-
Notifications
You must be signed in to change notification settings - Fork 4
Expand file tree
/
Copy pathindex.html
More file actions
783 lines (731 loc) · 27.2 KB
/
Copy pathindex.html
File metadata and controls
783 lines (731 loc) · 27.2 KB
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
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8" />
<meta
name="description"
content="SynParaSpeech: Automated Synthesis of Paralinguistic Datasets for Speech Generation and Understanding"
/>
<meta
property="og:title"
content="SynParaSpeech: A Large-Scale Bilingual Paralinguistic Dataset"
/>
<meta
property="og:description"
content="118.75 hours of Chinese speech with 6 paralinguistic categories and precise timestamps"
/>
<meta
property="og:url"
content="https://github.com/ShawnPi233/SynParaSpeech"
/>
<meta
name="twitter:title"
content="SynParaSpeech: Automated Synthesis of Paralinguistic Datasets for Speech Generation and Understanding"
/>
<meta
name="twitter:description"
content="Automated synthesis framework for laughter, sighs, throat clearing, gasps, tsk, and pause sounds"
/>
<meta
name="keywords"
content="paralinguistic, speech synthesis, speech understanding, dataset"
/>
<meta name="viewport" content="width=device-width, initial-scale=1" />
<title>SynParaSpeech: Automated Synthesis of Paralinguistic Datasets</title>
<link rel="icon" type="image/x-icon" href="statics/images/logo.ico" />
<link
href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro"
rel="stylesheet"
/>
<link rel="stylesheet" href="statics/css/bulma.min.css" />
<link rel="stylesheet" href="statics/css/bulma-carousel.min.css" />
<link rel="stylesheet" href="statics/css/bulma-slider.min.css" />
<link rel="stylesheet" href="statics/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="statics/css/index.css" />
<style>
.js-sticky-table {
position: relative; /* 让内部 th 的 sticky 参照此容器 */
overflow: auto; /* 出现滚动条 */
max-height: 60vh; /* 可视区域高度，可自己调 */
border: 1px solid #e0e0e0;
border-radius: 8px;
}
.js-sticky-table table {
margin: 0 !important; /* 去掉 Bulma 默认 margin */
border: none;
}
.js-sticky-table thead th {
position: sticky;
top: 0; /* 粘到容器顶部 */
z-index: 10;
background: #3498db !important;
color: white !important;
box-shadow: 0 2px 2px -1px rgba(0,0,0,.15); /* 轻微阴影，更醒目 */
}
#event-table thead tr:last-of-type th {
position: sticky;
top: 0;
z-index: 20;
background: #3498db !important;
color: white !important;
box-shadow: 0 2px 2px -1px rgba(0,0,0,.15);
}
.shadow-card {
box-shadow: 0 4px 8px rgba(0, 0, 0, 0.1);
border-radius: 8px;
background: white;
padding: 2rem;
margin-bottom: 2rem;
overflow: hidden;
width: 100% !important;
max-width: 100% !important;
margin-left: auto !important;
margin-right: auto !important;
}
.container-wider {
margin-left: auto;
margin-right: auto;
}
/* 表格容器样式 */
.table-container {
width: 100%;
overflow-x: auto; /* 横向滚动 */
-webkit-overflow-scrolling: touch;
position: relative;
}
#audio-table thead th {
position: sticky;
top: 0;
background: #3498db !important;
color: white !important;
font-weight: 600;
z-index: 5;
text-align: center !important;
}
/* ========= 第一列（Text 列） ========= */
#audio-table th:first-child,
#audio-table td:first-child {
width: 320px; /* 默认宽度，可调 */
min-width: 300px;
max-width: 400px;
text-align: left !important;
white-space: normal; /* 自动换行 */
word-break: break-word;
}
/* ========= 其它模型列 ========= */
#audio-table th:not(:first-child),
#audio-table td:not(:first-child) {
width: 180px;
min-width: 180px;
max-width: 180px;
overflow: hidden;
text-overflow: ellipsis; /* 超出显示省略号 */
}
/* ========= 音频控件适配 ========= */
#audio-table td audio {
width: 100%;
max-width: 150px;
display: block;
margin: 0 auto;
}
.table-container table thead th {
position: sticky;
top: 0;
z-index: 20;
background: #3498db !important;
text-align: center !important;
vertical-align: middle !important;
color: white !important;
font-weight: 600;
border-bottom: 1px solid #e0e0e0;
}
#Methodology .content ol {
padding-left: 1.2rem; /* 缩进一点点对齐标题 */
margin-left: 0; /* 去掉默认外边距 */
}
td audio {
width: 100%; /* 音频组件占满单元格宽度 */
max-width: 150px; /* 音频最大宽度，避免过大 */
}
/* 奇偶列颜色 */
/* table td:nth-child(even) {
background-color: #f9f9f9;
}
table td:nth-child(odd) {
background-color: #ffffff;
} */
.section {
padding-top: 1rem; /* 默认是 3rem */
padding-bottom: 1rem; /* 默认是 3rem */
}
.hero-body {
padding-top: 2rem; /* 默认是 3rem */
padding-bottom: 2rem; /* 默认是 3rem */
}
/* 添加可拖动指示器 */
.table-container::after {
position: absolute;
right: 10px;
bottom: 10px;
font-size: 16px;
color: #999;
background: rgba(255, 255, 255, 0.8);
border-radius: 50%;
width: 25px;
height: 25px;
display: flex;
align-items: center;
justify-content: center;
z-index: 10;
pointer-events: none;
}
/* 统一所有表格圆角样式 - 关键修改 */
.table,
#audio-table,
#event-table {
width: 100%;
border-collapse: collapse;
min-width: 800px;
border-radius: 12px; /* 统一为12px圆角 */
overflow: hidden; /* 确保圆角生效 */
}
/* 修复表头圆角显示 - 关键修改 */
.table thead th:first-child,
#audio-table thead th:first-child,
#event-table thead th:first-child {
border-top-left-radius: 12px;
}
.table thead th:last-child,
#audio-table thead th:last-child,
#event-table thead th:last-child {
border-top-right-radius: 12px;
}
th,
td {
border: 1px solid #ddd;
padding: 8px;
text-align: center;
vertical-align: top;
}
.table th,
.table td {
padding: 0.75rem;
text-align: center !important;
vertical-align: middle;
border: 1px solid #e0e0e0;
word-break: break-word;
min-width: 120px;
max-width: 300px;
}
.table th:first-child {
background-color: #3498db;
color: white;
z-index: 3;
}
.table th {
background-color: #3498db !important;
color: white !important;
font-weight: 600;
position: sticky;
top: 0;
z-index: 1;
}
audio {
width: 100%;
max-width: 170px;
box-sizing: border-box;
margin: 0 auto;
display: block;
}
@media screen and (min-width: 1024px) {
.container-wider {
max-width: 1440px;
}
}
@media screen and (min-width: 1216px) {
.container-wider {
max-width: 1600px;
}
}
@media screen and (max-width: 1023px) {
.container-wider {
max-width: 90%;
}
.table th,
.table td {
padding: 0.5rem;
font-size: 0.9rem;
}
.table th:first-child,
.table td:first-child {
min-width: 150px;
}
}
@media screen and (max-width: 768px) {
.table th,
.table td {
padding: 0.4rem;
font-size: 0.85rem;
min-width: 100px;
}
.table th:first-child,
.table td:first-child {
min-width: 120px;
}
}
#BibTeX pre {
font-size: 20px;
}
</style>
<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="statics/js/fontawesome.all.min.js"></script>
<script src="statics/js/bulma-carousel.min.js"></script>
<script src="statics/js/bulma-slider.min.js"></script>
<script src="statics/js/index.js"></script>
</head>
<body>
<section class="hero">
<div class="hero-body">
<div class="container-wider">
<div class="columns is-centered">
<div class="column has-text-centered">
<h1 class="title is-1 publication-title">SynParaSpeech</h1>
<h2 class="subtitle is-3">
Automated Synthesis of Paralinguistic Datasets for Speech
Generation and Understanding
</h2>
<div class="is-size-5 publication-authors">
<span class="author-block">Bingsong Bai<sup>1</sup>*, Qihang Lu<sup>1</sup>*, Wenbing Yang<sup>1</sup>*,</span>
<span class="author-block">Zihan Sun<sup>2</sup>, Yueran Hou<sup>2</sup>, Peilei Jia<sup>2</sup>, Songbai Pu<sup>2</sup>,</span>
<span class="author-block">Ruibo Fu<sup>3</sup>, Yingming Gao<sup>1</sup>, Ya Li<sup>1</sup>†, Jun Gao<sup>2</sup>†</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block"><sup>1</sup>School of Artificial Intelligence, Beijing University of Posts and Telecommunications, China</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block"><sup>2</sup>Hello Group Inc., China</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block"><sup>3</sup>Institute of Automation, Chinese Academy of Sciences, China</span>
</div>
<div class="mt-4">
<a href="https://arxiv.org/abs/2509.14946"
target="_blank"
rel="noopener noreferrer"
>
<img
src="https://img.shields.io/badge/arXiv-2509.14946-b31b1b.svg?logo=arxiv&logoColor=white"
alt="arXiv:2509.14946"
/>
</a>
<a
href="https://github.com/ShawnPi233/SynParaSpeech"
target="_blank"
rel="noopener noreferrer"
>
<img
src="https://img.shields.io/badge/GitHub-Repository-blue?logo=github"
alt="GitHub Repository"
/>
</a>
<a href="https://huggingface.co/datasets/shawnpi/SynParaSpeech"><img src="https://img.shields.io/badge/%F0%9F%A4%97%20Dataset%20-%20Access-orange" alt="Dataset Access"></a>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="hero">
<div class="hero-body">
<div class="container-wider">
<div class="column has-text-left">
<h2 class="title is-3 has-text-left">Abstract</h2>
<div class="content has-text-justified">
<p>
Paralinguistic sounds like laughter and sighs are crucial for
realistic speech synthesis. They make natural talks more
engaging and authentic. But curt methods depend heavily on
private datasets, and open ones have problems: missing speech,
timestamps, or not matching real life. To mitigate these issues,
we propose an automated synthesis framework for large-scale
paralinguistic datasets and introduce SynParaSpeech. It includes
6 paralinguistic categories, 118.75 hours of Chinese speech, and
precise timestamp annotations. Our work contributes the first
automated synthesis method for such datasets, the release of
SynParaSpeech, improved paralinguistic speech synthesis models
via fine-tuning, and enhanced paralinguistic event detection
through prompt tuning.
</p>
<div class="mt-6">
<h3 class="title is-5 has-text-left">Contents</h3>
<ul class="list-disc pl-6 mt-2">
<li>
<a href="#Methodology" class="link"
>Automated Synthesis Pipeline</a
>
</li>
<li>
<a href="#Dataset-Overview" class="link"
>Dataset Overview</a
>
</li>
<li>
<a href="#Paralinguistic-TTS" class="link"
>Paralinguistic TTS Improvement</a
>
</li>
<li>
<a href="#Event-Detection" class="link"
>Paralinguistic Event Detection</a
>
</li>
</ul>
</div>
</div>
</div>
</div>
</div>
</section>
<section id="Methodology" class="section">
<div class="container-wider">
<div class="shadow-card">
<h2 class="title is-3 has-text-left">Automated Synthesis Pipeline</h2>
<div class="content has-text-left">
<ol class="list-decimal pl-6">
<li>
<strong>Labeled Text Synthesis</strong>: ASR models (Whisper,
Paraformer) generate transcriptions with VAD-based timestamp
correction. LLMs insert paralinguistic tags at appropriate
positions.
</li>
<li>
<strong>Audio Synthesis</strong>: Paralinguistic audio clips are
converted to match speech timbre using SeedVC, then inserted
into speech segments at annotated timestamps.
</li>
<li>
<strong>Verification</strong>: Manual checks ensure naturalness,
timbre consistency, audio quality, and timing alignment.
</li>
</ol>
<figure class="image is-centered mt-4">
<img
src="statics/figs/synparaspeech.png"
alt="SynParaSpeech Pipeline"
style="max-width: 1400px"
/>
</figure>
</div>
</div>
</div>
</section>
<section id="Dataset-Overview" class="section">
<div class="container-wider">
<div class="shadow-card">
<h2 class="title is-3 has-text-left">Dataset Overview</h2>
<div class="content has-text-justified">
<p>
SynParaSpeech covers 6 paralinguistic categories (laughter, sigh,
throat clearing, gasp, tsk, pause), with 118.75 hours of audio and
precise timestamp annotations. The dataset is constructed via an
automated pipeline combining ASR transcription, LLM-based
paralinguistic tagging, voice conversion, and manual verification.
</p>
<div class="table-responsive mt-4">
<div class="table-container">
<table class="table">
<thead>
<tr>
<th>Category</th>
<th>Hours</th>
<th>Clips</th>
<th>Avg.(s)</th>
<th>Share</th>
</tr>
</thead>
<tbody>
<tr>
<td>Sigh</td>
<td>28.22</td>
<td>17,706</td>
<td>5.74</td>
<td>23.76 %</td>
</tr>
<tr>
<td>Throat Clearing</td>
<td>25.45</td>
<td>18,827</td>
<td>4.87</td>
<td>21.43 %</td>
</tr>
<tr>
<td>Laugh</td>
<td>20.84</td>
<td>13,023</td>
<td>5.76</td>
<td>17.55 %</td>
</tr>
<tr>
<td>Pause</td>
<td>18.30</td>
<td>9,643</td>
<td>6.83</td>
<td>15.41 %</td>
</tr>
<tr>
<td>Tsk</td>
<td>14.82</td>
<td>11,941</td>
<td>4.47</td>
<td>12.48 %</td>
</tr>
<tr>
<td>Gasp</td>
<td>11.11</td>
<td>8,846</td>
<td>4.52</td>
<td>9.36 %</td>
</tr>
<tr>
<td><strong>Total</strong></td>
<td><strong>118.75</strong></td>
<td><strong>79,986</strong></td>
<td><strong>5.34</strong></td>
<td><strong>100.00 %</strong></td>
</tr>
</tbody>
</table>
</div>
</div>
</div>
</div>
</div>
</section>
<section id="Paralinguistic-TTS" class="section">
<div class="container-wider">
<div class="shadow-card">
<h2 class="title is-3 has-text-left">
Paralinguistic TTS Improvement
</h2>
<p class="has-text-left">
Fine-tuning with SynParaSpeech enhances paralinguistic generation
quality in CosyVoice2 and F5-TTS
</p>
<div class="table-responsive">
<div class="table-container" style="margin-top: 30px">
<table id="audio-table">
<thead>
<tr id="table-head"></tr>
</thead>
<tbody id="table-body"></tbody>
</table>
</div>
</div>
</div>
</div>
</section>
<section id="Event-Detection" class="section">
<div class="container-wider">
<div class="shadow-card">
<h2 class="title is-3 has-text-left">Paralinguistic Event Detection</h2>
<p class="has-text-left">
Prompt tuning with SynParaSpeech improves event localization and
classification accuracy
</p>
<div class="js-sticky-table" style="margin-top: 30px">
<table id="event-table" class="table is-narrow is-striped">
<thead id="ved-thead">
<tr>
<th>Kimi Audio<br>0-shot</th>
<th>Kimi Audio<br> 1-shot</th>
<th>Kimi Audio<br> 3-shot</th>
<th>Kimi Audio<br> 5-shot</th>
<th>Kimi Audio<br> 7-shot</th>
<th>Qwen 2.5 omni 0-shot</th>
<th>Qwen 2.5 omni 1-shot</th>
<th>Qwen 2.5 omni 3-shot</th>
<th>Qwen 2.5 omni 5-shot</th>
<th>Qwen 2.5 omni 7-shot</th>
</tr>
</thead>
<tbody id="detection-body">
<!-- 数据由 JS 动态填充 -->
</tbody>
</table>
</div>
</div>
</div>
</div>
</section>
<section class="section" id="BibTeX">
<div class="container-wider content">
<h2 class="title">BibTeX</h2>
<pre><code>@article{bai2025synparaspeech,
title={SynParaSpeech: Automated Synthesis of Paralinguistic Datasets for Speech Generation and Understanding},
author={Bingsong Bai and Qihang Lu and Wenbing Yang and Zihan Sun and Yueran Hou and Peilei Jia and Songbai Pu and Ruibo Fu and Yingming Gao and Ya Li and Jun Gao},
journal={arXiv preprint arXiv:2509.14946},
year={2025}
}
</code></pre>
</div>
</section>
<footer class="footer">
<div class="container">
<div class="columns has-text-leftd">
<div class="column is-8">
<div class="content">
<p>
<br />Licensed under
<a
rel="license"
href="https://creativecommons.org/licenses/by-nc-nd/4.0/"
target="_blank"
>CC BY-NC-ND 4.0</a
>.
</p>
</div>
</div>
</div>
</div>
</footer>
<script>
const tableHead = document.getElementById("table-head");
const tableBody = document.getElementById("table-body");
// 模型显示名称
const modelDisplayNames = {
gt:"Reference",
cosyvoice_base: "Baseline CosyVoice2",
cosyvoice_sft: "CosyVoice2<br>+ SynParaSpeech",
cosyvoice_sft_dpo: "CosyVoice2<br>+ SynParaSpeech<br>+ DPO-Staged",
"cosyvoice_sft+dpo": "CosyVoice2<br>+ SynParaSpeech<br>+ DPO-Joint",
F5_base: "Baseline F5-TTS",
F5_sft: "F5-TTS<br>+ SynParaSpeech",
cosyvoice_nv: "CosyVoice2<br>+ NVS",
F5_nv: "F5-TTS<br>+ NVS"
};
let tts_data = {};
// 初始化表头（固定）
function initTableHeader() {
tableHead.innerHTML = '<th style="text-align:center !important;">Text</th>';
Object.keys(modelDisplayNames).forEach((model) => {
const th = document.createElement("th");
th.innerHTML = modelDisplayNames[model];
tableHead.appendChild(th);
});
}
// 渲染类别下拉菜单
// 在populateTable函数中添加高亮逻辑
function populateTable() {
tableBody.innerHTML = "";
const referenceModel = "gt";
const categories = Object.keys(tts_data[referenceModel]);
// 标签颜色映射
const tagColors = {
'[laugh]': '#B3E5FC', // 明亮的天蓝色
'[sigh]': '#bbdefb', // 标准的蓝色
'[gasp]': '#B3E5FC', // 清新的亮蓝色
'[throat_clearing]': '#81D4FA', // 柔和的浅蓝色
'[tsk]': '#81D4FA', // 淡雅的淡蓝色
'[pause]': '#B3E5FC' // 非常淡的蓝色
};
categories.forEach((category, categoryIndex) => {
// 添加类别标题行
const categoryRow = document.createElement("tr");
const categoryCell = document.createElement("td");
categoryCell.colSpan = Object.keys(modelDisplayNames).length + 1;
categoryCell.style.backgroundColor = "#e0e0e0";
categoryCell.style.fontWeight = "bold";
categoryCell.style.textAlign = "center";
categoryCell.textContent = `Category：${category}`;
categoryRow.appendChild(categoryCell);
tableBody.appendChild(categoryRow);
const texts = tts_data[referenceModel][category].texts;
texts.forEach((text, idx) => {
const tr = document.createElement("tr");
// 文本列 - 添加标签高亮
const tdText = document.createElement("td");
// 高亮所有标签
let highlightedText = text;
Object.keys(tagColors).forEach(tag => {
if (text.includes(tag)) {
highlightedText = highlightedText.replace(
new RegExp(tag.replace(/[.*+?^${}()|[\]\\]/g, '\\$&'), 'g'),
`<span class="highlighted-tag" style="background-color: ${tagColors[tag]}; padding: 2px 4px; border-radius: 3px; font-weight: bold;">${tag}</span>`
);
}
});
tdText.innerHTML = highlightedText;
tr.appendChild(tdText);
// 各模型音频列
Object.keys(modelDisplayNames).forEach((model) => {
const td = document.createElement("td");
if (tts_data[model] && tts_data[model][category]) {
const audios = tts_data[model][category].audios;
if (audios && audios[idx]) {
const audio = document.createElement("audio");
audio.controls = true;
audio.src = audios[idx];
td.appendChild(audio);
}
}
tr.appendChild(td);
});
tableBody.appendChild(tr);
});
});
}
/* ========== TTS 表格 ========== */
fetch("assets/tts_data.json")
.then(res => res.json())
.then(data => {
tts_data = data;
initTableHeader();
populateTable();
});
/* ========== Event-Detection 表格 ========== */
fetch("assets/ved_data.json")
.then(res => res.json())
.then(data => {
const tbody = document.getElementById('detection-body');
const shots = [0, 1, 3, 5, 7];
data.forEach(item => {
/* ---- GT 行（占满 10 列）---- */
const gtRow = document.createElement('tr');
gtRow.className = 'has-background-light';
gtRow.innerHTML = `
<td colspan="10">
<div class="is-flex is-align-items-center">
<div style="min-width:260px;">
<audio controls style="width:100%;">
<source src="${item.audio}" type="audio/wav">
</audio>
</div>
<div class="ml-4">
<strong>GT Text：</strong><span>${item.true_text}</span>
</div>
</div>
</td>`;
tbody.appendChild(gtRow);
/* ---- 结果行 ---- */
const resRow = document.createElement('tr');
shots.forEach(s => { // Kimi 0→7
const td = document.createElement('td');
td.textContent = item[`kimi_${s}`] || '';
resRow.appendChild(td);
});
shots.forEach(s => { // Qwen 0→7
const td = document.createElement('td');
td.textContent = item[`qwen_${s}`] || '';
resRow.appendChild(td);
});
tbody.appendChild(resRow);
});
});
</script>
</body>
</html>