Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,87 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
---
|
4 |
+
|
5 |
+
|
6 |
+
---
|
7 |
+
|
8 |
+
Data source:
|
9 |
+
https://openslr.org/93/
|
10 |
+
|
11 |
+
Audios and transcriptions are extracted to wav and txt files.
|
12 |
+
|
13 |
+
---
|
14 |
+
|
15 |
+
|
16 |
+
|
17 |
+
|
18 |
+
|
19 |
+
<div class="container">
|
20 |
+
<h2 class="slrStyle"> AISHELL-4
|
21 |
+
</h2>
|
22 |
+
<p class="resource"> <b>Identifier:</b> SLR111 </p>
|
23 |
+
<p class="resource"> <b>Summary:</b> A Free Mandarin Multi-channel Meeting Speech Corpus, provided by Beijing Shell Shell Technology Co.,Ltd
|
24 |
+
</p>
|
25 |
+
<p class="resource"> <b>Category:</b> Speech
|
26 |
+
</p>
|
27 |
+
<p class="resource"> <b>License:</b> CC BY-SA 4.0
|
28 |
+
</p>
|
29 |
+
<p class="resource"> <b>Downloads (use a mirror closer to you):</b> <br>
|
30 |
+
<a href="https://www.openslr.org/resources/111/train_L.tar.gz"> train_L.tar.gz </a> [7.0G] ( Training set of large room, 8-channel microphone array speech
|
31 |
+
) Mirrors:
|
32 |
+
<a href="https://us.openslr.org/resources/111/train_L.tar.gz"> [US] </a>
|
33 |
+
<a href="https://openslr.elda.org/resources/111/train_L.tar.gz"> [EU] </a>
|
34 |
+
<a href="https://openslr.magicdatatech.com/resources/111/train_L.tar.gz"> [CN] </a>
|
35 |
+
<br><a href="https://www.openslr.org/resources/111/train_M.tar.gz"> train_M.tar.gz </a> [25G] ( Training set of medium room, 8-channel microphone array speech
|
36 |
+
) Mirrors:
|
37 |
+
<a href="https://us.openslr.org/resources/111/train_M.tar.gz"> [US] </a>
|
38 |
+
<a href="https://openslr.elda.org/resources/111/train_M.tar.gz"> [EU] </a>
|
39 |
+
<a href="https://openslr.magicdatatech.com/resources/111/train_M.tar.gz"> [CN] </a>
|
40 |
+
<br><a href="https://www.openslr.org/resources/111/train_S.tar.gz"> train_S.tar.gz </a> [14G] ( Training set of small room, 8-channel microphone array speech
|
41 |
+
) Mirrors:
|
42 |
+
<a href="https://us.openslr.org/resources/111/train_S.tar.gz"> [US] </a>
|
43 |
+
<a href="https://openslr.elda.org/resources/111/train_S.tar.gz"> [EU] </a>
|
44 |
+
<a href="https://openslr.magicdatatech.com/resources/111/train_S.tar.gz"> [CN] </a>
|
45 |
+
<br><a href="https://www.openslr.org/resources/111/test.tar.gz"> test.tar.gz </a> [5.2G] ( Test set
|
46 |
+
) Mirrors:
|
47 |
+
<a href="https://us.openslr.org/resources/111/test.tar.gz"> [US] </a>
|
48 |
+
<a href="https://openslr.elda.org/resources/111/test.tar.gz"> [EU] </a>
|
49 |
+
<a href="https://openslr.magicdatatech.com/resources/111/test.tar.gz"> [CN] </a>
|
50 |
+
<br><br></p><p class="resource"><b>About this resource:</b></p><div id="about">The AISHELL-4 is a sizable real-recorded Mandarin speech dataset
|
51 |
+
collected by 8-channel circular microphone array for speech processing
|
52 |
+
in conference scenarios. The dataset consists of 211 recorded meeting
|
53 |
+
sessions, each containing 4 to 8 speakers, with a total length of 120
|
54 |
+
hours. This dataset aims to bridge the advanced research on
|
55 |
+
multi-speaker processing and the practical application scenario in
|
56 |
+
three aspects. With real recorded meetings, AISHELL-4 provides
|
57 |
+
realistic acoustics and rich natural speech characteristics in
|
58 |
+
conversation such as short pause, speech overlap, quick speaker turn,
|
59 |
+
noise, etc. Meanwhile, the accurate transcription and speaker voice
|
60 |
+
activity are provided for each meeting in AISHELL-4. This allows the
|
61 |
+
researchers to explore different aspects in meeting processing,
|
62 |
+
ranging from individual tasks such as speech front-end processing,
|
63 |
+
speech recognition and speaker diarization, to multi-modality modeling
|
64 |
+
and joint optimization of relevant tasks. We also release a
|
65 |
+
PyTorch-based training and evaluation framework as a baseline system to
|
66 |
+
promote reproducible research in this field. The baseline system code
|
67 |
+
and generated samples are available
|
68 |
+
<a href="https://github.com/felixfuyihui/AISHELL-4">here</a>.
|
69 |
+
<p>
|
70 |
+
|
71 |
+
You can cite the data
|
72 |
+
using the following BibTeX entry:
|
73 |
+
</p><pre>
|
74 |
+
@inproceedings{AISHELL-4_2021,
|
75 |
+
title={AISHELL-4: An Open Source Dataset for Speech Enhancement, Separation, Recognition and Speaker Diarization in Conference Scenario},
|
76 |
+
author={Yihui Fu, Luyao Cheng, Shubo Lv, Yukai Jv, Yuxiang Kong, Zhuo Chen, Yanxin Hu, Lei Xie, Jian Wu, Hui Bu, Xin Xu, Jun Du, Jingdong Chen},
|
77 |
+
booktitle={Interspeech},
|
78 |
+
url={https://arxiv.org/abs/2104.03603},
|
79 |
+
year={2021}
|
80 |
+
}
|
81 |
+
</pre>
|
82 |
+
</div><p class="resource"> <b>External URL:</b> <a href="http://www.aishelltech.com/aishell_4"> http://www.aishelltech.com/aishell_4 </a> Full description from the company website
|
83 |
+
</p>
|
84 |
+
|
85 |
+
<div style="height:300px"> </div>
|
86 |
+
|
87 |
+
</div>
|