LibriSpeech
~1000-hour English read speech dataset from LibriVox audiobooks, one of the most widely used ASR benchmark datasets
Duration
1000 hours
Languages
1
Sample Rate
16 kHz
Published
2015-04
Description
1Read speech from LibriVox audiobook project
2Carefully segmented and aligned
3Provides clean and other subsets divided by audio quality and speaker characteristics
4CC BY 4.0 open-source license
5One of the most widely used benchmark datasets in ASR
Language Details
| Language | Duration |
|---|---|
| English | 1000 hours |
Publisher