What is Aliasing | Anti-Aliasing Technique in signal processing
This page covers Aliasing basics and mention Anti-Aliasing Technique. The aliasing definition and its use in digital signal processing (DSP) are described.
Aliasing occurs due to inadequate sampling used in A to D conversion. Figure-1 a) depicts analog continous spectrum band limited to +B on upper side and -B on lower side. Let us understand sampling of this signal.
When this signal is sampled using Fs > 2B then it will separate spectral replications at folding frequency of about +/- fs/2. This is shown in figure-1 (b). The relationship fs >= 2B is referred as Nyquist Criteria.
When sampling frequency of fs = 1.5*B is choosen then, it will result into spectral replications. This type of sampling is known as undersampling. Here lower and upper edge of spectral replications exist around +fs and -fs. This overlaps with original sampled spectrum centered around 0 Hz. This condition is known as aliasing. This is shown in figure-1.
In practical conditions, analog signal exists with noise energy. When this signal is sampled at rate greater than 2B then it will prevent replications overlapp on each other. But resultant discrete spectrum will have noise energy between -fs/2 and +fs/2.
Hence in order to prevent noise energy, analog low pass anti alias filter is used to attenuate unwanted signal energy above +B and below -B Hz.
With the use of anti-alias filter spectral aliasing can be avoided at the output of A to D Converter. This is shown in figure-2 above. Cutoff frequency of anti-aliasing filter should match bandwidth of signal of interest.
What is Aliasing Frequency ?
➤Signal having bandwidth (B) should be sampled with sampling frequency greater than 2B in order to
avoid aliasing. This condition is called "Nyquist Criterian".
➤At given sampling frequency (fs), signal to be sampled should have BW smaller than Nyquist frequency (Fs/2).
➤Fs/2 is also known as folding frequency or aliasing frequency.