MATLAB Implementation for Speech Signal Processing and Prosody Adjustment Using PSOLA Algorithm

This MATLAB program implements speech signal processing for prosody adjustment using the Pitch Synchronous Overlap and Add (PSOLA) algorithm, which can be applied to various speech processing tasks. The program performs comprehensive speech signal analysis and synthesis, enabling modification of speech rhythm and pitch through prosody adjustment. The core PSOLA algorithm operates through three key implementation phases: pitch detection using autocorrelation or cepstral analysis, pitch-synchronous segmentation of the speech signal, and overlap-add synthesis with modified pitch periods. The algorithm maintains speech naturalness while adjusting speech duration and fundamental frequency. Key MATLAB functions implemented include: - Pitch detection using short-time autocorrelation analysis - Waveform segmentation at pitch pulse locations - Time-scale modification through overlap-add synthesis with resampled segments - Pitch scaling via spectral envelope preservation during synthesis Users can flexibly adjust various speech signal parameters through the program's interactive interface, including pitch scaling factors, time-stretching ratios, and spectral characteristics. The implementation allows for real-time parameter modification and preview of processing results, supporting diverse speech transformation effects such as voice conversion, speech rate adjustment, and expressive speech synthesis.