I like your videos and your explanations, Bredo. 
And they totally agree with what I say. Sample rates (given good filters, which I EXPECT from my RME Hammerfall DSP!) beyond a certain point (48 kHz?) don’t matter anymore, because there is no audible information lost by bandlimiting the signal.
Also, this perfectly explains, why oversampling is necessary in certain audio processes: to avoid foldback during the calculation process. But after the calculations are done, the signal can be downsampled again.
I think that we can close the case, seriously. 