Google translate:

===

Hello,

I VoxForge the project for quite some time and I have made several contributions but mostwere never processed.

I see that the number of records for the French body has not moved for a long time and most of the bids were not processed audio books (the last treaties books seem to date back to 2009 ... )

Hence my question : is the VoxForge project in existence?

I'm really starting to contribute to the realization of a French acoustic model but if no one takes care of the project , I prefer not to waste my time (both use it to contribute to an active project).

In short, if the project still exists and that it simply lacks manpower to process the data , I am ready to give a hand . I tried to search a bit but I did not find how to participate in the processing of data sent ...

Thank you in advance for information .

===

my replies:

>I have made several contributions but mostwere never processed

yes, unfortunately there are problems with your recordings.  For example for prompt number 278:

fr-sb-278 C’est pourquoi il paraît prudent de lui en faire grâce.

your recording has a leading edge spike (see attachment ) that needs to be removed before it can be added to the corpus - this has to be done manually... when I have time.

>processed audio books (the last treaties books seem to date back to 2009 ... )

Yes, if you could like to segment them, that would be greatly appreciated.  We need the audio to be segmented into 10-20 second audio files with accompanying test transcriptions.  See here to give you an idea of how the process might be made semi-automated (using HTK). Thanks!

The Lium group would probably be your best bet for French segmentation (using CMU Sphinx).

>is the VoxForge project in existence?

yes!

> if the project still exists and that it simply lacks manpower to process

>the data , I am ready to give a hand .

segmentation of your audio book recordings would be very helpful.

thanks,

ken

Hello ken,

Thank you for these very good news (and for the english translation of my message).

Sorry for my bad recordings, I used the builtin microphone of my laptop and it isn't very good.

I did some new recordings yesterday with a headset microphone, I hope they will have a better quality without spike.

For the segmentation of audio books, I started to prepare the environment on my computer. I hope I will be able to process them but I am far from an expert in audio processing...

Fingers crossed !

Hi Grumpf,

Thank you for your submissions!

I am travelling right now, and will only be able to look at them when I get back Sept 10ish

>but I am far from an expert in audio processing

truthfully, neither am I  :)

thanks for helping out,

Ken

Hello Ken,

You should be back now. I have splitted a chapter of "Les enfants du capitaine Grant" according to your guide.

I got several issues with the FTP yesterday (stuck between 3 or 5 Mo each time) so I uploaded it on Gigasize: http://www.gigasize.com/get/g542jqs8psd

Could take a look at this archive and tell me if it's the good way to proceed. If so, i will proceed the other capters of the book and probably some other books.

I can also write a little french guide on how to contribute with audiobooks if you want.

Grumpf

> I uploaded it on Gigasize

got it.  I will run it through the speech submission scripts to validate it tonight.

Hi Ken,

Do you have news with the speech submission scripts ?

Is there an issue with my files ?

I'm waiting for your return before starting the cutting of other audio books.

Grumpf

Salut,

Je me permets de UP cette discussion puisque intéressé par le sujet.

Comment peut-on faire marcrher la RV sous FreeBSD par exemple ?

Comment ça peut marcher pour gnome 3

Pour géant ou vim ?

Je sais que parfois un handicap est dur à surmonter et c'est là que l'informatique prend tout son sens.

Et même pour une personne valide comme moi une interface prenant en compte des commandes vocales, shortcut ou autres peux permettre un gain de temps.

Cdlt,

Orka.