-
Notifications
You must be signed in to change notification settings - Fork 11
/
Copy pathindex.html
577 lines (517 loc) · 36.5 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
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<link href="main.css" rel="stylesheet" media="all">
<meta name="description" content="Joint-AEC-Beamformer" />
<meta name="keywords" content="acoustic echo cancellation, muti-channel processing, deep neural networks, beamforming, speech enhancement">
<script>
function buttonSwitch(id, text) {
old_src = document.getElementById(id).src;
ind = old_src.lastIndexOf('/');
document.getElementById(id).src = old_src.substr(0, ind + 1) + text;
}
</script>
<title>Joint-AEC-and-Beamforming</title>
</head>
<body>
<div id="top_arrow" style="position: fixed; bottom: 10px; right: 10px;">
<a href="#title"><img src="./figures/top_arrow.jpg" style="border: 0pt none ; width: 26px; height: 26px;"/></a>
</div>
<!-- <p style="margin-bottom:3cm;">" "</p> -->
<h2 id="title" class="auto-style1"><center> Joint Neural AEC and Beamforming with Double-Talk Detection </center></h2>
<p class="auto-style7" align="center">
<a href="https://github.com/vkothapally" target="_blank">Vinay Kothapally</a>
<sup>1</sup>
<a href="https://scholar.google.com/citations?user=nCmKPM4AAAAJ&hl=en&oi=ao" target="_blank">Yong Xu</a>
<sup>2</sup>
<a href="https://scholar.google.com/citations?user=pxFJcEEAAAAJ&hl=en&oi=ao" target="_blank">Meng Yu</a>
<sup>3</sup>
<a href="https://scholar.google.com/citations?user=4nGncN4AAAAJ&hl=en" target="_blank">Shi-Xiong Zhang</a>
<sup>4</sup>
<a href="https://scholar.google.com/citations?user=tMY31_gAAAAJ&hl=en" target="_blank">Dong Yu</a>
<sup>5</sup>
</p>
<p class="auto-style7" align="center">
Center for Robust Speech Systems (CRSS), The University of Texas at Dallas, Texas, USA<sup>1</sup>,
Tencent AI Lab, Bellevue, WA, USA<sup>2,3,4,5</sup>
</p>
<!--<p class="auto-style7" align="center"> </p>-->
<p align=left> </p>
<p align="center">
<!-- Image with description -->
<table style="width:1250px" align="center">
<tr><td><center><img width=1250px alt="" src="figures/proposed_network.png"></center></td></tr>
<tr><td><p align="center" class="auto-style5-j">Figure 1. Architectural detials of all deep learning based joint echo cancellation and beamformer trained using time-domain scale-invariant SNR.</p></td></tr>
</table>
<!-- Abstract -->
<p class="auto-style5"> </p>
<p class="style2", style="color:blue;"><strong><span class="auto-style6"><u>Abstract</u></span></strong></p>
<p align="justify" class="auto-style5-j">
Acoustic echo cancellation (AEC) in full-duplex communication systems eliminates acoustic feedback. However, nonlinear distortions induced by audio devices, background noise, reverberation, and double-talk reduce the efficiency of conventional AEC systems. Several hybrid AEC models were proposed to address this, which use deep learning models to suppress residual echo from standard adaptive filtering. This paper proposes deep learning-based joint AEC and beamforming model (JAECBF) building on our previous self-attentive recurrent neural network (RNN) beamformer. The proposed network consists of two modules: (i) multi-channel neural-AEC, and (ii) joint AEC-RNN beamformer with a double-talk detection (DTD) that computes time-frequency (T-F) beamforming weights. We train the proposed model in an end-to-end approach to eliminate background noise and echoes from far-end audio devices, which include nonlinear distortions. From experimental evaluations, we find the proposed network outperforms other multi-channel AEC and denoising systems in terms of speech recognition rate and overall speech quality.
</p>
<!-- Key Contributions -->
<p class="auto-style5"> </p>
<p class="style2", style="color:blue;"><strong><span class="auto-style6"><u>Key contributions</u></span></strong></p>
<p align="justify" class="auto-style5-j">
Our main contributions towards the proposed all-deep-learning (ADL) Joint AEC and Beamforming system are summarized as follows:
<ol>
<li>We propose using a joint spatial covariance matrix computed using microphone signals and far-end speech as input features, which accounts for cross-correlation between far-end speech and multiple microphones, essential for designing an efficient multi-channel AEC system.</li>
<li>We extend our recently proposed generalized spatio-temporal RNN beamformer (GRNNBF) to a joint spatio-temporal AEC-beamformer (JAECBF) for handling AEC and beamforming simultaneously using original and AEC processed signals.</li>
<li>We employ a double-talk detection (DTD) module based on the multi-head attention and recurrent networks, that computes attention over time while leveraging double-talk detection to suppress far-end residuals.</li>
</ol>
</p>
<!-- Pyhton Packages -->
<p class="auto-style5"> </p>
<p class="style2", style="color:blue;"><strong><span class="auto-style6"><u>Python Scripts</u></span></strong></p>
<ol>
<li>Proposed Joint AEC Beamformer model script (<a href="https://github.com/vkothapally/JOINT-AEC-AND-BEAMFORMING/blob/main/src/model.py" target="_blank">model.py</a>)</li>
<li>Run "python model.py" to print architectural details for the proposed netowrk
</ol>
<!-- Model -->
<p class="auto-style5"> </p>
<p class="style2", style="color:blue;"><strong><span class="auto-style6"><u>Model Architecture</u></span></strong></p>
<p align="justify" class="auto-style5-j">
<center><img src="figures/Model_Arch_v2.png" width="940" /></center>
</p>
<!-- Results -->
<p class="auto-style5"> </p>
<p class="style2", style="color:blue;"><strong><span class="auto-style6"><u>Results on speech quality metrics and word error rate (WER)</u></span></strong></p>
<p align="justify" class="auto-style5-j">
<center><img src="results/Results_Table_AEC_BF.png" width="940" /></center>
</p>
<!-- Performance -->
<p class="auto-style5"> </p>
<p id="performance" , class="auto-style4", style="color:blue;"><strong><u>Enhanced audio samples from various Systems in the study</u></strong></p>
<!-- Audio 01 -->
<center>
<table width="1500" border="1">
<td width="300">
<p><center><img src="specs/01_NP.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/01_NP.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>No Processing</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/01_Speex_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/01_Speex_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>SpeexDSP + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/01_FTLSTM_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/01_FTLSTM_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>FTLSTM + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/01_Proposed.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/01_Proposed.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Proposed JAECBF w. DTD</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/01_CleanRef.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/01_CleanRef.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Reverberant Near-End Clean Speech</strong></center></p>
</td>
</table>
</center>
<!-- Audio 02 -->
<center>
<table width="1500" border="1">
<td width="300">
<p><center><img src="specs/02_NP.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/02_NP.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>No Processing</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/02_Speex_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/02_Speex_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>SpeexDSP + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/02_FTLSTM_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/02_FTLSTM_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>FTLSTM + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/02_Proposed.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/02_Proposed.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Proposed JAECBF w. DTD</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/02_CleanRef.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/02_CleanRef.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Reverberant Near-End Clean Speech</strong></center></p>
</td>
</table>
</center>
<!-- Audio 03 -->
<center>
<table width="1500" border="1">
<td width="300">
<p><center><img src="specs/03_NP.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/03_NP.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>No Processing</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/03_Speex_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/03_Speex_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>SpeexDSP + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/03_FTLSTM_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/03_FTLSTM_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>FTLSTM + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/03_Proposed.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/03_Proposed.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Proposed JAECBF w. DTD</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/03_CleanRef.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/03_CleanRef.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Reverberant Near-End Clean Speech</strong></center></p>
</td>
</table>
</center>
<!-- Audio 04 -->
<center>
<table width="1500" border="1">
<td width="300">
<p><center><img src="specs/04_NP.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/04_NP.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>No Processing</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/04_Speex_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/04_Speex_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>SpeexDSP + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/04_FTLSTM_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/04_FTLSTM_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>FTLSTM + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/04_Proposed.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/04_Proposed.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Proposed JAECBF w. DTD</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/04_CleanRef.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/04_CleanRef.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Reverberant Near-End Clean Speech</strong></center></p>
</td>
</table>
</center>
<!-- Audio 05 -->
<center>
<table width="1500" border="1">
<td width="300">
<p><center><img src="specs/05_NP.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/05_NP.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>No Processing</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/05_Speex_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/05_Speex_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>SpeexDSP + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/05_FTLSTM_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/05_FTLSTM_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>FTLSTM + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/05_Proposed.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/05_Proposed.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Proposed JAECBF w. DTD</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/05_CleanRef.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/05_CleanRef.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Reverberant Near-End Clean Speech</strong></center></p>
</td>
</table>
</center>
<!-- Audio 06 -->
<center>
<table width="1500" border="1">
<td width="300">
<p><center><img src="specs/06_NP.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/06_NP.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>No Processing</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/06_Speex_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/06_Speex_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>SpeexDSP + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/06_FTLSTM_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/06_FTLSTM_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>FTLSTM + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/06_Proposed.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/06_Proposed.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Proposed JAECBF w. DTD</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/06_CleanRef.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/06_CleanRef.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Reverberant Near-End Clean Speech</strong></center></p>
</td>
</table>
</center>
<!-- Audio 07 -->
<center>
<table width="1500" border="1">
<td width="300">
<p><center><img src="specs/07_NP.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/07_NP.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>No Processing</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/07_Speex_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/07_Speex_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>SpeexDSP + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/07_FTLSTM_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/07_FTLSTM_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>FTLSTM + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/07_Proposed.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/07_Proposed.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Proposed JAECBF w. DTD</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/07_CleanRef.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/07_CleanRef.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Reverberant Near-End Clean Speech</strong></center></p>
</td>
</table>
</center>
<!-- Audio 08 -->
<center>
<table width="1500" border="1">
<td width="300">
<p><center><img src="specs/08_NP.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/08_NP.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>No Processing</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/08_Speex_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/08_Speex_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>SpeexDSP + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/08_FTLSTM_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/08_FTLSTM_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>FTLSTM + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/08_Proposed.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/08_Proposed.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Proposed JAECBF w. DTD</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/08_CleanRef.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/08_CleanRef.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Reverberant Near-End Clean Speech</strong></center></p>
</td>
</table>
</center>
<!-- Audio 09 -->
<center>
<table width="1500" border="1">
<td width="300">
<p><center><img src="specs/09_NP.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/09_NP.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>No Processing</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/09_Speex_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/09_Speex_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>SpeexDSP + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/09_FTLSTM_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/09_FTLSTM_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>FTLSTM + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/09_Proposed.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/09_Proposed.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Proposed JAECBF w. DTD</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/09_CleanRef.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/09_CleanRef.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Reverberant Near-End Clean Speech</strong></center></p>
</td>
</table>
</center>
<!-- Audio 10 -->
<center>
<table width="1500" border="1">
<td width="300">
<p><center><img src="specs/10_NP.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/10_NP.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>No Processing</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/10_Speex_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/10_Speex_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>SpeexDSP + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/10_FTLSTM_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/10_FTLSTM_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>FTLSTM + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/10_Proposed.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/10_Proposed.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Proposed JAECBF w. DTD</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/10_CleanRef.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/10_CleanRef.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Reverberant Near-End Clean Speech</strong></center></p>
</td>
</table>
</center>
<!-- Audio 11 -->
<center>
<table width="1500" border="1">
<td width="300">
<p><center><img src="specs/11_NP.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/11_NP.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>No Processing</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/11_Speex_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/11_Speex_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>SpeexDSP + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/11_FTLSTM_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/11_FTLSTM_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>FTLSTM + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/11_Proposed.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/11_Proposed.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Proposed JAECBF w. DTD</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/11_CleanRef.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/11_CleanRef.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Reverberant Near-End Clean Speech</strong></center></p>
</td>
</table>
</center>
<!-- Audio 12 -->
<center>
<table width="1500" border="1">
<td width="300">
<p><center><img src="specs/12_NP.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/12_NP.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>No Processing</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/12_Speex_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/12_Speex_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>SpeexDSP + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/12_FTLSTM_JAECBF.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/12_FTLSTM_JAECBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>FTLSTM + JAECBF</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/12_Proposed.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/12_Proposed.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Proposed JAECBF w. DTD</strong></center></p>
</td>
<td width="300">
<p><center><img src="specs/12_CleanRef.png" width="290" /></center></p>
<p><center><audio width="290" controls><source src="audio/12_CleanRef.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Reverberant Near-End Clean Speech</strong></center></p>
</td>
</table>
</center>
<!--
<p class="auto-style5"> </p>
<p id="performance" , class="auto-style4", style="color:blue;"><strong><u>Enhanced audio samples from various Systems in the study</u></strong></p>
<center>
<table width="1200" border="1">
<td width="320">
<p><center><img src="specs/01_No_processing.png" height="271" /></center></p>
<p><center><audio width="315" controls><source src="audio/01_No_Processing.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>No Processing</strong></center></p>
</td>
<td width="320">
<p><center><img src="specs/01_JRNNBF.png" height="271" /></center></p>
<p><center><audio width="315" controls><source src="audio/01_JRNNBF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>JRNNBF</strong></center></p>
</td>
</table>
</center>
<center>
<table width="1200" border="1">
<td width="320">
<p><center><img src="specs/01_SpeexDSP_JRNN_AEC_BF.png" height="271" /></center></p>
<p><center><audio width="315" controls><source src="audio/01_SpeexDSP_JRNN_AEC_BF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>SpeexDSP + JAECBF</strong></center></p>
</td>
<td width="320">
<p><center><img src="specs/01_FTLSTM_JRNN_AEC_BF.png" height="271" /></center></p>
<p><center><audio width="315" controls><source src="audio/01_FTLSTM_JRNN_AEC_BF.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>FT-LSTM + JAECBF </strong></center></p>
</td>
</table>
</center>
<center>
<table width="1200" border="1">
<td width="320">
<p><center><img src="specs/01_Proposed.png" height="271" /></center></p>
<p><center><audio width="315" controls><source src="audio/01_Proposed.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Proposed JAECBF w. DTD</strong></center></p>
</td>
<td width="320">
<p><center><img src="specs/01_Near_End_Reference.png" height="271" /></center></p>
<p><center><audio width="315" controls><source src="audio/01_Near_End.wav" type="audio/wav" >Your browser does not support the audio element.</audio></center></p>
<p><center><strong>Near-End Reverberant Speech</strong></center></p>
</td>
</table>
</center> -->
<p><strong>----------------------------------------------------------------------------------------------------------------------------------------------------</strong></p>
<p><strong>More audio samples can be released upon approval of the submission and company's sharing policy.</strong></p>
<p><strong>----------------------------------------------------------------------------------------------------------------------------------------------------</strong></p>
<p class="auto-style5"> </p>
<p id="reference" , class="auto-style4", style="color:blue;"><strong><u>References</u></strong></p>
<p id="ref_1" class="auto-style5">
[1] Q. Wang and et al., “A frequency-domain nonlinear echo processing algorithm for high quality hands-free voice communication devices,” Multimedia Tools and Applications, vol. 80, no. 7, 2021.
<br> [2] Y. Park and et al., “Frequency domain acoustic echo suppression based on soft decision,” IEEE Signal Processing Letters, vol. 16, no. 1, pp. 53–56, 2008.
<br> [3] S. Zhang and et al., “F-T-LSTM Based Complex Network for Joint Acoustic Echo Cancellation and Speech Enhancement,” in Interspeech, 2021, pp. 4758–4762.
<br> [4] I. Amir and et al., “Nonlinear Acoustic Echo Cancellation with Deep Learning,” in Interspeech, 2021.
<br> [5] J. Valin, “Speex: A free codec for free speech,” ArXiv, vol.abs/1602.08668, 2016.
<br> [6] L. Ma and et al., “Acoustic echo cancellation by combining adaptive digital filter and recurrent neural network,” arXiv preprint arXiv:2005.09237, 2020.
<br> [7] L. Ma and et al., “Echofilter: End-to-end neural network for acoustic echo cancellation,” arXiv preprint arXiv:2105.14666, 2021.
<br> [8] X. Zhou and et al., “Residual acoustic echo suppression based on efficient multi-task convolutional neural network,” arXiv preprint arXiv:2009.13931, 2020.
<br> [9] C. Zhang and et al., “A robust and cascaded acoustic echo cancellation based on deep learning.,” in Interspeech, 2020, pp. 3940–3944.
<br> [10] H. Zhang and et al., “A Deep Learning Approach to Multi-Channel and Multi-Microphone Acoustic Echo Cancellation,” in Interspeech, 2021, pp. 1139–1143.
<br> [11] L. Ma and et al., “Multi-scale attention neural network for acoustic echo cancellation,” arXiv preprint arXiv:2106.00010, 2021.
<br> [12] R. Peng and et al., “Acoustic Echo Cancellation Using Deep Complex Neural Network with Nonlinear Magnitude Compression and Phase Information,” in Interspeech, 2021, pp.4768–4772.
<br> [13] M. M. Halimeh and et al., “Combining adaptive filtering and complex-valued deep postfiltering for acoustic echo cancellation,” in ICASSP, 2021, pp. 121–125.
<br> [14] Y. Xu and et al., “Generalized spatio-temporal rnn beamformer for target speech separation,” Interspeech, 2021.
<br> [15] Zhuohuang Zhang, Yong Xu, and et al., “ADL-MVDR: All deep learning MVDR beamformer for target speech separation,” in ICASSP, 2021, pp. 6089–6093.
<br> [16] W. Mack and et al., “Deep filtering: Signal extraction and reconstruction using complex time-frequency filters,” IEEE Signal Processing Letters, vol. 27, pp. 61–65, 2019.
<br> [17] M. M. Halimeh and et al., “Combining adaptive filtering and complex-valued deep postfiltering for acoustic echo cancella-tion,” in ICASSP, 2021, pp. 121–125.
</p>
<p class="auto-style1"><font color="#999999">Last update: March 24, 2022</font></p>
</body>
</html>