A Program for Pitch Period Detection Using Autocorrelation Method

Resource Overview

A self-developed program utilizing autocorrelation method for pitch period extraction in speech signals, featuring signal preprocessing and pitch tracking capabilities.

Detailed Documentation

This is a custom-developed program that employs the autocorrelation method to detect pitch periods in speech signals. The core functionality involves analyzing and processing speech data to extract fundamental frequency information for better understanding and application. The implementation includes framing the speech signal into short segments, computing autocorrelation coefficients to identify periodicity peaks, and applying peak detection algorithms to determine pitch periods. The program generates a structured data file containing pitch period information for further analysis and visualization. Additional features incorporate digital filtering techniques for signal enhancement and noise reduction algorithms to improve detection accuracy under noisy conditions. The system utilizes windowing functions (e.g., Hamming window) for frame smoothing and includes threshold-based voicing detection to distinguish between voiced and unvoiced segments.