I'm not at all convinced that they are problematic for human listeners, but I also notice that there are many people who do regard them as problematic. Peter Craven, for example, who has introduced the apodizing filter in digital audio. Anyone who prefers high sample rate digital audio apparently also cares about what happens above 20 kHz (although it could also mean that they have pets or intersample over issues (high sample rates help to reduce intersample overs) or don't understand the Whittaker/Nyquist/Shannon sampling theorem). Peter Craven certainly does understand the Whittaker/Nyquist/Shannon sampling theorem and he is a lot smarter than I will ever be, so I tend to take what he writes seriously, even though he had very little evidence to back up his point of view.
Linear-phase FIR filters have their own very peculiar imperfections, but that is way off topic. Those imperfections can be made arbitrarily small anyway.
Linear-phase FIR filters have their own very peculiar imperfections, but that is way off topic. Those imperfections can be made arbitrarily small anyway.
Maybe the used a first order filter with an Fo at 26k? If they did, it would be down 1,37dB at 14k...
//
Just under 100 dB. That is, I just calculated -100 dB/2log(22050/20000) and rounded the result to arrive at -700 dB/octave.
Last edited: