3 ms·
I needed to build a audio corpus to train acoustic models on. It actually worked pretty good. I was able to build up about 100 hours worth. It was unbelievab
by Caligula 15y ago
I needed to build a audio corpus to train acoustic models on. It actually worked pretty good. I was able to build up about 100 hours worth. It was unbelievably tedious because I had to verify the audio even after I used amazon turk a second time to do a first verification pass.
The other thing is that I don't necessarily want perfect quality audio. I wanted whatever microphone audio normal people have so it probably worked out better that I used amazon turk.
- abe_duarte 15y agoVery interesting, I´m thinking that voicebunny would be useful to these types of projects because you can ask recordings on specific language accents. Also Voicebunny also has a quality check and a guarantee included in the service.