Open youngsuen opened 8 years ago
Hi, We tried this praat voice morphing algorithm in skype voice stream. We cut the voice stream into small segments(say 100ms time period for each segment). Then we put the small segments into the voice morphing algorithm to change the sound style. Lastly, the small segments will be redirect back to skype for transmition. As human cannot discern the latency between 100ms, it seems like real time. You can applied more hack ways but it is a trade-off.
Hi,I reviewed your Voice Morphing project, but I found it hard to be implemented in real time, because the praat pitch estimation and pitch pulses postion estimation both need a long time data. Can you give some idea on how to implement Voice Morphing in a frame by frame method?