Digital Signal Processing – Pitch Detection (MATLAB)
Project Description:
The pitch of an audio signal is an important parameter in the analysis of the information contained
in the signal. The pitch of a typical information-bearing signal such as speech varies constantly and
can sometimes be difficult to identify without resorting to computationally intensive analysis
techniques. Another common difficulty with robust pitch detection is that its value can sometimes
be close to or equal to the frequency of the first formant, particularly for female or young voices.
The purpose of the proposed project is to investigate the key design criteria for an efficient real-time
pitch detector. These criteria will include the principles and software implementation of the lowlevel
signal capture, analysis and display as well as an intuitive user-interface design.
Expected Procedure:
1. Familiarize yourself with the principles of pitch detection (and document it).
2. Familiarize yourself with the existing pitch detectors in the Opus codec (and document it).
3. Based on 1 and 2 above, design an efficient pitch detector which you think will be a good
addition to the Opus codec.
4. Implement the above design in Matlab and document the process.