We try to retrain sphinx2 to decode Mandarin(Chinese) phones only, not the words. Will the accuracy be enough for retrieval How should we do step by step? Including recompile sphinx2 in Windows98 platform, wav to mfcc, and audio corpus creation, etc.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
We will be releasing the whole training process, as well as scripts to do it, shortly (by the end of the year). This would allow you to do all-phone decode for Mandarin if you had enough training data.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
We try to retrain sphinx2 to decode Mandarin(Chinese) phones only, not the words. Will the accuracy be enough for retrieval How should we do step by step? Including recompile sphinx2 in Windows98 platform, wav to mfcc, and audio corpus creation, etc.
We will be releasing the whole training process, as well as scripts to do it, shortly (by the end of the year). This would allow you to do all-phone decode for Mandarin if you had enough training data.