Download the UCL Speech Filing System from here, and do the following exercises:
1. Generate 4 kHz white-noise, and display its waveform, wideband and narrowband spectrograms.
2. Record a segment of speech or open a speech file, and then
a. Display its waveform, wideband and narrowband spectrograms, source spectrum and filter response. (Hint: Tools - Speech - Display - Cross-section)
b. Down-sample the signal to 2 kHz and listen to it.
c. Estimate pitch of the original speech signal.
d. Add pink noise to the original speech signal and save the generated signal as a .wav file.