5
$\begingroup$

Does anybody know any open voice/speech annotated database for Voice Activity Detection? Or any other corpus for speech evaluation purposes?

The ones I know are private: AURORA, TIMIT, SpeechDat-Car, Speecon, Catalan Speecon.

The other option is to create hand-labeled annotations for own recordings, or use a method like the Short-Term Energy Contour. But in the first case the amount of files to process would be insurmountable, and in the second, the energy based methods only work fine for free-noise recordings.

Improve this question
$\endgroup$

3 Answers 3

Reset to default
4
$\begingroup$

You can take some big speech corpus like TEDLIUM and add the noise you like:

http://www-lium.univ-lemans.fr/en/content/ted-lium-corpus

The advantage of TEDLIUM is that it's a set of continuous recordings with speech timings, not just a collection of utterances.

Improve this answer
$\endgroup$
1
  • $\begingroup$ Wow thanks. I heard also that The IViE Corpus is also a good choice, but is not as timed as the one you suggested. $\endgroup$
    – JMFS
    Commented Jan 7, 2015 at 19:15
7
$\begingroup$

Researchers from the Johns Hopkins University have recently released a corpus of music, speech, and noise which, according to them, is suitable for training models for voice activity detection and music/speech discrimination.

See https://arxiv.org/pdf/1510.08484.pdf for details.

Improve this answer
$\endgroup$
3
$\begingroup$

Free VAD Data set recorded in real environment with ground truth label:

refer: https://github.com/jtkim-kaist/VAD-Toolkit

Improve this answer
$\endgroup$
1
  • $\begingroup$ Kaist has been earning a good reputation these years. Is this toolkit related to the work of Yang-Hann Kim? I wonder because I took his MOOC in Acoustics. $\endgroup$
    – JMFS
    Commented Nov 9, 2017 at 16:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.