Develop and evaluate a suite of speech signal analysis plugins. These plugins should give the features or characteristics of speech segments in the time and frequency domains. The suite should give at least two time-domain features (e.g. RMS values, peak, ZCR, etc.) More features (up to 4) may help achieve higher marks. In the frequency domain, the suite should give a straightforward FFT spectrum and a power spectrum density estimation using the Welch method. In coding and testing the algorithms, you should start with a speech segment sampled at the most common CD sampling rate of 44.1 kHz and then down sample it to 16 kHz or 8 KHz of your choice. All subsequent processing should be done with the down sampled speech signals. There is no requirement of developing specific graphic user interfaces (GUIs) for the plugins, but graphic presentations produced using Matlab should be used to clearly illustrate your results where applicable, and these illustrations should be included in the AES conference paper type of written work.
Please refer to the assessment criteria attached separately for a breakdown of grades.
A written submission in the form of a short conference paper (2000-4000) should be produced to accompany your Matlab codes. The Matlab codes should include detailed comments and be added to the end of the main body of the conference paper.
This submission should include appropriately labelled graphs of your plugin performance (e.g., frequency and time domain plots). In the written submission please include: