Why generally while combining both interpolation and decimation,
we perform interpolation first followed by decimation???

 

Why generally while combining both interpolation and decimation,
we perform interpolation first followed by decimation???

Decimation reduces the information content and signal bandwidth. If you interpolate after decimation, you will be less accurate than if you do it before decimation. Presumably, the purpose of interpolation would be to get the approximate data point values at times that you don't have that data so that, when you decimate, you have the necessary datapoints needed in the decimated signal. Interpolation of the undecimated data might provide very good estimates of those values. If you do the reverse, you will have erased key information content needed to provide the most accurate estimation of the required data point values.