VoxForge
I did a quick search through the forums and didn't see any mention of this. The whitehouse.gov has a corpus of transcribed presidents' speeches along with the audio files. Have these been incorporated in Voxforge?
--- (Edited on 7/29/2010 12:13 pm [GMT-0500] by donnied) ---
>The whitehouse.gov has a corpus of transcribed presidents' speeches
>along with the audio files.
Thanks for letting us know!
>Have these been incorporated in Voxforge?
No, not yet - though I am not sure what kind of audio compression they are using.
--- (Edited on 9/12/2010 12:24 am [GMT-0400] by kmaclean) ---
> No, not yet - though I am not sure what kind of audio compression they are using.
I've just looked at one audio which was MPEG layer-2, layer-3 at 128kbps and the corresponding video which was AAC at 113.6 kbps.
I'd guess that the AAC at 133.6 kbps is fine for ASR. In general, if there is high quality video then the sound is also very good.
Tony
--- (Edited on 12-September-2010 10:29 am [GMT+0100] by TonyR) ---
Just to say that I'm working on this now - to avoid duplicating effort in case anyone else is also working on it. I think the main problem will be working out if the Transcript is accurate.
Tony
--- (Edited on 9/13/2010 8:14 am [GMT-0500] by Visitor) ---