There are some Quicktime movies I showed in some lectures which you can try to view.
You will need a Quicktime player in your browser to see them. I don't guarantee they will work on your computer -
although I will try to help!!  I use a Mac, but if you have PC problems, let me know.

Feedback
In the lecture (p3) I showed a system with feedback which had a transfer function
    y(t)/x(t) = H/(1-GH)
In this demo, Gain = H and beta = G. Beta, which is the feedback fraction is varied between -1 and +1, holding H constant. The blue lines represents the input sine wave, while the red line is the system output. The system becomes unstable at G = 1/H, although you will see that the amplifier saturates at values  nearby.

Aliasing
I demonstrated what happens when you increase the frequency of a sine wave and sample it at a constant rate. The values are indicated on the figure. The inset shows the waveform reconstructed from the samples by joining them up. The red histogram shows the value of the frequency of the reconstructed wave (the program uses a Fast Fourier Transform method).
The value it calculates is displayed in red, as "Computed frequency". See how it is fooled (compared to the actual value) once the sampling rate falls below twice the true frequency. The tone (if you can hear it) is the audio signal corresponding to the histogrammed frequency. I haven't quite managed to get it in phase with the waveform.
The reason for this is shown in the following demo, which relates to the Nyquist limit discussed in the lecture.

Sampling
If a signal is sampled at less than twice the maximum frequency component contained within it, the Nyquist criterion tells us we cannot reconstruct it perfectly.  This demo shows what happens if we sample a simple sine wave at varying rates. You will see that once the Nyquist limit is passed the reconstructed waveform (simply joining the samples together here) rapidly begins to show artefacts unrelated to the original signal.
If this were an audio signal, the undersampling of high frequency components would lead to low frequency tones being observed, probably sounding very strange.
 
 

Geoff Hall