-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
665 lines (619 loc) · 43.5 KB
/
index.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
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<!-- Meta tags for social media banners, these should be filled in appropriatly as they are your "business card" -->
<!-- Replace the content tag with appropriate information -->
<meta name="description" content="DESCRIPTION META TAG">
<meta property="og:title" content="SOCIAL MEDIA TITLE TAG"/>
<meta property="og:description" content="SOCIAL MEDIA DESCRIPTION TAG TAG"/>
<meta property="og:url" content="URL OF THE WEBSITE"/>
<!-- Path to banner image, should be in the path listed below. Optimal dimenssions are 1200X630-->
<meta property="og:image" content="static/image/your_banner_image.png" />
<meta property="og:image:width" content="1200"/>
<meta property="og:image:height" content="630"/>
<meta name="twitter:title" content="TWITTER BANNER TITLE META TAG">
<meta name="twitter:description" content="TWITTER BANNER DESCRIPTION META TAG">
<!-- Path to banner image, should be in the path listed below. Optimal dimenssions are 1200X600-->
<meta name="twitter:image" content="static/images/your_twitter_banner_image.png">
<meta name="twitter:card" content="summary_large_image">
<!-- Keywords for your paper to be indexed by-->
<meta name="keywords" content="KEYWORDS SHOULD BE PLACED HERE">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title> FINALLY: Fast and Universal Speech Enhancement With Studio-like Quality</title>
<link rel="ico" type="image/x-icon" href="funny_se_image.jpg">
<link href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro"
rel="stylesheet">
<link rel="stylesheet" href="static/css/bulma.min.css">
<link rel="stylesheet" href="static/css/bulma-carousel.min.css">
<link rel="stylesheet" href="static/css/bulma-slider.min.css">
<link rel="stylesheet" href="static/css/fontawesome.all.min.css">
<link rel="stylesheet"
href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css">
<link rel="stylesheet" href="static/css/index.css">
<script src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script>
<script src="https://documentcloud.adobe.com/view-sdk/main.js"></script>
<script defer src="static/js/fontawesome.all.min.js"></script>
<script src="static/js/bulma-carousel.min.js"></script>
<script src="static/js/bulma-slider.min.js"></script>
<script src="static/js/index.js"></script>
</head>
<body>
<section class="hero">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column has-text-centered">
<h1 class="title is-1 publication-title">FINALLY <br> Fast and Universal Speech Enhancement With Studio-like Quality</h1>
<div class="is-size-5 publication-authors">
<!-- Paper authors -->
<span class="author-block">
<a href="https://scholar.google.com/citations?user=Q4w7mjcAAAAJ&hl=en" target="_blank">Nicholas Babaev</a><sup>*</sup>,
</span>
<span class="author-block">
<a href="THIRD AUTHOR PERSONAL LINK" target="_blank">Kirill Tamogashev</a><sup>*</sup>,
</span>
<span class="author-block">
<a href="https://scholar.google.com/citations?hl=ru&user=VfXlrqoAAAAJ" target="_blank">Azat Saginbaev</a>,
</span>
<span class="author-block">
<a href="https://scholar.google.com/citations?user=1opT3A0AAAAJ&hl=en" target="_blank">Ivan Shchekotov</a>,
</span>
<span class="author-block">
<a href="https://scholar.google.com/citations?user=OvIic1kAAAAJ&hl=en" target="_blank">Hanbin Bae</a>,
</span>
<span class="author-block">
<a href="https://scholar.google.co.kr/citations?user=rcCzgZAAAAAJ&hl=en" target="_blank">Hosang Sung</a>,
</span>
<span class="author-block">
<a href="THIRD AUTHOR PERSONAL LINK" target="_blank">WonJun Lee</a>,
</span>
<span class="author-block">
<a href="THIRD AUTHOR PERSONAL LINK" target="_blank">Hoon-Young Cho</a>,
</span>
<span class="author-block">
<a href="https://scholar.google.com/citations?user=jIl4FywAAAAJ&hl=en" target="_blank">Pavel Andreev</a><sup>*</sup>
</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block">Samsung Research <br>NeurIPS 2024</span>
<span class="eql-cntrb"><small><br><sup>*</sup>Equal Contribution. Correspondence to Pavel Andreev (p.andreev@samsung.com)</small></span>
</div>
<div class="column has-text-centered">
<div class="publication-links">
<span class="link-block">
<a href="https://arxiv.org/abs/2410.05920v1" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="ai ai-arxiv"></i>
</span>
<span>arXiv</span>
</a>
</span>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<!-- Paper abstract -->
<section class="section hero is-light">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Abstract</h2>
<div class="content has-text-justified">
<p>
In this paper, we address the challenge of speech enhancement in real-world recordings, which often contain various forms of distortion, such as background noise, reverberation, and microphone artifacts.
We revisit the use of Generative Adversarial Networks (GANs) for speech enhancement and theoretically show that GANs are naturally inclined to seek the point of maximum density within the conditional clean speech distribution, which, as we argue, is essential for speech enhancement task.
We study various feature extractors for perceptual loss to facilitate the stability of adversarial training, developing a methodology for probing the structure of the feature space.
This leads us to integrate WavLM-based perceptual loss into MS-STFT adversarial training pipeline, creating an effective and stable training procedure for the speech enhancement model.
The resulting speech enhancement model, which we refer to as FINALLY, builds upon the HiFi++ architecture, augmented with a WavLM encoder and a novel training pipeline.
Empirical results on various datasets confirm our model's ability to produce clear, high-quality speech at 48 kHz, achieving state-of-the-art performance in the field of speech enhancement.
</p>
</div>
</div>
</div>
</div>
</section>
<!-- End paper abstract -->
<!-- Model architecture -->
<section class="section hero">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-full">
<div class="content">
<h2 class="title is-3"><center>Model Architecure</center></h2>
<div class="level-set has-text-justified">
<p>
The model is a six-component neural network consisting of WavLM-large, SpectralUNet, Upsampler, WaveUNet, SpectralMaskNet and the Upsample WaveUNet modules.
SpectralUNet is responsible for initial preprocessing of audio in the spectral domain using two-dimensional convolutions.
Additionaly, the SSL features, obtained from WavLM-large, are added to spectral ones.
The Upsampler is a HiFi-GAN generator-based module that increases the temporal resolution of the input tensor, mapping it to the waveform domain.
WaveUNet performs post-processing in the waveform domain and improves the output of the Upsampler by incorporating phase information gleaned directly from the raw input waveform.
SpectralMaskNet is applied to perform spectrum-based post-processing and thus, remove any possible artifacts that remained after WaveUNet.
Thus, the model alternates between time and frequency domains, allowing for effective audio restoration.
Finally, the Upsample WaveUNet is a learnable upsampler of the signal sampling rate, consisting of
the WaveUNet with an additional convolutional upsampling block in the decoder that upsamples the temporal resolution by 3 times.
<br>
</div>
<img src="images/architecture.png" alt="Normal and Anomalous Representations" class="center-image">
</div>
</div>
</div>
</div>
</section>
<section class="section hero is-light">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-full">
<div class="content">
<h2 class="title is-3"><center>Real data samples</center></h2>
<div class="level-set has-text-justified">
<p>
In this section we illustrate the performance of FINALLY on read world samples.
As you can hear, our model superbly removes all
existing distortions, producing high quality audio samples
<br>
</div>
</div>
</div>
</div>
</div>
</section>
<!-- Real data demo -->
<section class="hero is-light">
<div id="results-carousel" class="carousel results-carousel">
<div class="item item-video1">
<table>
<tr>
<td style="text-align: center; width: 50%;"><h2 class="subtitle has-text-centered">Input</h2></td>
<td style="text-align: center; width: 50%;"><h2 class="subtitle has-text-centered">Output</h2></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><img align="center" src="images/real_data_input_0.png" class="illustration" onclick="zoom(this)"/></td>
<td style="text-align: center; width: 25%;"><img align="center" src="images/real_data_output_0.png" class="illustration" onclick="zoom(this)"/></td>
</tr>
<tr>
<td style="text-align: center; width: 50%;"><audio controls style="width: 50%;"><source src="samples/real_data/input/0.wav"></audio></td>
<td style="text-align: center; width: 50%;"><audio controls style="width: 50%;"><source src="samples/real_data/output/0.wav"></audio></td>
</table>
</div>
<div class="item item-video2">
<table>
<tr>
<td style="text-align: center; width: 50%;">
<h2 class="subtitle has-text-centered">Input</h2>
</td>
<td style="text-align: center; width: 50%;">
<h2 class="subtitle has-text-centered">Output</h2>
</td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><img align="center" src="images/real_data_input_1.png" class="illustration" onclick="zoom(this)"/></td>
<td style="text-align: center; width: 25%;"><img align="center" src="images/real_data_output_1.png" class="illustration" onclick="zoom(this)"/></td>
</tr>
<tr>
<td style="text-align: center; width: 50%;"><audio controls style="width: 50%;"><source src="samples/real_data/input/1.wav"></audio></td>
<td style="text-align: center; width: 50%;"><audio controls style="width: 50%;"><source src="samples/real_data/output/1.wav"></audio></td>
</tr>
</table>
</div>
<div class="item item-video3">
<table>
<tr>
<td style="text-align: center; width: 50%;">
<h2 class="subtitle has-text-centered">Input</h2>
</td>
<td style="text-align: center; width: 50%;">
<h2 class="subtitle has-text-centered">Output</h2>
</td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><img align="center" src="images/real_data_input_2.png" class="illustration" onclick="zoom(this)"/></td>
<td style="text-align: center; width: 25%;"><img align="center" src="images/real_data_output_2.png" class="illustration" onclick="zoom(this)"/></td>
</tr>
<tr>
<td style="text-align: center; width: 50%;"><audio controls style="width: 50%;"><source src="samples/real_data/input/2.wav"></audio></td>
<td style="text-align: center; width: 50%;"><audio controls style="width: 50%;"><source src="samples/real_data/output/2.wav"></audio></td>
</tr>
</table>
</div>
</div>
</div>
</div>
</section>
<!-- End video carousel -->
<section class="section hero">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-full">
<div class="content">
<h2 class="title is-3"><center>Comparison with UNIVERSE diffusion model</center></h2>
<div class="level-set has-text-justified">
<p>
Additionally, we provide comparison of our model with UNIVERSE on their validation data. Our model is less prone to hallucinations,
while delivering high perceptual quality.
<br>
</div>
</div>
</div>
</div>
</div>
</section>
<!-- Real data demo -->
<section class="hero is-small">
<div id="results-carousel" class="carousel results-carousel">
<div class="item item-video1">
<table>
<tr>
<tr>
<td style="text-align: center; width: 50%;"><h2 class="subtitle has-text-centered">Input</h2></td>
<td style="text-align: center; width: 50%;"><h2 class="subtitle has-text-centered">Ground Truth</h2></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><img align="center" src="images/0015_input.png" class="illustration" onclick="zoom(this)"/></td>
<td style="text-align: center; width: 25%;"><img align="center" src="images/0015_gt.png" class="illustration" onclick="zoom(this)"/></td>
</tr>
<tr>
<td style="text-align: center; width: 50%;"><audio controls style="width: 50%;"><source src="samples/ours_vs_universe/input/0015.wav"></audio></td>
<td style="text-align: center; width: 50%;"><audio controls style="width: 50%;"><source src="samples/ours_vs_universe/gt/0015.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 50%;"><h2 class="subtitle has-text-centered">Finally (<b>ours</b>)</h2></td>
<td style="text-align: center; width: 50%;"><h2 class="subtitle has-text-centered">UNIVERSE</h2></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><img align="center" src="images/0015_ours.png" class="illustration" onclick="zoom(this)"/></td>
<td style="text-align: center; width: 25%;"><img align="center" src="images/0015_universe.png" class="illustration" onclick="zoom(this)"/></td>
</tr>
<tr>
<td style="text-align: center; width: 50%;"><audio controls style="width: 50%;"><source src="samples/ours_vs_universe/ours/0015.wav"></audio></td>
<td style="text-align: center; width: 50%;"><audio controls style="width: 50%;"><source src="samples/ours_vs_universe/universe/0015.wav"></audio></td>
</tr>
</table>
</div>
<div class="item item-video2">
<table>
<tr>
<td style="text-align: center; width: 50%;"><h2 class="subtitle has-text-centered">Input</h2></td>
<td style="text-align: center; width: 50%;"><h2 class="subtitle has-text-centered">Ground Truth</h2></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><img align="center" src="images/0090_input.png" class="illustration" onclick="zoom(this)"/></td>
<td style="text-align: center; width: 25%;"><img align="center" src="images/0090_gt.png" class="illustration" onclick="zoom(this)"/></td>
</tr>
<tr>
<td style="text-align: center; width: 50%;"><audio controls style="width: 50%;"><source src="samples/ours_vs_universe/input/0090.wav"></audio></td>
<td style="text-align: center; width: 50%;"><audio controls style="width: 50%;"><source src="samples/ours_vs_universe/gt/0090.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 50%;"><h2 class="subtitle has-text-centered">Finally (<b>ours</b>)</h2></td>
<td style="text-align: center; width: 50%;"><h2 class="subtitle has-text-centered">UNIVERSE</h2></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><img align="center" src="images/0090_ours.png" class="illustration" onclick="zoom(this)"/></td>
<td style="text-align: center; width: 25%;"><img align="center" src="images/0090_universe.png" class="illustration" onclick="zoom(this)"/></td>
</tr>
<tr>
<td style="text-align: center; width: 50%;"><audio controls style="width: 50%;"><source src="samples/ours_vs_universe/ours/0090.wav"></audio></td>
<td style="text-align: center; width: 50%;"><audio controls style="width: 50%;"><source src="samples/ours_vs_universe/universe/0090.wav"></audio></td>
</tr>
</table>
</div>
<div class="item item-video3">
<table>
<tr>
<td style="text-align: center; width: 50%;"><h2 class="subtitle has-text-centered">Input</h2></td>
<td style="text-align: center; width: 50%;"><h2 class="subtitle has-text-centered">Ground Truth</h2></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><img align="center" src="images/0061_input.png" class="illustration" onclick="zoom(this)"/></td>
<td style="text-align: center; width: 25%;"><img align="center" src="images/0061_gt.png" class="illustration" onclick="zoom(this)"/></td>
</tr>
<tr>
<td style="text-align: center; width: 50%;"><audio controls style="width: 50%;"><source src="samples/ours_vs_universe/input/0061.wav"></audio></td>
<td style="text-align: center; width: 50%;"><audio controls style="width: 50%;"><source src="samples/ours_vs_universe/gt/0061.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 50%;"><h2 class="subtitle has-text-centered">Finally (<b>ours</b>)</h2></td>
<td style="text-align: center; width: 50%;"><h2 class="subtitle has-text-centered">UNIVERSE</h2></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><img align="center" src="images/0061_ours.png" class="illustration" onclick="zoom(this)"/></td>
<td style="text-align: center; width: 25%;"><img align="center" src="images/0061_universe.png" class="illustration" onclick="zoom(this)"/></td>
</tr>
<tr>
<td style="text-align: center; width: 50%;"><audio controls style="width: 50%;"><source src="samples/ours_vs_universe/ours/0061.wav"></audio></td>
<td style="text-align: center; width: 50%;"><audio controls style="width: 50%;"><source src="samples/ours_vs_universe/universe/0061.wav"></audio></td>
</tr>
</table>
</div>
</div>
</div>
</div>
</section>
<!-- End video carousel -->
<section class="section hero is-light">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-full">
<div class="content">
<h2 class="title is-3"><center>Additional Comparison with HiFi-GAN-2 and UNIVERSE</center></h2>
<div class="level-set has-text-justified">
<p>
<br>
</div>
</div>
</div>
</div>
</div>
</section>
<!-- Real data demo -->
<section class="hero is-light">
<div id="results-carousel" class="carousel results-carousel">
<div class="item item-video1">
<table style="width: 100%;">
<!-- <caption>Comparison with HiFi-GAN-2</caption> -->
<caption style="text-align: center; width: 102%;"><h1 class="subtitle has-text-centered">Comparison with HiFi-GAN-2</h1></caption>
<tr>
<td style="text-align: center; width: 34%;"><h5 class="subtitle has-text-centered">Input</h5></td>
<td style="text-align: center; width: 34%;"><h5 class="subtitle has-text-centered">HiFi-GAN-2</h5></td>
<td style="text-align: center; width: 34%;"><h5 class="subtitle has-text-centered">Ours</h5></td>
</tr>
<tr>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/input/f75.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/hifigan2/f75.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/ours/f75.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/input/f81.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/hifigan2/f81.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/ours/f81.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/input/f82.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/hifigan2/f82.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/ours/f82.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/input/f88.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/hifigan2/f88.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/ours/f88.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/input/m76.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/hifigan2/m76.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/ours/m76.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/input/m77.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/hifigan2/m77.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/ours/m77.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/input/m80.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/hifigan2/m80.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/ours/m80.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/input/m81.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/hifigan2/m81.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/ours/m81.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/input/m83.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/hifigan2/m83.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/ours/m83.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/input/m98.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/hifigan2/m98.wav"></audio></td>
<td style="text-align: center; width: 34%;"><audio controls style="width: 66%;"><source src="samples/ours_vs_hifigan2/ours/m98.wav"></audio></td>
</tr>
</table>
</div>
<div class="item item-video3">
<table style="width: 100%;">
<caption style="text-align: center; width: 100%;"><h1 class="subtitle has-text-centered">Comparison with UNIVERSE</h1></caption>
<tr>
<td style="text-align: center; width: 25%;"><h3 class="subtitle has-text-centered">Input</h3></td>
<td style="text-align: center; width: 25%;"><h3 class="subtitle has-text-centered">UNIVERSE</h3></td>
<td style="text-align: center; width: 25%;"><h3 class="subtitle has-text-centered">Ours</h3></td>
<td style="text-align: center; width: 25%;"><h3 class="subtitle has-text-centered">Ground truth</h3></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/input/0098.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/universe/0098.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/ours/0098.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/gt/0098.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/input/0096.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/universe/0096.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/ours/0096.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/gt/0096.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/input/0095.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/universe/0095.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/ours/0095.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/gt/0095.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/input/0090.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/universe/0090.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/ours/0090.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/gt/0090.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/input/0083.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/universe/0083.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/ours/0083.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/gt/0083.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/input/0076.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/universe/0076.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/ours/0076.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/gt/0076.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/input/0065.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/universe/0065.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/ours/0065.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/gt/0065.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/input/0061.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/universe/0061.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/ours/0061.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/gt/0061.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/input/0060.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/universe/0060.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/ours/0060.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/gt/0060.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/input/0034.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/universe/0034.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/ours/0034.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/ours_vs_universe/gt/0034.wav"></audio></td>
</tr>
</table>
</div>
</div>
</section>
<section class="section hero">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-full">
<div class="content">
<h2 class="title is-3"><center>Examples of clusters obtained during LMOS studies</center></h2>
<div class="level-set has-text-justified">
<p>
As we mentioned in our paper, we generated clusters with the help of VITS.
In this part we provide the examples of different clusters.
As it can be heard, the diversity of samples from one cluster is not caused by the phrase, speaker or phoneme duration mismatch.
WavLM tends to preserve this structure whilst the L2 distance, for instance, usually not.
<br>
</div>
<div class="level-set has-text-justified">
<table style="width: 100%;">
<tr>
<td style="text-align: center; width: 25%;">Cluster 1</td>
<td style="text-align: center; width: 25%;">Cluster 2</td>
<td style="text-align: center; width: 25%;">Cluster 3</td>
<td style="text-align: center; width: 25%;">Cluster 4</td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_005-8/0.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_009-44/0.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_024-62/0.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_044-27/0.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_005-8/1.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_009-44/1.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_024-62/1.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_044-27/1.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_005-8/2.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_009-44/2.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_024-62/2.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_044-27/2.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_005-8/3.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_009-44/3.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_024-62/3.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_044-27/3.wav"></audio></td>
</tr>
<tr>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_005-8/4.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_009-44/4.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_024-62/4.wav"></audio></td>
<td style="text-align: center; width: 25%;"><audio controls style="width: 90%;"><source src="samples/clusters/p226_044-27/4.wav"></audio></td>
</tr>
</table>
</div>
</div>
</div>
</div>
</div>
</section>
<!--BibTex citation -->
<section class="section" id="Data">
<div class="container is-max-desktop content">
<h2 class="title">Data</h2>
To ensure a fair comparison with our work, we provide samples from five datasets enhanced by our model:
<ul>
<li><b>VoxCeleb</b>: 50 audio clips from VoxCeleb1 (Nagrani et al., 2017), covering the
Speech Transmission Index (STI) range of 0.75-0.99, balanced between male and female speakers.
<a href="https://drive.google.com/file/d/1hYtEimp0LqPpyWQF0Wp0q5QK8RHmus6n/view?usp=drive_link" style="color: blue;">Link</a>
</li>
<li><b>UNIVERSE</b>: 100 audio clips randomly generated by the authors of
UNIVERSE (Serrà et al., 2022) from clean utterances sampled from VCTK and Harvard sentences,
alongside noises from DEMAND and FSDnoisy18k. The data includes various simulated distortions
like band limiting, reverberation, codec, and transmission artifacts. For more details,
refer to (Serrà et al., 2022).
<a href="https://drive.google.com/file/d/1Js_dWZGZfC0NO9-6uV3Z7cqIkfDlNZMV/view?usp=drive_link" style="color: blue;">Link</a>
</li>
<li><b>VCTK-DEMAND</b>: Validation samples from the Valentini denoising
benchmark (Valentini-Botinhao et al., 2017). This dataset facilitates
broad comparisons across various speech enhancement models, with a test
set of 824 utterances containing artificially simulated noisy samples from 2
speakers at 4 SNR levels (17.5, 12.5, 7.5, and 2.5 dB).
<a href="https://drive.google.com/file/d/1f7j45fxbohdoQdzqkkNXQnitjhV0lZAU/view?usp=drive_link" style="color: blue;">Link</a>
</li>
<li><b>LibriTTS</b>: A multi-speaker corpus of English speech at 24kHz sampling rate, originally intended for TTS. We provide an enhanced
version for 100 randomly selected samples from the test-other set.
<a href="https://drive.google.com/file/d/1A6YDP469zFlplrJYQsL1FDQ6t0CM84dC/view?usp=drive_link" style="color: blue;">Link</a>
</li>
<li><b>Deep Noise Suppression Challenge</b>: We provide enhanced version for <code style="color: black;">dns5-blind-testset</code> data
for both headset and non-headset tracks. For more information about the challenge, please, refere to the
<a href="https://github.com/microsoft/DNS-Challenge?tab=readme-ov-file" style="color: blue;">DNS github page</a>. Our enhanced data is available
through the <a href="https://drive.google.com/file/d/1n6jP9UBg6RlC_haAVjEcVOl2GOPsiwS5/view?usp=drive_link" style="color: blue;">link</a>.
</li>
</ul>
</div>
</section>
<!--End BibTex citation -->
<!--BibTex citation -->
<section class="section" id="BibTeX">
<div class="container is-max-desktop content">
<h2 class="title">BibTeX</h2>
<pre><code>@misc{babaev2024finallyfastuniversalspeech,
title={FINALLY: fast and universal speech enhancement with studio-like quality},
author={Nicholas Babaev and Kirill Tamogashev and Azat Saginbaev and Ivan Shchekotov and Hanbin Bae and Hosang Sung and WonJun Lee and Hoon-Young Cho and Pavel Andreev},
year={2024},
eprint={2410.05920},
archivePrefix={arXiv},
primaryClass={cs.SD},
url={https://arxiv.org/abs/2410.05920},
}</code></pre>
</div>
</section>
<!--End BibTex citation -->
<footer class="footer">
<div class="container">
<div class="columns is-centered">
<div class="column is-8">
<div class="content">
<p>
<br> This website is licensed under a <a rel="license" href="https://creativecommons.org/licenses/by-nc/4.0/" target="_blank"> Creative Commons Attribution-NonCommercial 4.0 International License </a>.
</p>
</div>
</div>
</div>
</div>
</footer>
<!-- Statcounter tracking code -->
<!-- You can add a tracker to track page visits by creating an account at statcounter.com -->
<!-- End of Statcounter Code -->
</body>
</html>