Hello friends, I'm now into another university project, topic of which is under speech recognition. We will be using the c5402 dsk. The first step in the real-time recognition is frame blocking which is just dividing the speech into frames of the same number of samples each. Each frame overlaps adjacent frames over some number of samples. my question now is how can i obtain, in an efficient manner, a frame (or buffer) of samples? I have read the codec header file and found the following functions to be of probable use: codec_read() - the data array & its size are specified codec_cont_capture() - the data array, its size, pointer to function after the capture is completed, and the status of the read are specified However, no example codes for these functions can be found. After reading the Vol. 1 manual, the use of the ABU (autobuffering unit) seems attractive.But again, no sample code could be found for this nice feature. the manual gave a clear procedure as to how to use it but i feel doing those from scratch will prove to be not a good idea. Im in the physics field and I think going into the details would just eat away some much needed time. But still i'm willing and happy to learn those engineering stuff. Am I on the right track? which way should I go? thanks, marlo __________________________________________________ |