Thanks for the feedback! Yes, this is the 'easiest' thing to do and I also thought that it might be the most useful as well. My assumption is: if you speak too fast, you are NOT following any speaking technique. I think that is ok to speed up sometimes, but if I speed up most of the day, eventually I would lose control. It applies FFT to PCM audio streams in real-time to count syllables. The more complicated aspect is the filtering required to defect only the user speaking during conversations (or noisy environments).