VoxForge
Hi!
I'm planning to build an spanish acousting model using the recorded data on voxforge and HTK. But before I need to know:
It is planned to release an oficial spanish acousting model for julius? Actually there is almost 25 hours of recording data. Is far from 200h of 2000h which would be ideally, but maybe it's enough for small applications.
>I'm planning to build an spanish acousting model using the recorded data on voxforge and HTK.
that would be great, please supply link, or we can host it here
>It is planned to release an oficial spanish acousting model for julius?
not in the near future
thanks,
Ken
Ok, I will share it as soon as I build and test it!
I have seen that there are quite recordings that are not valid (too noisy, the speaker says something diferent regarding what he/she should read,...) Do you have some tools to help cleaning the data?
Thank you!
>Do you have some tools to help cleaning the data?
only limited checking is performed for non-English submissions... if there are errors, please let me know and I will update the repository.
thanks,
Ken