <span class="nn">tarfile</span>
<span class="kn">import</span> <span class="nn">multiprocessing</span>
<span class="kn">import</span> <span class="nn">scipy</span>
<span class="kn">import</span> <span class="nn">librosa</span>
<span class="kn">import</span> <span class="nn">boto3</span>
<span class="kn">from</span> <span class="nn">botocore</span> <span class="kn">import</span> <span class="n">UNSIGNED</span>
<span class="kn">from</span> <span class="nn">botocore.config</span> <span class="kn">import</span> <span class="n">Config</span>
<span class="kn">import</span> <span class="nn">requests</span>
<span class="kn">import</span> <span class="nn">matplotlib</span>
<span class="kn">import</span> <span class="nn">matplotlib.pyplot</span> <span class="k">as</span> <span class="nn">plt</span>
<span class="kn">import</span> <span class="nn">pandas</span> <span class="k">as</span> <span class="nn">pd</span>
<span class="kn">import</span> <span class="nn">time</span>
<span class="kn">from</span> <span class="nn">IPython.display</span> <span class="kn">import</span> <span class="n">Audio</span><span class="p">,</span> <span class="n">display</span>
<span class="p">[</span><span class="n">width</span><span class="p">,</span> <span class="n">height</span><span class="p">]</span> <span class="o">=</span> <span class="n">matplotlib</span><span class="o">.</span><span class="n">rcParams</span><span class="p">[</span><span class="s1">'figure.figsize'</span><span class="p">]</span>
<span class="k">if</span> <span class="n">width</span> <span class="o"><</span> <span class="mi">10</span><span class="p">:</span>
<span class="n">matplotlib</span><span class="o">.</span><span class="n">rcParams</span><span class="p">[</span><span class="s1">'figure.figsize'</span><span class="p">]</span> <span class="o">=</span> <span class="p">[</span><span class="n">width</span> <span class="o">*</span> <span class="mf">2.5</span><span class="p">,</span> <span class="n">height</span><span class="p">]</span>
<span class="n">_SAMPLE_DIR</span> <span class="o">=</span> <span class="s2">"_sample_data"</span>
<span class="n">SAMPLE_WAV_URL</span> <span class="o">=</span> <span class="s2">"https://pytorch-tutorial-assets.s3.amazonaws.com/steam-train-whistle-daniel_simon.wav"</span>
<span class="n">SAMPLE_WAV_PATH</span> <span class="o">=</span> <span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">_SAMPLE_DIR</span><span class="p">,</span> <span class="s2">"steam.wav"</span><span class="p">)</span>
<span class="n">SAMPLE_WAV_SPEECH_URL</span> <span class="o">=</span> <span class="s2">"https://pytorch-tutorial-assets.s3.amazonaws.com/VOiCES_devkit/source-16k/train/sp0307/Lab41-SRI-VOiCES-src-sp0307-ch127535-sg0042.wav"</span>
<span class="n">SAMPLE_WAV_SPEECH_PATH</span> <span class="o">=</span> <span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">_SAMPLE_DIR</span><span class="p">,</span> <span class="s2">"speech.wav"</span><span class="p">)</span>
<span class="n">SAMPLE_RIR_URL</span> <span class="o">=</span> <span class="s2">"https://pytorch-tutorial-assets.s3.amazonaws.com/VOiCES_devkit/distant-16k/room-response/rm1/impulse/Lab41-SRI-VOiCES-rm1-impulse-mc01-stu-clo.wav"</span>
<span class="n">SAMPLE_RIR_PATH</span> <span class="o">=</span> <span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">_SAMPLE_DIR</span><span class="p">,</span> <span class="s2">"rir.wav"</span><span class="p">)</span>
<span class="n">SAMPLE_NOISE_URL</span> <span class="o">=</span> <span class="s2">"https://pytorch-tutorial-assets.s3.amazonaws.com/VOiCES_devkit/distant-16k/distractors/rm1/babb/Lab41-SRI-VOiCES-rm1-babb-mc01-stu-clo.wav"</span>
<span class="n">SAMPLE_NOISE_PATH</span> <span class="o">=</span> <span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">_SAMPLE_DIR</span><span class="p">,</span> <span class="s2">"bg.wav"</span><span class="p">)</span>
<span class="n">SAMPLE_MP3_URL</span> <span class="o">=</span> <span class="s2">"https://pytorch-tutorial-assets.s3.amazonaws.com/steam-train-whistle-daniel_simon.mp3"</span>
<span class="n">SAMPLE_MP3_PATH</span> <span class="o">=</span> <span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">_SAMPLE_DIR</span><span class="p">,</span> <span class="s2">"steam.mp3"</span><span class="p">)</span>
<span class="n">SAMPLE_GSM_URL</span> <span class="o">=</span> <span class="s2">"https://pytorch-tutorial-assets.s3.amazonaws.com/steam-train-whistle-daniel_simon.gsm"</span>
<span class="n">SAMPLE_GSM_PATH</span> <span class="o">=</span> <span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">_SAMPLE_DIR</span><span class="p">,</span> <span class="s2">"steam.gsm"</span><span class="p">)</span>
<span class="n">SAMPLE_TAR_URL</span> <span class="o">=</span> <span class="s2">"https://pytorch-tutorial-assets.s3.amazonaws.com/VOiCES_devkit.tar.gz"</span>
<span class="n">SAMPLE_TAR_PATH</span> <span class="o">=</span> <span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">_SAMPLE_DIR</span><span class="p">,</span> <span class="s2">"sample.tar.gz"</span><span class="p">)</span>
<span class="n">SAMPLE_TAR_ITEM</span> <span class="o">=</span> <span class="s2">"VOiCES_devkit/source-16k/train/sp0307/Lab41-SRI-VOiCES-src-sp0307-ch127535-sg0042.wav"</span>
<span class="n">S3_BUCKET</span> <span class="o">=</span> <span class="s2">"pytorch-tutorial-assets"</span>
<span class="n">S3_KEY</span> <span class="o">=</span> <span class="s2">"VOiCES_devkit/source-16k/train/sp0307/Lab41-SRI-VOiCES-src-sp0307-ch127535-sg0042.wav"</span>
<span class="n">YESNO_DATASET_PATH</span> <span class="o">=</span> <span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">_SAMPLE_DIR</span><span class="p">,</span> <span class="s2">"yes_no"</span><span class="p">)</span>
<span class="n">os</span><span class="o">.</span><span class="n">makedirs</span><span class="p">(</span><span class="n">YESNO_DATASET_PATH</span><span class="p">,</span> <span class="n">exist_ok</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="n">os</span><span class="o">.</span><span class="n">makedirs</span><span class="p">(</span><span class="n">_SAMPLE_DIR</span><span class="p">,</span> <span class="n">exist_ok</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">_fetch_data</span><span class="p">():</span>
<span class="n">uri</span> <span class="o">=</span> <span class="p">[</span>
<span class="p">(</span><span class="n">SAMPLE_WAV_URL</span><span class="p">,</span> <span class="n">SAMPLE_WAV_PATH</span><span class="p">),</span>
<span class="p">(</span><span class="n">SAMPLE_WAV_SPEECH_URL</span><span class="p">,</span> <span class="n">SAMPLE_WAV_SPEECH_PATH</span><span class="p">),</span>
<span class="p">(</span><span class="n">SAMPLE_RIR_URL</span><span class="p">,</span> <span class="n">SAMPLE_RIR_PATH</span><span class="p">),</span>
<span class="p">(</span><span class="n">SAMPLE_NOISE_URL</span><span class="p">,</span> <span class="n">SAMPLE_NOISE_PATH</span><span class="p">),</span>
<span class="p">(</span><span class="n">SAMPLE_MP3_URL</span><span class="p">,</span> <span class="n">SAMPLE_MP3_PATH</span><span class="p">),</span>
<span class="p">(</span><span class="n">SAMPLE_GSM_URL</span><span class="p">,</span> <span class="n">SAMPLE_GSM_PATH</span><span class="p">),</span>
<span class="p">(</span><span class="n">SAMPLE_TAR_URL</span><span class="p">,</span> <span class="n">SAMPLE_TAR_PATH</span><span class="p">),</span>
<span class="p">]</span>
<span class="k">for</span> <span class="n">url</span><span class="p">,</span> <span class="n">path</span> <span class="ow">in</span> <span class="n">uri</span><span class="p">:</span>
<span class="k">with</span> <span class="nb">open</span><span class="p">(</span><span class="n">path</span><span class="p">,</span> <span class="s1">'wb'</span><span class="p">)</span> <span class="k">as</span> <span class="n">file_</span><span class="p">:</span>
<span class="n">file_</span><span class="o">.</span><span class="n">write</span><span class="p">(</span><span class="n">requests</span><span class="o">.</span><span class="n">get</span><span class="p">(</span><span class="n">url</span><span class="p">)</span><span class="o">.</span><span class="n">content</span><span class="p">)</span>
<span class="n">_fetch_data</span><span class="p">()</span>
<span class="k">def</span> <span class="nf">_download_yesno</span><span class="p">():</span>
<span class="k">if</span> <span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">exists</span><span class="p">(</span><span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">YESNO_DATASET_PATH</span><span class="p">,</span> <span class="s2">"waves_yesno.tar.gz"</span><span class="p">)):</span>
<span class="k">return</span>
<span class="n">torchaudio</span><span class="o">.</span><span class="n">datasets</span><span class="o">.</span><span class="n">YESNO</span><span class="p">(</span><span class="n">root</span><span class="o">=</span><span class="n">YESNO_DATASET_PATH</span><span class="p">,</span> <span class="n">download</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="n">YESNO_DOWNLOAD_PROCESS</span> <span class="o">=</span> <span class="n">multiprocessing</span><span class="o">.</span><span class="n">Process</span><span class="p">(</span><span class="n">target</span><span class="o">=</span><span class="n">_download_yesno</span><span class="p">)</span>
<span class="n">YESNO_DOWNLOAD_PROCESS</span><span class="o">.</span><span class="n">start</span><span class="p">()</span>
<span class="k">def</span> <span class="nf">_get_sample</span><span class="p">(</span><span class="n">path</span><span class="p">,</span> <span class="n">resample</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="n">effects</span> <span class="o">=</span> <span class="p">[</span>
<span class="p">[</span><span class="s2">"remix"</span><span class="p">,</span> <span class="s2">"1"</span><span class="p">]</span>
<span class="p">]</span>
<span class="k">if</span> <span class="n">resample</span><span class="p">:</span>
<span class="n">effects</span><span class="o">.</span><span class="n">extend</span><span class="p">([</span>
<span class="p">[</span><span class="s2">"lowpass"</span><span class="p">,</span> <span class="sa">f</span><span class="s2">"</span><span class="si">{</span><span class="n">resample</span> <span class="o">//</span> <span class="mi">2</span><span class="si">}</span><span class="s2">"</span><span class="p">],</span>
<span class="p">[</span><span class="s2">"rate"</span><span class="p">,</span> <span class="sa">f</span><span class="s1">'</span><span class="si">{</span><span class="n">resample</span><span class="si">}</span><span class="s1">'</span><span class="p">],</span>
<span class="p">])</span>
<span class="k">return</span> <span class="n">torchaudio</span><span class="o">.</span><span class="n">sox_effects</span><span class="o">.</span><span class="n">apply_effects_file</span><span class="p">(</span><span class="n">path</span><span class="p">,</span> <span class="n">effects</span><span class="o">=</span><span class="n">effects</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">get_speech_sample</span><span class="p">(</span><span class="o">*</span><span class="p">,</span> <span class="n">resample</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="k">return</span> <span class="n">_get_sample</span><span class="p">(</span><span class="n">SAMPLE_WAV_SPEECH_PATH</span><span class="p">,</span> <span class="n">resample</span><span class="o">=</span><span class="n">resample</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">get_sample</span><span class="p">(</span><span class="o">*</span><span class="p">,</span> <span class="n">resample</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="k">return</span> <span class="n">_get_sample</span><span class="p">(</span><span class="n">SAMPLE_WAV_PATH</span><span class="p">,</span> <span class="n">resample</span><span class="o">=</span><span class="n">resample</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">get_rir_sample</span><span class="p">(</span><span class="o">*</span><span class="p">,</span> <span class="n">resample</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">processed</span><span class="o">=</span><span class="kc">False</span><span class="p">):</span>
<span class="n">rir_raw</span><span class="p">,</span> <span class="n">sample_rate</span> <span class="o">=</span> <span class="n">_get_sample</span><span class="p">(</span><span class="n">SAMPLE_RIR_PATH</span><span class="p">,</span> <span class="n">resample</span><span class="o">=</span><span class="n">resample</span><span class="p">)</span>
<span class="k">if</span> <span class="ow">not</span> <span class="n">processed</span><span class="p">:</span>
<span class="k">return</span> <span class="n">rir_raw</span><span class="p">,</span> <span class="n">sample_rate</span>
<span class="n">rir</span> <span class="o">=</span> <span class="n">rir_raw</span><span class="p">[:,</span> <span class="nb">int</span><span class="p">(</span><span class="n">sample_rate</span><span class="o">*</span><span class="mf">1.01</span><span class="p">):</span><span class="nb">int</span><span class="p">(</span><span class="n">sample_rate</span><span class="o">*</span><span class="mf">1.3</span><span class="p">)]</span>
<span class="n">rir</span> <span class="o">=</span> <span class="n">rir</span> <span class="o">/</span> <span class="n">torch</span><span class="o">.</span><span class="n">norm</span><span class="p">(</span><span class="n">rir</span><span class="p">,</span> <span class="n">p</span><span class="o">=</span><span class="mi">2</span><span class="p">)</span>
<span class="n">rir</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">flip</span><span class="p">(</span><span class="n">rir</span><span class="p">,</span> <span class="p">[</span><span class="mi">1</span><span class="p">])</span>
<span class="k">return</span> <span class="n">rir</span><span class="p">,</span> <span class="n">sample_rate</span>
<span class="k">def</span> <span class="nf">get_noise_sample</span><span class="p">(</span><span class="o">*</span><span class="p">,</span> <span class="n">resample</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="k">return</span> <span class="n">_get_sample</span><span class="p">(</span><span class="n">SAMPLE_NOISE_PATH</span><span class="p">,</span> <span class="n">resample</span><span class="o">=</span><span class="n">resample</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">print_stats</span><span class="p">(</span><span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">src</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="k">if</span> <span class="n">src</span><span class="p">:</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">"-"</span> <span class="o">*</span> <span class="mi">10</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">"Source:"</span><span class="p">,</span> <span class="n">src</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">"-"</span> <span class="o">*</span> <span class="mi">10</span><span class="p">)</span>
<span class="k">if</span> <span class="n">sample_rate</span><span class="p">:</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">"Sample Rate:"</span><span class="p">,</span> <span class="n">sample_rate</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">"Shape:"</span><span class="p">,</span> <span class="nb">tuple</span><span class="p">(</span><span class="n">waveform</span><span class="o">.</span><span class="n">shape</span><span class="p">))</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">"Dtype:"</span><span class="p">,</span> <span class="n">waveform</span><span class="o">.</span><span class="n">dtype</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="sa">f</span><span class="s2">" - Max: </span><span class="si">{</span><span class="n">waveform</span><span class="o">.</span><span class="n">max</span><span class="p">()</span><span class="o">.</span><span class="n">item</span><span class="p">()</span><span class="si">:</span><span class="s2">6.3f</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="sa">f</span><span class="s2">" - Min: </span><span class="si">{</span><span class="n">waveform</span><span class="o">.</span><span class="n">min</span><span class="p">()</span><span class="o">.</span><span class="n">item</span><span class="p">()</span><span class="si">:</span><span class="s2">6.3f</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="sa">f</span><span class="s2">" - Mean: </span><span class="si">{</span><span class="n">waveform</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span><span class="o">.</span><span class="n">item</span><span class="p">()</span><span class="si">:</span><span class="s2">6.3f</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="sa">f</span><span class="s2">" - Std Dev: </span><span class="si">{</span><span class="n">waveform</span><span class="o">.</span><span class="n">std</span><span class="p">()</span><span class="o">.</span><span class="n">item</span><span class="p">()</span><span class="si">:</span><span class="s2">6.3f</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span>
<span class="nb">print</span><span class="p">()</span>
<span class="nb">print</span><span class="p">(</span><span class="n">waveform</span><span class="p">)</span>
<span class="nb">print</span><span class="p">()</span>
<span class="k">def</span> <span class="nf">plot_waveform</span><span class="p">(</span><span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span><span class="p">,</span> <span class="n">title</span><span class="o">=</span><span class="s2">"Waveform"</span><span class="p">,</span> <span class="n">xlim</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">ylim</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="n">waveform</span> <span class="o">=</span> <span class="n">waveform</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span>
<span class="n">num_channels</span><span class="p">,</span> <span class="n">num_frames</span> <span class="o">=</span> <span class="n">waveform</span><span class="o">.</span><span class="n">shape</span>
<span class="n">time_axis</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">num_frames</span><span class="p">)</span> <span class="o">/</span> <span class="n">sample_rate</span>
<span class="n">figure</span><span class="p">,</span> <span class="n">axes</span> <span class="o">=</span> <span class="n">plt</span><span class="o">.</span><span class="n">subplots</span><span class="p">(</span><span class="n">num_channels</span><span class="p">,</span> <span class="mi">1</span><span class="p">)</span>
<span class="k">if</span> <span class="n">num_channels</span> <span class="o">==</span> <span class="mi">1</span><span class="p">:</span>
<span class="n">axes</span> <span class="o">=</span> <span class="p">[</span><span class="n">axes</span><span class="p">]</span>
<span class="k">for</span> <span class="n">c</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">num_channels</span><span class="p">):</span>
<span class="n">axes</span><span class="p">[</span><span class="n">c</span><span class="p">]</span><span class="o">.</span><span class="n">plot</span><span class="p">(</span><span class="n">time_axis</span><span class="p">,</span> <span class="n">waveform</span><span class="p">[</span><span class="n">c</span><span class="p">],</span> <span class="n">linewidth</span><span class="o">=</span><span class="mi">1</span><span class="p">)</span>
<span class="n">axes</span><span class="p">[</span><span class="n">c</span><span class="p">]</span><span class="o">.</span><span class="n">grid</span><span class="p">(</span><span class="kc">True</span><span class="p">)</span>
<span class="k">if</span> <span class="n">num_channels</span> <span class="o">></span> <span class="mi">1</span><span class="p">:</span>
<span class="n">axes</span><span class="p">[</span><span class="n">c</span><span class="p">]</span><span class="o">.</span><span class="n">set_ylabel</span><span class="p">(</span><span class="sa">f</span><span class="s1">'Channel </span><span class="si">{</span><span class="n">c</span><span class="o">+</span><span class="mi">1</span><span class="si">}</span><span class="s1">'</span><span class="p">)</span>
<span class="k">if</span> <span class="n">xlim</span><span class="p">:</span>
<span class="n">axes</span><span class="p">[</span><span class="n">c</span><span class="p">]</span><span class="o">.</span><span class="n">set_xlim</span><span class="p">(</span><span class="n">xlim</span><span class="p">)</span>
<span class="k">if</span> <span class="n">ylim</span><span class="p">:</span>
<span class="n">axes</span><span class="p">[</span><span class="n">c</span><span class="p">]</span><span class="o">.</span><span class="n">set_ylim</span><span class="p">(</span><span class="n">ylim</span><span class="p">)</span>
<span class="n">figure</span><span class="o">.</span><span class="n">suptitle</span><span class="p">(</span><span class="n">title</span><span class="p">)</span>
<span class="n">plt</span><span class="o">.</span><span class="n">show</span><span class="p">(</span><span class="n">block</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">plot_specgram</span><span class="p">(</span><span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span><span class="p">,</span> <span class="n">title</span><span class="o">=</span><span class="s2">"Spectrogram"</span><span class="p">,</span> <span class="n">xlim</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="n">waveform</span> <span class="o">=</span> <span class="n">waveform</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span>
<span class="n">num_channels</span><span class="p">,</span> <span class="n">num_frames</span> <span class="o">=</span> <span class="n">waveform</span><span class="o">.</span><span class="n">shape</span>
<span class="n">time_axis</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">num_frames</span><span class="p">)</span> <span class="o">/</span> <span class="n">sample_rate</span>
<span class="n">figure</span><span class="p">,</span> <span class="n">axes</span> <span class="o">=</span> <span class="n">plt</span><span class="o">.</span><span class="n">subplots</span><span class="p">(</span><span class="n">num_channels</span><span class="p">,</span> <span class="mi">1</span><span class="p">)</span>
<span class="k">if</span> <span class="n">num_channels</span> <span class="o">==</span> <span class="mi">1</span><span class="p">:</span>
<span class="n">axes</span> <span class="o">=</span> <span class="p">[</span><span class="n">axes</span><span class="p">]</span>
<span class="k">for</span> <span class="n">c</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">num_channels</span><span class="p">):</span>
<span class="n">axes</span><span class="p">[</span><span class="n">c</span><span class="p">]</span><span class="o">.</span><span class="n">specgram</span><span class="p">(</span><span class="n">waveform</span><span class="p">[</span><span class="n">c</span><span class="p">],</span> <span class="n">Fs</span><span class="o">=</span><span class="n">sample_rate</span><span class="p">)</span>
<span class="k">if</span> <span class="n">num_channels</span> <span class="o">></span> <span class="mi">1</span><span class="p">:</span>
<span class="n">axes</span><span class="p">[</span><span class="n">c</span><span class="p">]</span><span class="o">.</span><span class="n">set_ylabel</span><span class="p">(</span><span class="sa">f</span><span class="s1">'Channel </span><span class="si">{</span><span class="n">c</span><span class="o">+</span><span class="mi">1</span><span class="si">}</span><span class="s1">'</span><span class="p">)</span>
<span class="k">if</span> <span class="n">xlim</span><span class="p">:</span>
<span class="n">axes</span><span class="p">[</span><span class="n">c</span><span class="p">]</span><span class="o">.</span><span class="n">set_xlim</span><span class="p">(</span><span class="n">xlim</span><span class="p">)</span>
<span class="n">figure</span><span class="o">.</span><span class="n">suptitle</span><span class="p">(</span><span class="n">title</span><span class="p">)</span>
<span class="n">plt</span><span class="o">.</span><span class="n">show</span><span class="p">(</span><span class="n">block</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">play_audio</span><span class="p">(</span><span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span><span class="p">):</span>
<span class="n">waveform</span> <span class="o">=</span> <span class="n">waveform</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span>
<span class="n">num_channels</span><span class="p">,</span> <span class="n">num_frames</span> <span class="o">=</span> <span class="n">waveform</span><span class="o">.</span><span class="n">shape</span>
<span class="k">if</span> <span class="n">num_channels</span> <span class="o">==</span> <span class="mi">1</span><span class="p">:</span>
<span class="n">display</span><span class="p">(</span><span class="n">Audio</span><span class="p">(</span><span class="n">waveform</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="n">rate</span><span class="o">=</span><span class="n">sample_rate</span><span class="p">))</span>
<span class="k">elif</span> <span class="n">num_channels</span> <span class="o">==</span> <span class="mi">2</span><span class="p">:</span>
<span class="n">display</span><span class="p">(</span><span class="n">Audio</span><span class="p">((</span><span class="n">waveform</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="n">waveform</span><span class="p">[</span><span class="mi">1</span><span class="p">]),</span> <span class="n">rate</span><span class="o">=</span><span class="n">sample_rate</span><span class="p">))</span>
<span class="k">else</span><span class="p">:</span>
<span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="s2">"Waveform with more than 2 channels are not supported."</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">inspect_file</span><span class="p">(</span><span class="n">path</span><span class="p">):</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">"-"</span> <span class="o">*</span> <span class="mi">10</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">"Source:"</span><span class="p">,</span> <span class="n">path</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">"-"</span> <span class="o">*</span> <span class="mi">10</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="sa">f</span><span class="s2">" - File size: </span><span class="si">{</span><span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">getsize</span><span class="p">(</span><span class="n">path</span><span class="p">)</span><span class="si">}</span><span class="s2"> bytes"</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="sa">f</span><span class="s2">" - </span><span class="si">{</span><span class="n">torchaudio</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="n">path</span><span class="p">)</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">plot_spectrogram</span><span class="p">(</span><span class="n">spec</span><span class="p">,</span> <span class="n">title</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">ylabel</span><span class="o">=</span><span class="s1">'freq_bin'</span><span class="p">,</span> <span class="n">aspect</span><span class="o">=</span><span class="s1">'auto'</span><span class="p">,</span> <span class="n">xmax</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="n">fig</span><span class="p">,</span> <span class="n">axs</span> <span class="o">=</span> <span class="n">plt</span><span class="o">.</span><span class="n">subplots</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="mi">1</span><span class="p">)</span>
<span class="n">axs</span><span class="o">.</span><span class="n">set_title</span><span class="p">(</span><span class="n">title</span> <span class="ow">or</span> <span class="s1">'Spectrogram (db)'</span><span class="p">)</span>
<span class="n">axs</span><span class="o">.</span><span class="n">set_ylabel</span><span class="p">(</span><span class="n">ylabel</span><span class="p">)</span>
<span class="n">axs</span><span class="o">.</span><span class="n">set_xlabel</span><span class="p">(</span><span class="s1">'frame'</span><span class="p">)</span>
<span class="n">im</span> <span class="o">=</span> <span class="n">axs</span><span class="o">.</span><span class="n">imshow</span><span class="p">(</span><span class="n">librosa</span><span class="o">.</span><span class="n">power_to_db</span><span class="p">(</span><span class="n">spec</span><span class="p">),</span> <span class="n">origin</span><span class="o">=</span><span class="s1">'lower'</span><span class="p">,</span> <span class="n">aspect</span><span class="o">=</span><span class="n">aspect</span><span class="p">)</span>
<span class="k">if</span> <span class="n">xmax</span><span class="p">:</span>
<span class="n">axs</span><span class="o">.</span><span class="n">set_xlim</span><span class="p">((</span><span class="mi">0</span><span class="p">,</span> <span class="n">xmax</span><span class="p">))</span>
<span class="n">fig</span><span class="o">.</span><span class="n">colorbar</span><span class="p">(</span><span class="n">im</span><span class="p">,</span> <span class="n">ax</span><span class="o">=</span><span class="n">axs</span><span class="p">)</span>
<span class="n">plt</span><span class="o">.</span><span class="n">show</span><span class="p">(</span><span class="n">block</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">plot_mel_fbank</span><span class="p">(</span><span class="n">fbank</span><span class="p">,</span> <span class="n">title</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="n">fig</span><span class="p">,</span> <span class="n">axs</span> <span class="o">=</span> <span class="n">plt</span><span class="o">.</span><span class="n">subplots</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="mi">1</span><span class="p">)</span>
<span class="n">axs</span><span class="o">.</span><span class="n">set_title</span><span class="p">(</span><span class="n">title</span> <span class="ow">or</span> <span class="s1">'Filter bank'</span><span class="p">)</span>
<span class="n">axs</span><span class="o">.</span><span class="n">imshow</span><span class="p">(</span><span class="n">fbank</span><span class="p">,</span> <span class="n">aspect</span><span class="o">=</span><span class="s1">'auto'</span><span class="p">)</span>
<span class="n">axs</span><span class="o">.</span><span class="n">set_ylabel</span><span class="p">(</span><span class="s1">'frequency bin'</span><span class="p">)</span>
<span class="n">axs</span><span class="o">.</span><span class="n">set_xlabel</span><span class="p">(</span><span class="s1">'mel bin'</span><span class="p">)</span>
<span class="n">plt</span><span class="o">.</span><span class="n">show</span><span class="p">(</span><span class="n">block</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">get_spectrogram</span><span class="p">(</span>
<span class="n">n_fft</span> <span class="o">=</span> <span class="mi">400</span><span class="p">,</span>
<span class="n">win_len</span> <span class="o">=</span> <span class="kc">None</span><span class="p">,</span>
<span class="n">hop_len</span> <span class="o">=</span> <span class="kc">None</span><span class="p">,</span>
<span class="n">power</span> <span class="o">=</span> <span class="mf">2.0</span><span class="p">,</span>
<span class="p">):</span>
<span class="n">waveform</span><span class="p">,</span> <span class="n">_</span> <span class="o">=</span> <span class="n">get_speech_sample</span><span class="p">()</span>
<span class="n">spectrogram</span> <span class="o">=</span> <span class="n">T</span><span class="o">.</span><span class="n">Spectrogram</span><span class="p">(</span>
<span class="n">n_fft</span><span class="o">=</span><span class="n">n_fft</span><span class="p">,</span>
<span class="n">win_length</span><span class="o">=</span><span class="n">win_len</span><span class="p">,</span>
<span class="n">hop_length</span><span class="o">=</span><span class="n">hop_len</span><span class="p">,</span>
<span class="n">center</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span>
<span class="n">pad_mode</span><span class="o">=</span><span class="s2">"reflect"</span><span class="p">,</span>
<span class="n">power</span><span class="o">=</span><span class="n">power</span><span class="p">,</span>
<span class="p">)</span>
<span class="k">return</span> <span class="n">spectrogram</span><span class="p">(</span><span class="n">waveform</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">plot_pitch</span><span class="p">(</span><span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span><span class="p">,</span> <span class="n">pitch</span><span class="p">):</span>
<span class="n">figure</span><span class="p">,</span> <span class="n">axis</span> <span class="o">=</span> <span class="n">plt</span><span class="o">.</span><span class="n">subplots</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="mi">1</span><span class="p">)</span>
<span class="n">axis</span><span class="o">.</span><span class="n">set_title</span><span class="p">(</span><span class="s2">"Pitch Feature"</span><span class="p">)</span>
<span class="n">axis</span><span class="o">.</span><span class="n">grid</span><span class="p">(</span><span class="kc">True</span><span class="p">)</span>
<span class="n">end_time</span> <span class="o">=</span> <span class="n">waveform</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">/</span> <span class="n">sample_rate</span>
<span class="n">time_axis</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">end_time</span><span class="p">,</span> <span class="n">waveform</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">])</span>
<span class="n">axis</span><span class="o">.</span><span class="n">plot</span><span class="p">(</span><span class="n">time_axis</span><span class="p">,</span> <span class="n">waveform</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="n">linewidth</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">color</span><span class="o">=</span><span class="s1">'gray'</span><span class="p">,</span> <span class="n">alpha</span><span class="o">=</span><span class="mf">0.3</span><span class="p">)</span>
<span class="n">axis2</span> <span class="o">=</span> <span class="n">axis</span><span class="o">.</span><span class="n">twinx</span><span class="p">()</span>
<span class="n">time_axis</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">end_time</span><span class="p">,</span> <span class="n">pitch</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">])</span>
<span class="n">ln2</span> <span class="o">=</span> <span class="n">axis2</span><span class="o">.</span><span class="n">plot</span><span class="p">(</span>
<span class="n">time_axis</span><span class="p">,</span> <span class="n">pitch</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="n">linewidth</span><span class="o">=</span><span class="mi">2</span><span class="p">,</span> <span class="n">label</span><span class="o">=</span><span class="s1">'Pitch'</span><span class="p">,</span> <span class="n">color</span><span class="o">=</span><span class="s1">'green'</span><span class="p">)</span>
<span class="n">axis2</span><span class="o">.</span><span class="n">legend</span><span class="p">(</span><span class="n">loc</span><span class="o">=</span><span class="mi">0</span><span class="p">)</span>
<span class="n">plt</span><span class="o">.</span><span class="n">show</span><span class="p">(</span><span class="n">block</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">plot_kaldi_pitch</span><span class="p">(</span><span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span><span class="p">,</span> <span class="n">pitch</span><span class="p">,</span> <span class="n">nfcc</span><span class="p">):</span>
<span class="n">figure</span><span class="p">,</span> <span class="n">axis</span> <span class="o">=</span> <span class="n">plt</span><span class="o">.</span><span class="n">subplots</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="mi">1</span><span class="p">)</span>
<span class="n">axis</span><span class="o">.</span><span class="n">set_title</span><span class="p">(</span><span class="s2">"Kaldi Pitch Feature"</span><span class="p">)</span>
<span class="n">axis</span><span class="o">.</span><span class="n">grid</span><span class="p">(</span><span class="kc">True</span><span class="p">)</span>
<span class="n">end_time</span> <span class="o">=</span> <span class="n">waveform</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">/</span> <span class="n">sample_rate</span>
<span class="n">time_axis</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">end_time</span><span class="p">,</span> <span class="n">waveform</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">])</span>
<span class="n">axis</span><span class="o">.</span><span class="n">plot</span><span class="p">(</span><span class="n">time_axis</span><span class="p">,</span> <span class="n">waveform</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="n">linewidth</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">color</span><span class="o">=</span><span class="s1">'gray'</span><span class="p">,</span> <span class="n">alpha</span><span class="o">=</span><span class="mf">0.3</span><span class="p">)</span>
<span class="n">time_axis</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">end_time</span><span class="p">,</span> <span class="n">pitch</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">])</span>
<span class="n">ln1</span> <span class="o">=</span> <span class="n">axis</span><span class="o">.</span><span class="n">plot</span><span class="p">(</span><span class="n">time_axis</span><span class="p">,</span> <span class="n">pitch</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="n">linewidth</span><span class="o">=</span><span class="mi">2</span><span class="p">,</span> <span class="n">label</span><span class="o">=</span><span class="s1">'Pitch'</span><span class="p">,</span> <span class="n">color</span><span class="o">=</span><span class="s1">'green'</span><span class="p">)</span>
<span class="n">axis</span><span class="o">.</span><span class="n">set_ylim</span><span class="p">((</span><span class="o">-</span><span class="mf">1.3</span><span class="p">,</span> <span class="mf">1.3</span><span class="p">))</span>
<span class="n">axis2</span> <span class="o">=</span> <span class="n">axis</span><span class="o">.</span><span class="n">twinx</span><span class="p">()</span>
<span class="n">time_axis</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">end_time</span><span class="p">,</span> <span class="n">nfcc</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">1</span><span class="p">])</span>
<span class="n">ln2</span> <span class="o">=</span> <span class="n">axis2</span><span class="o">.</span><span class="n">plot</span><span class="p">(</span>
<span class="n">time_axis</span><span class="p">,</span> <span class="n">nfcc</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="n">linewidth</span><span class="o">=</span><span class="mi">2</span><span class="p">,</span> <span class="n">label</span><span class="o">=</span><span class="s1">'NFCC'</span><span class="p">,</span> <span class="n">color</span><span class="o">=</span><span class="s1">'blue'</span><span class="p">,</span> <span class="n">linestyle</span><span class="o">=</span><span class="s1">'--'</span><span class="p">)</span>
<span class="n">lns</span> <span class="o">=</span> <span class="n">ln1</span> <span class="o">+</span> <span class="n">ln2</span>
<span class="n">labels</span> <span class="o">=</span> <span class="p">[</span><span class="n">l</span><span class="o">.</span><span class="n">get_label</span><span class="p">()</span> <span class="k">for</span> <span class="n">l</span> <span class="ow">in</span> <span class="n">lns</span><span class="p">]</span>
<span class="n">axis</span><span class="o">.</span><span class="n">legend</span><span class="p">(</span><span class="n">lns</span><span class="p">,</span> <span class="n">labels</span><span class="p">,</span> <span class="n">loc</span><span class="o">=</span><span class="mi">0</span><span class="p">)</span>
<span class="n">plt</span><span class="o">.</span><span class="n">show</span><span class="p">(</span><span class="n">block</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
<span class="n">DEFAULT_OFFSET</span> <span class="o">=</span> <span class="mi">201</span>
<span class="n">SWEEP_MAX_SAMPLE_RATE</span> <span class="o">=</span> <span class="mi">48000</span>
<span class="n">DEFAULT_LOWPASS_FILTER_WIDTH</span> <span class="o">=</span> <span class="mi">6</span>
<span class="n">DEFAULT_ROLLOFF</span> <span class="o">=</span> <span class="mf">0.99</span>
<span class="n">DEFAULT_RESAMPLING_METHOD</span> <span class="o">=</span> <span class="s1">'sinc_interpolation'</span>
<span class="k">def</span> <span class="nf">_get_log_freq</span><span class="p">(</span><span class="n">sample_rate</span><span class="p">,</span> <span class="n">max_sweep_rate</span><span class="p">,</span> <span class="n">offset</span><span class="p">):</span>
<span class="sd">"""Get freqs evenly spaced out in log-scale, between [0, max_sweep_rate // 2]</span>
<span class="sd"> offset is used to avoid negative infinity `log(offset + x)`.</span>
<span class="sd"> """</span>
<span class="n">half</span> <span class="o">=</span> <span class="n">sample_rate</span> <span class="o">//</span> <span class="mi">2</span>
<span class="n">start</span><span class="p">,</span> <span class="n">stop</span> <span class="o">=</span> <span class="n">math</span><span class="o">.</span><span class="n">log</span><span class="p">(</span><span class="n">offset</span><span class="p">),</span> <span class="n">math</span><span class="o">.</span><span class="n">log</span><span class="p">(</span><span class="n">offset</span> <span class="o">+</span> <span class="n">max_sweep_rate</span> <span class="o">//</span> <span class="mi">2</span><span class="p">)</span>
<span class="k">return</span> <span class="n">torch</span><span class="o">.</span><span class="n">exp</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="n">start</span><span class="p">,</span> <span class="n">stop</span><span class="p">,</span> <span class="n">sample_rate</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">torch</span><span class="o">.</span><span class="n">double</span><span class="p">))</span> <span class="o">-</span> <span class="n">offset</span>
<span class="k">def</span> <span class="nf">_get_inverse_log_freq</span><span class="p">(</span><span class="n">freq</span><span class="p">,</span> <span class="n">sample_rate</span><span class="p">,</span> <span class="n">offset</span><span class="p">):</span>
<span class="sd">"""Find the time where the given frequency is given by _get_log_freq"""</span>
<span class="n">half</span> <span class="o">=</span> <span class="n">sample_rate</span> <span class="o">//</span> <span class="mi">2</span>
<span class="k">return</span> <span class="n">sample_rate</span> <span class="o">*</span> <span class="p">(</span><span class="n">math</span><span class="o">.</span><span class="n">log</span><span class="p">(</span><span class="mi">1</span> <span class="o">+</span> <span class="n">freq</span> <span class="o">/</span> <span class="n">offset</span><span class="p">)</span> <span class="o">/</span> <span class="n">math</span><span class="o">.</span><span class="n">log</span><span class="p">(</span><span class="mi">1</span> <span class="o">+</span> <span class="n">half</span> <span class="o">/</span> <span class="n">offset</span><span class="p">))</span>
<span class="k">def</span> <span class="nf">_get_freq_ticks</span><span class="p">(</span><span class="n">sample_rate</span><span class="p">,</span> <span class="n">offset</span><span class="p">,</span> <span class="n">f_max</span><span class="p">):</span>
<span class="c1"># Given the original sample rate used for generating the sweep,</span>
<span class="c1"># find the x-axis value where the log-scale major frequency values fall in</span>
<span class="n">time</span><span class="p">,</span> <span class="n">freq</span> <span class="o">=</span> <span class="p">[],</span> <span class="p">[]</span>
<span class="k">for</span> <span class="n">exp</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="mi">2</span><span class="p">,</span> <span class="mi">5</span><span class="p">):</span>
<span class="k">for</span> <span class="n">v</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="mi">10</span><span class="p">):</span>
<span class="n">f</span> <span class="o">=</span> <span class="n">v</span> <span class="o">*</span> <span class="mi">10</span> <span class="o">**</span> <span class="n">exp</span>
<span class="k">if</span> <span class="n">f</span> <span class="o"><</span> <span class="n">sample_rate</span> <span class="o">//</span> <span class="mi">2</span><span class="p">:</span>
<span class="n">t</span> <span class="o">=</span> <span class="n">_get_inverse_log_freq</span><span class="p">(</span><span class="n">f</span><span class="p">,</span> <span class="n">sample_rate</span><span class="p">,</span> <span class="n">offset</span><span class="p">)</span> <span class="o">/</span> <span class="n">sample_rate</span>
<span class="n">time</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">t</span><span class="p">)</span>
<span class="n">freq</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">f</span><span class="p">)</span>
<span class="n">t_max</span> <span class="o">=</span> <span class="n">_get_inverse_log_freq</span><span class="p">(</span><span class="n">f_max</span><span class="p">,</span> <span class="n">sample_rate</span><span class="p">,</span> <span class="n">offset</span><span class="p">)</span> <span class="o">/</span> <span class="n">sample_rate</span>
<span class="n">time</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">t_max</span><span class="p">)</span>
<span class="n">freq</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">f_max</span><span class="p">)</span>
<span class="k">return</span> <span class="n">time</span><span class="p">,</span> <span class="n">freq</span>
<span class="k">def</span> <span class="nf">plot_sweep</span><span class="p">(</span><span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span><span class="p">,</span> <span class="n">title</span><span class="p">,</span> <span class="n">max_sweep_rate</span><span class="o">=</span><span class="n">SWEEP_MAX_SAMPLE_RATE</span><span class="p">,</span> <span class="n">offset</span><span class="o">=</span><span class="n">DEFAULT_OFFSET</span><span class="p">):</span>
<span class="n">x_ticks</span> <span class="o">=</span> <span class="p">[</span><span class="mi">100</span><span class="p">,</span> <span class="mi">500</span><span class="p">,</span> <span class="mi">1000</span><span class="p">,</span> <span class="mi">5000</span><span class="p">,</span> <span class="mi">10000</span><span class="p">,</span> <span class="mi">20000</span><span class="p">,</span> <span class="n">max_sweep_rate</span> <span class="o">//</span> <span class="mi">2</span><span class="p">]</span>
<span class="n">y_ticks</span> <span class="o">=</span> <span class="p">[</span><span class="mi">1000</span><span class="p">,</span> <span class="mi">5000</span><span class="p">,</span> <span class="mi">10000</span><span class="p">,</span> <span class="mi">20000</span><span class="p">,</span> <span class="n">sample_rate</span><span class="o">//</span><span class="mi">2</span><span class="p">]</span>
<span class="n">time</span><span class="p">,</span> <span class="n">freq</span> <span class="o">=</span> <span class="n">_get_freq_ticks</span><span class="p">(</span><span class="n">max_sweep_rate</span><span class="p">,</span> <span class="n">offset</span><span class="p">,</span> <span class="n">sample_rate</span> <span class="o">//</span> <span class="mi">2</span><span class="p">)</span>
<span class="n">freq_x</span> <span class="o">=</span> <span class="p">[</span><span class="n">f</span> <span class="k">if</span> <span class="n">f</span> <span class="ow">in</span> <span class="n">x_ticks</span> <span class="ow">and</span> <span class="n">f</span> <span class="o"><=</span> <span class="n">max_sweep_rate</span> <span class="o">//</span> <span class="mi">2</span> <span class="k">else</span> <span class="kc">None</span> <span class="k">for</span> <span class="n">f</span> <span class="ow">in</span> <span class="n">freq</span><span class="p">]</span>
<span class="n">freq_y</span> <span class="o">=</span> <span class="p">[</span><span class="n">f</span> <span class="k">for</span> <span class="n">f</span> <span class="ow">in</span> <span class="n">freq</span> <span class="k">if</span> <span class="n">f</span> <span class="o">>=</span> <span class="mi">1000</span> <span class="ow">and</span> <span class="n">f</span> <span class="ow">in</span> <span class="n">y_ticks</span> <span class="ow">and</span> <span class="n">f</span> <span class="o"><=</span> <span class="n">sample_rate</span> <span class="o">//</span> <span class="mi">2</span><span class="p">]</span>
<span class="n">figure</span><span class="p">,</span> <span class="n">axis</span> <span class="o">=</span> <span class="n">plt</span><span class="o">.</span><span class="n">subplots</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="mi">1</span><span class="p">)</span>
<span class="n">axis</span><span class="o">.</span><span class="n">specgram</span><span class="p">(</span><span class="n">waveform</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span><span class="o">.</span><span class="n">numpy</span><span class="p">(),</span> <span class="n">Fs</span><span class="o">=</span><span class="n">sample_rate</span><span class="p">)</span>
<span class="n">plt</span><span class="o">.</span><span class="n">xticks</span><span class="p">(</span><span class="n">time</span><span class="p">,</span> <span class="n">freq_x</span><span class="p">)</span>
<span class="n">plt</span><span class="o">.</span><span class="n">yticks</span><span class="p">(</span><span class="n">freq_y</span><span class="p">,</span> <span class="n">freq_y</span><span class="p">)</span>
<span class="n">axis</span><span class="o">.</span><span class="n">set_xlabel</span><span class="p">(</span><span class="s1">'Original Signal Frequency (Hz, log scale)'</span><span class="p">)</span>
<span class="n">axis</span><span class="o">.</span><span class="n">set_ylabel</span><span class="p">(</span><span class="s1">'Waveform Frequency (Hz)'</span><span class="p">)</span>
<span class="n">axis</span><span class="o">.</span><span class="n">xaxis</span><span class="o">.</span><span class="n">grid</span><span class="p">(</span><span class="kc">True</span><span class="p">,</span> <span class="n">alpha</span><span class="o">=</span><span class="mf">0.67</span><span class="p">)</span>
<span class="n">axis</span><span class="o">.</span><span class="n">yaxis</span><span class="o">.</span><span class="n">grid</span><span class="p">(</span><span class="kc">True</span><span class="p">,</span> <span class="n">alpha</span><span class="o">=</span><span class="mf">0.67</span><span class="p">)</span>
<span class="n">figure</span><span class="o">.</span><span class="n">suptitle</span><span class="p">(</span><span class="sa">f</span><span class="s1">'</span><span class="si">{</span><span class="n">title</span><span class="si">}</span><span class="s1"> (sample rate: </span><span class="si">{</span><span class="n">sample_rate</span><span class="si">}</span><span class="s1"> Hz)'</span><span class="p">)</span>
<span class="n">plt</span><span class="o">.</span><span class="n">show</span><span class="p">(</span><span class="n">block</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">get_sine_sweep</span><span class="p">(</span><span class="n">sample_rate</span><span class="p">,</span> <span class="n">offset</span><span class="o">=</span><span class="n">DEFAULT_OFFSET</span><span class="p">):</span>
<span class="n">max_sweep_rate</span> <span class="o">=</span> <span class="n">sample_rate</span>
<span class="n">freq</span> <span class="o">=</span> <span class="n">_get_log_freq</span><span class="p">(</span><span class="n">sample_rate</span><span class="p">,</span> <span class="n">max_sweep_rate</span><span class="p">,</span> <span class="n">offset</span><span class="p">)</span>
<span class="n">delta</span> <span class="o">=</span> <span class="mi">2</span> <span class="o">*</span> <span class="n">math</span><span class="o">.</span><span class="n">pi</span> <span class="o">*</span> <span class="n">freq</span> <span class="o">/</span> <span class="n">sample_rate</span>
<span class="n">cummulative</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">cumsum</span><span class="p">(</span><span class="n">delta</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="mi">0</span><span class="p">)</span>
<span class="n">signal</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">sin</span><span class="p">(</span><span class="n">cummulative</span><span class="p">)</span><span class="o">.</span><span class="n">unsqueeze</span><span class="p">(</span><span class="n">dim</span><span class="o">=</span><span class="mi">0</span><span class="p">)</span>
<span class="k">return</span> <span class="n">signal</span>
<span class="k">def</span> <span class="nf">benchmark_resample</span><span class="p">(</span>
<span class="n">method</span><span class="p">,</span>
<span class="n">waveform</span><span class="p">,</span>
<span class="n">sample_rate</span><span class="p">,</span>
<span class="n">resample_rate</span><span class="p">,</span>
<span class="n">lowpass_filter_width</span><span class="o">=</span><span class="n">DEFAULT_LOWPASS_FILTER_WIDTH</span><span class="p">,</span>
<span class="n">rolloff</span><span class="o">=</span><span class="n">DEFAULT_ROLLOFF</span><span class="p">,</span>
<span class="n">resampling_method</span><span class="o">=</span><span class="n">DEFAULT_RESAMPLING_METHOD</span><span class="p">,</span>
<span class="n">beta</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span>
<span class="n">librosa_type</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span>
<span class="n">iters</span><span class="o">=</span><span class="mi">5</span>
<span class="p">):</span>
<span class="k">if</span> <span class="n">method</span> <span class="o">==</span> <span class="s2">"functional"</span><span class="p">:</span>
<span class="n">begin</span> <span class="o">=</span> <span class="n">time</span><span class="o">.</span><span class="n">time</span><span class="p">()</span>
<span class="k">for</span> <span class="n">_</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iters</span><span class="p">):</span>
<span class="n">F</span><span class="o">.</span><span class="n">resample</span><span class="p">(</span><span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span><span class="p">,</span> <span class="n">resample_rate</span><span class="p">,</span> <span class="n">lowpass_filter_width</span><span class="o">=</span><span class="n">lowpass_filter_width</span><span class="p">,</span>
<span class="n">rolloff</span><span class="o">=</span><span class="n">rolloff</span><span class="p">,</span> <span class="n">resampling_method</span><span class="o">=</span><span class="n">resampling_method</span><span class="p">)</span>
<span class="n">elapsed</span> <span class="o">=</span> <span class="n">time</span><span class="o">.</span><span class="n">time</span><span class="p">()</span> <span class="o">-</span> <span class="n">begin</span>
<span class="k">return</span> <span class="n">elapsed</span> <span class="o">/</span> <span class="n">iters</span>
<span class="k">elif</span> <span class="n">method</span> <span class="o">==</span> <span class="s2">"transforms"</span><span class="p">:</span>
<span class="n">resampler</span> <span class="o">=</span> <span class="n">T</span><span class="o">.</span><span class="n">Resample</span><span class="p">(</span><span class="n">sample_rate</span><span class="p">,</span> <span class="n">resample_rate</span><span class="p">,</span> <span class="n">lowpass_filter_width</span><span class="o">=</span><span class="n">lowpass_filter_width</span><span class="p">,</span>
<span class="n">rolloff</span><span class="o">=</span><span class="n">rolloff</span><span class="p">,</span> <span class="n">resampling_method</span><span class="o">=</span><span class="n">resampling_method</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">waveform</span><span class="o">.</span><span class="n">dtype</span><span class="p">)</span>
<span class="n">begin</span> <span class="o">=</span> <span class="n">time</span><span class="o">.</span><span class="n">time</span><span class="p">()</span>
<span class="k">for</span> <span class="n">_</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iters</span><span class="p">):</span>
<span class="n">resampler</span><span class="p">(</span><span class="n">waveform</span><span class="p">)</span>
<span class="n">elapsed</span> <span class="o">=</span> <span class="n">time</span><span class="o">.</span><span class="n">time</span><span class="p">()</span> <span class="o">-</span> <span class="n">begin</span>
<span class="k">return</span> <span class="n">elapsed</span> <span class="o">/</span> <span class="n">iters</span>
<span class="k">elif</span> <span class="n">method</span> <span class="o">==</span> <span class="s2">"librosa"</span><span class="p">:</span>
<span class="n">waveform_np</span> <span class="o">=</span> <span class="n">waveform</span><span class="o">.</span><span class="n">squeeze</span><span class="p">()</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span>
<span class="n">begin</span> <span class="o">=</span> <span class="n">time</span><span class="o">.</span><span class="n">time</span><span class="p">()</span>
<span class="k">for</span> <span class="n">_</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iters</span><span class="p">):</span>
<span class="n">librosa</span><span class="o">.</span><span class="n">resample</span><span class="p">(</span><span class="n">waveform_np</span><span class="p">,</span> <span class="n">sample_rate</span><span class="p">,</span> <span class="n">resample_rate</span><span class="p">,</span> <span class="n">res_type</span><span class="o">=</span><span class="n">librosa_type</span><span class="p">)</span>
<span class="n">elapsed</span> <span class="o">=</span> <span class="n">time</span><span class="o">.</span><span class="n">time</span><span class="p">()</span> <span class="o">-</span> <span class="n">begin</span>
<span class="k">return</span> <span class="n">elapsed</span> <span class="o">/</span> <span class="n">iters</span>
</pre></div>
</div>
</div>
</div>
<div class="section" id="audio-i-o">
<h1>Audio I/O<a class="headerlink" href="#audio-i-o" title="Permalink to this headline">¶</a></h1>
<p>torchaudio integrates <code class="docutils literal notranslate"><span class="pre">libsox</span></code> and provides a rich set of audio I/O.</p>
<div class="section" id="quering-audio-metadata">
<h2>Quering audio metadata<a class="headerlink" href="#quering-audio-metadata" title="Permalink to this headline">¶</a></h2>
<p><code class="docutils literal notranslate"><span class="pre">torchaudio.info</span></code> function fetches metadata of audio. You can provide
a path-like object or file-like object.</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">metadata</span> <span class="o">=</span> <span class="n">torchaudio</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="n">SAMPLE_WAV_PATH</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="n">metadata</span><span class="p">)</span>
</pre></div>
</div>
<p class="sphx-glr-script-out">Out:</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>AudioMetaData(sample_rate=44100, num_frames=109368, num_channels=2, bits_per_sample=16, encoding=PCM_S)
</pre></div>
</div>
<p>Where</p>
<ul class="simple">
<li><code class="docutils literal notranslate"><span class="pre">sample_rate</span></code> is the sampling rate of the audio</li>
<li><code class="docutils literal notranslate"><span class="pre">num_channels</span></code> is the number of channels</li>
<li><code class="docutils literal notranslate"><span class="pre">num_frames</span></code> is the number of frames per channel</li>
<li><code class="docutils literal notranslate"><span class="pre">bits_per_sample</span></code> is bit depth</li>
<li><code class="docutils literal notranslate"><span class="pre">encoding</span></code> is the sample coding format</li>
</ul>
<p>The values <code class="docutils literal notranslate"><span class="pre">encoding</span></code> can take are one of the following</p>
<ul class="simple">
<li><code class="docutils literal notranslate"><span class="pre">"PCM_S"</span></code>: Signed integer linear PCM</li>
<li><code class="docutils literal notranslate"><span class="pre">"PCM_U"</span></code>: Unsigned integer linear PCM</li>
<li><code class="docutils literal notranslate"><span class="pre">"PCM_F"</span></code>: Floating point linear PCM</li>
<li><code class="docutils literal notranslate"><span class="pre">"FLAC"</span></code>: Flac, <a class="reference external" href="https://xiph.org/flac/">Free Lossless Audio
Codec</a></li>
<li><code class="docutils literal notranslate"><span class="pre">"ULAW"</span></code>: Mu-law,
[<a class="reference external" href="https://en.wikipedia.org/wiki/%CE%9C-law_algorithm">wikipedia</a>]</li>
<li><code class="docutils literal notranslate"><span class="pre">"ALAW"</span></code>: A-law
[<a class="reference external" href="https://en.wikipedia.org/wiki/A-law_algorithm">wikipedia</a>]</li>
<li><code class="docutils literal notranslate"><span class="pre">"MP3"</span></code> : MP3, MPEG-1 Audio Layer III</li>
<li><code class="docutils literal notranslate"><span class="pre">"VORBIS"</span></code>: OGG Vorbis [<a class="reference external" href="https://xiph.org/vorbis/">xiph.org</a>]</li>
<li><code class="docutils literal notranslate"><span class="pre">"AMR_NB"</span></code>: Adaptive Multi-Rate
[<a class="reference external" href="https://en.wikipedia.org/wiki/Adaptive_Multi-Rate_audio_codec">wikipedia</a>]</li>
<li><code class="docutils literal notranslate"><span class="pre">"AMR_WB"</span></code>: Adaptive Multi-Rate Wideband
[<a class="reference external" href="https://en.wikipedia.org/wiki/Adaptive_Multi-Rate_Wideband">wikipedia</a>]</li>
<li><code class="docutils literal notranslate"><span class="pre">"OPUS"</span></code>: Opus [<a class="reference external" href="https://opus-codec.org/">opus-codec.org</a>]</li>
<li><code class="docutils literal notranslate"><span class="pre">"GSM"</span></code>: GSM-FR
[<a class="reference external" href="https://en.wikipedia.org/wiki/Full_Rate">wikipedia</a>]</li>
<li><code class="docutils literal notranslate"><span class="pre">"UNKNOWN"</span></code> None of avobe</li>
</ul>
<p><strong>Note</strong></p>
<ul class="simple">
<li><code class="docutils literal notranslate"><span class="pre">bits_per_sample</span></code> can be <code class="docutils literal notranslate"><span class="pre">0</span></code> for formats with compression and/or
variable bit rate. (such as mp3)</li>
<li><code class="docutils literal notranslate"><span class="pre">num_frames</span></code> can be <code class="docutils literal notranslate"><span class="pre">0</span></code> for GSM-FR format.</li>
</ul>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">metadata</span> <span class="o">=</span> <span class="n">torchaudio</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="n">SAMPLE_MP3_PATH</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="n">metadata</span><span class="p">)</span>
<span class="n">metadata</span> <span class="o">=</span> <span class="n">torchaudio</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="n">SAMPLE_GSM_PATH</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="n">metadata</span><span class="p">)</span>
</pre></div>
</div>
<p class="sphx-glr-script-out">Out:</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>AudioMetaData(sample_rate=44100, num_frames=110559, num_channels=2, bits_per_sample=0, encoding=MP3)
AudioMetaData(sample_rate=8000, num_frames=0, num_channels=1, bits_per_sample=0, encoding=GSM)
</pre></div>
</div>
<div class="section" id="querying-file-like-object">
<h3>Querying file-like object<a class="headerlink" href="#querying-file-like-object" title="Permalink to this headline">¶</a></h3>
<p><code class="docutils literal notranslate"><span class="pre">info</span></code> function works on file-like object as well.</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nb">print</span><span class="p">(</span><span class="s2">"Source:"</span><span class="p">,</span> <span class="n">SAMPLE_WAV_URL</span><span class="p">)</span>
<span class="k">with</span> <span class="n">requests</span><span class="o">.</span><span class="n">get</span><span class="p">(</span><span class="n">SAMPLE_WAV_URL</span><span class="p">,</span> <span class="n">stream</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span> <span class="k">as</span> <span class="n">response</span><span class="p">:</span>
<span class="n">metadata</span> <span class="o">=</span> <span class="n">torchaudio</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="n">response</span><span class="o">.</span><span class="n">raw</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="n">metadata</span><span class="p">)</span>
</pre></div>
</div>
<p class="sphx-glr-script-out">Out:</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Source: https://pytorch-tutorial-assets.s3.amazonaws.com/steam-train-whistle-daniel_simon.wav
AudioMetaData(sample_rate=44100, num_frames=109368, num_channels=2, bits_per_sample=16, encoding=PCM_S)
</pre></div>
</div>
<p><strong>Note</strong> When passing file-like object, <code class="docutils literal notranslate"><span class="pre">info</span></code> function does not read
all the data, instead it only reads the beginning portion of data.
Therefore, depending on the audio format, it cannot get the correct
metadata, including the format itself. The following example illustrates
this.</p>
<ul class="simple">
<li>Use <code class="docutils literal notranslate"><span class="pre">format</span></code> argument to tell what audio format it is.</li>
<li>The returned metadata has <code class="docutils literal notranslate"><span class="pre">num_frames</span> <span class="pre">=</span> <span class="pre">0</span></code></li>
</ul>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nb">print</span><span class="p">(</span><span class="s2">"Source:"</span><span class="p">,</span> <span class="n">SAMPLE_MP3_URL</span><span class="p">)</span>
<span class="k">with</span> <span class="n">requests</span><span class="o">.</span><span class="n">get</span><span class="p">(</span><span class="n">SAMPLE_MP3_URL</span><span class="p">,</span> <span class="n">stream</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span> <span class="k">as</span> <span class="n">response</span><span class="p">:</span>
<span class="n">metadata</span> <span class="o">=</span> <span class="n">torchaudio</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="n">response</span><span class="o">.</span><span class="n">raw</span><span class="p">,</span> <span class="nb">format</span><span class="o">=</span><span class="s2">"mp3"</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="sa">f</span><span class="s2">"Fetched </span><span class="si">{</span><span class="n">response</span><span class="o">.</span><span class="n">raw</span><span class="o">.</span><span class="n">tell</span><span class="p">()</span><span class="si">}</span><span class="s2"> bytes."</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="n">metadata</span><span class="p">)</span>
</pre></div>
</div>
<p class="sphx-glr-script-out">Out:</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Source: https://pytorch-tutorial-assets.s3.amazonaws.com/steam-train-whistle-daniel_simon.mp3
Fetched 4096 bytes.
AudioMetaData(sample_rate=44100, num_frames=0, num_channels=2, bits_per_sample=0, encoding=MP3)
</pre></div>
</div>
</div>
</div>
<div class="section" id="loading-audio-data-into-tensor">
<h2>Loading audio data into Tensor<a class="headerlink" href="#loading-audio-data-into-tensor" title="Permalink to this headline">¶</a></h2>
<p>To load audio data, you can use <code class="docutils literal notranslate"><span class="pre">torchaudio.load</span></code>.</p>
<p>This function accepts path-like object and file-like object.</p>
<p>The returned value is a tuple of waveform (<code class="docutils literal notranslate"><span class="pre">Tensor</span></code>) and sample rate
(<code class="docutils literal notranslate"><span class="pre">int</span></code>).</p>
<p>By default, the resulting tensor object has <code class="docutils literal notranslate"><span class="pre">dtype=torch.float32</span></code> and
its value range is normalized within <code class="docutils literal notranslate"><span class="pre">[-1.0,</span> <span class="pre">1.0]</span></code>.</p>
<p>For the list of supported format, please refer to <a class="reference external" href="https://pytorch.org/audio">the torchaudio
documentation</a>.</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span> <span class="o">=</span> <span class="n">torchaudio</span><span class="o">.</span><span class="n">load</span><span class="p">(</span><span class="n">SAMPLE_WAV_SPEECH_PATH</span><span class="p">)</span>
<span class="n">print_stats</span><span class="p">(</span><span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span><span class="o">=</span><span class="n">sample_rate</span><span class="p">)</span>
<span class="n">plot_waveform</span><span class="p">(</span><span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span><span class="p">)</span>
<span class="n">plot_specgram</span><span class="p">(</span><span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span><span class="p">)</span>
<span class="n">play_audio</span><span class="p">(</span><span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span><span class="p">)</span>
</pre></div>
</div>
<ul class="sphx-glr-horizontal">
<li><img alt="../_images/sphx_glr_audio_preprocessing_tutorial_001.png" class="sphx-glr-multi-img first" src="../_images/sphx_glr_audio_preprocessing_tutorial_001.png" />
</li>
<li><img alt="../_images/sphx_glr_audio_preprocessing_tutorial_002.png" class="sphx-glr-multi-img first" src="../_images/sphx_glr_audio_preprocessing_tutorial_002.png" />
</li>
</ul>
<p class="sphx-glr-script-out">Out:</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Sample Rate: 16000
Shape: (1, 54400)
Dtype: torch.float32
- Max: 0.668
- Min: -1.000
- Mean: 0.000
- Std Dev: 0.122
tensor([[0.0183, 0.0180, 0.0180, ..., 0.0018, 0.0019, 0.0032]])
<IPython.lib.display.Audio object>
</pre></div>
</div>
<div class="section" id="loading-from-file-like-object">
<h3>Loading from file-like object<a class="headerlink" href="#loading-from-file-like-object" title="Permalink to this headline">¶</a></h3>
<p><code class="docutils literal notranslate"><span class="pre">torchaudio</span></code>’s I/O functions now support file-like object. This
allows to fetch audio data and decode at the same time from the location
other than local file system. The following examples illustrates this.</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># Load audio data as HTTP request</span>
<span class="k">with</span> <span class="n">requests</span><span class="o">.</span><span class="n">get</span><span class="p">(</span><span class="n">SAMPLE_WAV_SPEECH_URL</span><span class="p">,</span> <span class="n">stream</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span> <span class="k">as</span> <span class="n">response</span><span class="p">:</span>
<span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span> <span class="o">=</span> <span class="n">torchaudio</span><span class="o">.</span><span class="n">load</span><span class="p">(</span><span class="n">response</span><span class="o">.</span><span class="n">raw</span><span class="p">)</span>
<span class="n">plot_specgram</span><span class="p">(</span><span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span><span class="p">,</span> <span class="n">title</span><span class="o">=</span><span class="s2">"HTTP datasource"</span><span class="p">)</span>
<span class="c1"># Load audio from tar file</span>
<span class="k">with</span> <span class="n">tarfile</span><span class="o">.</span><span class="n">open</span><span class="p">(</span><span class="n">SAMPLE_TAR_PATH</span><span class="p">,</span> <span class="n">mode</span><span class="o">=</span><span class="s1">'r'</span><span class="p">)</span> <span class="k">as</span> <span class="n">tarfile_</span><span class="p">:</span>
<span class="n">fileobj</span> <span class="o">=</span> <span class="n">tarfile_</span><span class="o">.</span><span class="n">extractfile</span><span class="p">(</span><span class="n">SAMPLE_TAR_ITEM</span><span class="p">)</span>
<span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span> <span class="o">=</span> <span class="n">torchaudio</span><span class="o">.</span><span class="n">load</span><span class="p">(</span><span class="n">fileobj</span><span class="p">)</span>
<span class="n">plot_specgram</span><span class="p">(</span><span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span><span class="p">,</span> <span class="n">title</span><span class="o">=</span><span class="s2">"TAR file"</span><span class="p">)</span>
<span class="c1"># Load audio from S3</span>
<span class="n">client</span> <span class="o">=</span> <span class="n">boto3</span><span class="o">.</span><span class="n">client</span><span class="p">(</span><span class="s1">'s3'</span><span class="p">,</span> <span class="n">config</span><span class="o">=</span><span class="n">Config</span><span class="p">(</span><span class="n">signature_version</span><span class="o">=</span><span class="n">UNSIGNED</span><span class="p">))</span>
<span class="n">response</span> <span class="o">=</span> <span class="n">client</span><span class="o">.</span><span class="n">get_object</span><span class="p">(</span><span class="n">Bucket</span><span class="o">=</span><span class="n">S3_BUCKET</span><span class="p">,</span> <span class="n">Key</span><span class="o">=</span><span class="n">S3_KEY</span><span class="p">)</span>
<span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span> <span class="o">=</span> <span class="n">torchaudio</span><span class="o">.</span><span class="n">load</span><span class="p">(</span><span class="n">response</span><span class="p">[</span><span class="s1">'Body'</span><span class="p">])</span>
<span class="n">plot_specgram</span><span class="p">(</span><span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span><span class="p">,</span> <span class="n">title</span><span class="o">=</span><span class="s2">"From S3"</span><span class="p">)</span>
</pre></div>
</div>
<ul class="sphx-glr-horizontal">
<li><img alt="../_images/sphx_glr_audio_preprocessing_tutorial_003.png" class="sphx-glr-multi-img first" src="../_images/sphx_glr_audio_preprocessing_tutorial_003.png" />
</li>
<li><img alt="../_images/sphx_glr_audio_preprocessing_tutorial_004.png" class="sphx-glr-multi-img first" src="../_images/sphx_glr_audio_preprocessing_tutorial_004.png" />
</li>
<li><img alt="../_images/sphx_glr_audio_preprocessing_tutorial_005.png" class="sphx-glr-multi-img first" src="../_images/sphx_glr_audio_preprocessing_tutorial_005.png" />
</li>
</ul>
</div>
<div class="section" id="tips-on-slicing">
<h3>Tips on slicing<a class="headerlink" href="#tips-on-slicing" title="Permalink to this headline">¶</a></h3>
<p>Providing <code class="docutils literal notranslate"><span class="pre">num_frames</span></code> and <code class="docutils literal notranslate"><span class="pre">frame_offset</span></code> arguments will slice the
resulting Tensor object while decoding.</p>
<p>The same result can be achieved using the regular Tensor slicing,
(i.e. <code class="docutils literal notranslate"><span class="pre">waveform[:,</span> <span class="pre">frame_offset:frame_offset+num_frames]</span></code>) however,
providing <code class="docutils literal notranslate"><span class="pre">num_frames</span></code> and <code class="docutils literal notranslate"><span class="pre">frame_offset</span></code> arguments is more
efficient.</p>
<p>This is because the function will stop data acquisition and decoding
once it finishes decoding the requested frames. This is advantageous
when the audio data are transfered via network as the data transfer will
stop as soon as the necessary amount of data is fetched.</p>
<p>The following example illustrates this;</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># Illustration of two different decoding methods.</span>
<span class="c1"># The first one will fetch all the data and decode them, while</span>
<span class="c1"># the second one will stop fetching data once it completes decoding.</span>
<span class="c1"># The resulting waveforms are identical.</span>
<span class="n">frame_offset</span><span class="p">,</span> <span class="n">num_frames</span> <span class="o">=</span> <span class="mi">16000</span><span class="p">,</span> <span class="mi">16000</span> <span class="c1"># Fetch and decode the 1 - 2 seconds</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">"Fetching all the data..."</span><span class="p">)</span>
<span class="k">with</span> <span class="n">requests</span><span class="o">.</span><span class="n">get</span><span class="p">(</span><span class="n">SAMPLE_WAV_SPEECH_URL</span><span class="p">,</span> <span class="n">stream</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span> <span class="k">as</span> <span class="n">response</span><span class="p">:</span>
<span class="n">waveform1</span><span class="p">,</span> <span class="n">sample_rate1</span> <span class="o">=</span> <span class="n">torchaudio</span><span class="o">.</span><span class="n">load</span><span class="p">(</span><span class="n">response</span><span class="o">.</span><span class="n">raw</span><span class="p">)</span>
<span class="n">waveform1</span> <span class="o">=</span> <span class="n">waveform1</span><span class="p">[:,</span> <span class="n">frame_offset</span><span class="p">:</span><span class="n">frame_offset</span><span class="o">+</span><span class="n">num_frames</span><span class="p">]</span>
<span class="nb">print</span><span class="p">(</span><span class="sa">f</span><span class="s2">" - Fetched </span><span class="si">{</span><span class="n">response</span><span class="o">.</span><span class="n">raw</span><span class="o">.</span><span class="n">tell</span><span class="p">()</span><span class="si">}</span><span class="s2"> bytes"</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">"Fetching until the requested frames are available..."</span><span class="p">)</span>
<span class="k">with</span> <span class="n">requests</span><span class="o">.</span><span class="n">get</span><span class="p">(</span><span class="n">SAMPLE_WAV_SPEECH_URL</span><span class="p">,</span> <span class="n">stream</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span> <span class="k">as</span> <span class="n">response</span><span class="p">:</span>
<span class="n">waveform2</span><span class="p">,</span> <span class="n">sample_rate2</span> <span class="o">=</span> <span class="n">torchaudio</span><span class="o">.</span><span class="n">load</span><span class="p">(</span>
<span class="n">response</span><span class="o">.</span><span class="n">raw</span><span class="p">,</span> <span class="n">frame_offset</span><span class="o">=</span><span class="n">frame_offset</span><span class="p">,</span> <span class="n">num_frames</span><span class="o">=</span><span class="n">num_frames</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="sa">f</span><span class="s2">" - Fetched </span><span class="si">{</span><span class="n">response</span><span class="o">.</span><span class="n">raw</span><span class="o">.</span><span class="n">tell</span><span class="p">()</span><span class="si">}</span><span class="s2"> bytes"</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">"Checking the resulting waveform ... "</span><span class="p">,</span> <span class="n">end</span><span class="o">=</span><span class="s2">""</span><span class="p">)</span>
<span class="k">assert</span> <span class="p">(</span><span class="n">waveform1</span> <span class="o">==</span> <span class="n">waveform2</span><span class="p">)</span><span class="o">.</span><span class="n">all</span><span class="p">()</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">"matched!"</span><span class="p">)</span>
</pre></div>
</div>
<p class="sphx-glr-script-out">Out:</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Fetching all the data...
- Fetched 108844 bytes
Fetching until the requested frames are available...
- Fetched 65580 bytes
Checking the resulting waveform ... matched!
</pre></div>
</div>
</div>
</div>
<div class="section" id="saving-audio-to-file">
<h2>Saving audio to file<a class="headerlink" href="#saving-audio-to-file" title="Permalink to this headline">¶</a></h2>
<p>To save audio data in the formats intepretable by common applications,
you can use <code class="docutils literal notranslate"><span class="pre">torchaudio.save</span></code>.</p>
<p>This function accepts path-like object and file-like object.</p>
<p>When passing file-like object, you also need to provide <code class="docutils literal notranslate"><span class="pre">format</span></code>
argument so that the function knows which format it should be using. In
case of path-like object, the function will detemine the format based on
the extension. If you are saving to a file without extension, you need
to provide <code class="docutils literal notranslate"><span class="pre">format</span></code> argument.</p>
<p>When saving as WAV format, the default encoding for <code class="docutils literal notranslate"><span class="pre">float32</span></code> Tensor
is 32-bit floating-point PCM. You can provide <code class="docutils literal notranslate"><span class="pre">encoding</span></code> and
<code class="docutils literal notranslate"><span class="pre">bits_per_sample</span></code> argument to change this. For example, to save data
in 16 bit signed integer PCM, you can do the following.</p>
<p><strong>Note</strong> Saving data in encodings with lower bit depth reduces the
resulting file size but loses precision.</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">waveform</span><span class="p">,</span> <span class="n">sample_rate</span> <span class="o">=</span> <span class="n">get_sample</span><span class="p">()</span>