Voice Morphing

The transition of one speech signal to another is called voice morphing. Speech morphing is similar to image morphing. In image morphing we can smoothly change the texture and shape to another one. In voice morphing we can change the sound of one person to another one. The voice morphing has many applications. By using TTs system we can produce many other different voices. For doing voice transformation we first split it into a number of frames. Then we will apply the transformation into each frame of the signal.

For doing the next step we must find two things that are overall envelope of the sound and pitch information. The most important is the manipulation of the pitch information. For a successful morphing there must be a smooth pitching. The two pitches must match. After the process the frames is converted back into the waveform. During the process we may loses some information and it has to be re-estimated. The voice morphing contains many processes like sampling. The main steps are

• Preprocessing – It involves acquisition in discrete form and windowing
• Pitch analysis – This process is used for extracting the pitch. Then format the information’s in the speech signal
• Morphing
• Signal re-estimation