A simple analog waveform is mathematical representation of a wave. It is a graph of any one or more characteristics of a wave against time such as amplitude against time. However, when we store this signal in a computer, we have to save it as a digital approximation. This can be done by taking samples along the waveform e.g. the height of the waveform at specific intervals. It is obvious that the more we sample more we would be able to reconstruct the true signal of the wave.

However, due to space limitations on computers we have to only store finite number of samples. Hence when we want to reproduce the signal, it can be close to the original waveform but not exactly the same. For example when we convert analog to sound to digital, more number of sampling will make the digital sound more like the real voice. A Theorem called 'Nyquist sampling' is used for arriving at the best sampling intervals in order for the signal to be reconstructed most effectively and efficiently.

However, due to space limitations on computers we have to only store finite number of samples. Hence when we want to reproduce the signal, it can be close to the original waveform but not exactly the same. For example when we convert analog to sound to digital, more number of sampling will make the digital sound more like the real voice. A Theorem called 'Nyquist sampling' is used for arriving at the best sampling intervals in order for the signal to be reconstructed most effectively and efficiently.