Hello PaulVK and Geozem,

Please allow me to explain how up-sampling and down-sampling works and why we need to apply effectively the same low pass (antialiasing) filters for both steps.

1.) Down-Sampling:

If we reduce the sample rate then the Nyquist theorem still applies, which means that the highest frequency sampled in the signal must be lower or equal to 1/2 x sample rate.

To assure that the signal must be low pass filtered and then down-sampled. By using dedicated decimation filters that happens in one step.

2.) Up-Sampling

To do an up-sampling we need to insert the right amount of samples.

That is a two step process:

a.) Insert zeros to achieve the target sample rate

b.) Apply a low pass (Antialiasing) filter to remove the mirror spectra created by inserting zeros. That filter has the same cut-off frequency as the filter used for down-sampling.

If we have to change the sample base (e.g. 44.1kHz => 48kHz) then we do an up- and down-sampling to achieve a rational fraction. That is the reason why the up-/down-sampling between the sample bases takes considerable more time.

So, it is not a matter whether we won't like to present an option to switch the filter off, in fact those filters are absolutely necessary to avoid any kind of aliasing.

Thanks

Stephan