The “Nyquist theorem” isn’t what you were taught (why digital used to suck)

▶ MY PLUGINS: https://apmastering.com/plugins
▶ MY COURSES: https://apmastering.com/courses

SHOPS I USE AND RECOMMEND:
▶ Sound Imports: http://soundimports.eu
▶ SchneidersLaden: https://schneidersladen.de

GEAR I USE AND RECOMMEND:
▶ Hypex amplifiers: https://hypex.nl
▶ Reaper: http://reaper.fm

STUDIOS I RECOMMEND:
▶ Studio Wong: http://studiowong.de

========== VIDEO DESCRIPTION ==========

Texas Instruments video: https://www.youtube.com/watch?v=U_Yv69IGAfQ

I’m sure I’m going to get a few people objecting to me saying the Monty Montgomery video isn’t absolutely perfect. Now, while I think it is MOSTLY perfect and I really like it, he does make one error of the kind I describe in this video. It’s a very small but important detail. When he talks about band limiting a square wave, he suggests his filter characteristics are:

Cutoff: 20.5kHz
Attenuation at 21kHz: 100db

This means, in the space of 500Hz his filter needs to roll off 100db. If calculated, this gets us a steepness of:

2832 dB per octave

That’s just stupid. He refers to it as being “quite steep” lol. This is the only oversight I can see from his video which is otherwise an absolutely great introduction to digital audio and the cardinal theorem of interpolation.

But this oversight kind of shows that Monty didn’t at the time either understand, or want to have to explain, the fact that modern audio interfaces (and most likely the one he was using for the video) are NOT Nyquist rate devices but rather have delta sigma front ends, which is why we can get such good sound at 44.1kHz CD quality. If you try to do exactly what he says in his video with a Nyquist rate device, you would get different, and poor quality results.

Loading