[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
rules of good speech for Cochlear implant users
I am researching on front end speech processing technique to help Cochlear implant users improve their speech recognition, based on the assumption that certain acoustic speech pattern may arouse optimal representation with CI users' daily worn speech processor. I am thinking if we can have some general rules that describe "good speech" for general cochlear implant users, that would help. Or if we can have some consistent clinical data to show this point, it possibly will help this research goal from engineering point of view also. Once I have those rules, certain speech manipulation technique (e.g., normalization etc. ) may try to optimize the acoustic speech pattern before the speech signal feed into the speech processor.
Let me throw couple of traits I am guessing that might fit:
1)larger temporal modulation index
2)slower speaking rate,
3)larger spectral modulation?
4)clear speaking vs. conversational style..
I especially interested in the rules that describe the spectral envelope. Those rules may help to set as a guideline for the speech manipulation. Without such rules, the speech manipulation can be only blind or empirical.
Any of your comments and response on this topic is greatly appreciated.
Ph.D student at USC