Oversampling

August 5, 2019

Rather than sampling at the Nyquist frequency, as in Example 5, or sampling at twice the Nyquist frequency as in Example 4, there can be sampling at 4 times or even 256 times the Nyquist rate.

Sampling higher than the Nyquist rate is called “over-sampling.” As the sample rate increases, the spectral images spread farther as well as “taller” in the DTFT frequency domain. The significance of over-sampling is discussed later in this course in the context of quantization noise and delta-sigma modulation.

Example 6

Figure 2.8 illustrates the spectral content of the waveform x(t) sampled at 𝖥s = 8  × 𝖿h rather than 4  × 𝖿h. There is still a repeat every 𝜔 = 2𝜋 as in Example 4, but the images also become more narrow and thus more “spread out” in the DTFT frequency domain.

Figure 2.8. The frequency content of an over-sampled waveform.