The ability to modify the apparent rate of speech is desirable in a number of applications. For example, one can reduce the bit rate required for medium band speech coding by time-scale compression of the input speech, followed by coding and transmission, followed by time-scale expansion to original signal at the receiver.

There are mainly two types of Speech Modifications namely, Time-Scale Modification and Pitch-Scale Modification. Which, further can be done in two ways, first method is Time-Domain and second method is Frequency-Domain.  In time-domain method the OLA (OverLap and Add) method is applied for synthesizing the speech. For more accuracy the synchronized OLA method is used, which is synchronized with pitch of the speech signal. Other method is based on Short Time Fourier Transform (STFT). STFT method is somewhat complex as it involves many calculations whereas in Time Domain method is faster because of less number of calculations.

In this project the Phase Vocoder Based time-scale modification, and Synchronized Overlap Add time-scale modification Method both are implemented. Both methods are also described theoretically, and after successfully implementation both methods are compared in three different aspects, namely, Quality at different rates, Complexity, and limitation of both methods.