分享一些英文和中文語音數據集,開啟語音識別之旅
1、TIMIT
簡介:語音識別語料庫
規模:6.3K個英語句子
地址:
https://catalog.ldc.upenn.edu/LDC93S1
http://www.fon.hum.uva.nl/david/ma_ssp/2012/
2、VoxForge
簡介:Free GPL Speech Audio
規模:
地址:http://www.voxforge.org/
3、2000 HUB5 English Evaluation Transcripts
簡介:英語電話對話語音數據集
規模:40 source speech data files
地址:https://catalog.ldc.upenn.edu/LDC2002T43
4、CHIME
簡介:CHiME-5競賽數據集
規模:The CHiME-5 data consists of 20 parties each recorded in a different home
地址:http://spandh.dcs.shef.ac.uk/chime_challenge/data.html
5、Yesno
簡介:Sixty recordings of one individual saying yes or no in Hebrew; each recording is eight words long
規模:60 .wav files, sampled at 8 kHz
地址:http://www.openslr.org/1/
6、Vystadial
簡介:English and Czech data, mirrored from the Vystadial project
規模:
地址:http://www.openslr.org/6/
7、TED-LIUM
簡介:English speech recognition training corpus from TED talks, created by Laboratoire d’Informatique de l’Université du Maine (LIUM)
規模:118 hours of speech
地址:http://www.openslr.org/7/
8、LibriSpeech ASR corpus
簡介:Large-scale (1000 hours) corpus of read English speech
規模:1000 hours of speech
地址:http://www.openslr.org/12/
9、The AMI Corpus
簡介:Acoustic speech data and meta-data from The AMI corpus
規模:100 hours of meeting recordings
地址:http://www.openslr.org/16/
10、THCHS-30
簡介:A Free Chinese Speech Corpus Released by CSLT@Tsinghua University
規模:
地址:http://www.openslr.org/18/
11、TED-LIUMv2
簡介:TED-LIUM corpus release 2, English speech recognition training corpus from TED talks, created by Laboratoire d’Informatique de l’Université du Maine (LIUM)
規模:
地址:http://www.openslr.org/19/
12、THUYG-20
簡介:A free Uyghur speech database Released by CSLT@Tsinghua University & Xinjiang University
規模:
地址:http://www.openslr.org/22/
13、Aishell
簡介:Mandarin data, provided by Beijing Shell Shell Technology Co.,Ltd
規模:
地址:http://www.openslr.org/33/
14、Free ST Chinese Mandarin Corpus
簡介:A free Chinese Mandarin corpus by Surfingtech (www.surfing.ai)
規模:containing utterances from 855 speakers, 102600 utterances
地址:http://www.openslr.org/38/
15、Free ST American English Corpus
簡介:A free American English corpus by Surfingtech (www.surfing.ai)
規模:containing utterances from 10 speakers, Each speaker has about 350 utterances
地址:http://www.openslr.org/45/
16、TED-LIUM Release 3
簡介:TED-LIUM corpus release 3
規模:
地址:http://www.openslr.org/51/
17、FSDD
簡介:一個簡單的語音數據集
規模:1.5K recordings
地址:https://github.com/Jakobovski/free-spoken-digit-dataset
18、FMA
簡介:A Dataset For Music Analysis
規模:106,574 tracks
地址:https://github.com/mdeff/fma
19、Ballroom
簡介:交際舞音頻
規模:698個樣本
地址:http://mtg.upf.edu/ismir2004/contest/tempoContest/node5.html
20、Million Song Dataset
簡介:歌曲集
規模:one million songs
地址:https://labrosa.ee.columbia.edu/millionsong/
![深度學習數據集彙總(語音數據集)](http://p2.ttnews.xyz/loading.gif)
閱讀更多 深度學習社區 的文章