Sampling and Binary Representation - Synthesis Lectures on Signal Processing - page 119

obtain the interpolated value. The operation is actually the inner or dot product of the chosen sinc matrix
row with a column vector selected from the signal to be interpolated.
It is only necessary to perform the interpolation for a limited number of samples to demonstrate
efficacy; 1000-2000 samples are enough at a sample rate of 8kHz to hear that the original and interpolated
versions have the same apparent frequency.
The interpolated samples should be lowpass filtered after interpolation to retain the original fre-
quency limit, which is 4 kHz. The following code may be used:
[b,a] = cheby1(12,0.5,7.9/11.025);
SincInterpSig = filter(b,a,SincInterpSig);
LinInterpSig = filter(b,a,LinInterpSig);