The speex manual says: "It is recommended to use a frame size in the order of 20 ms (or equal to the codec frame size) and make sure it is easy to perform an FFT of that size (powers of two are better than prime sizes)." What is the impact of using a non-power-of-two frame size here? Since 20 ms of 8000 (or 16000) is not a power of 2, I have to make a choice. It would be most convenient to pass my regular packet size (320 samples). Would there be any advantage for me to re-buffer the samples and pass them 256 (or 128) at a time? Thanks, by the way, for the very interesting and useful software! -Michael