Article is full of crap that is just plain wrong and misguided and the analogies suck. I'm not sayin 192KHz 24 bit is needed but the article is weird and says things that just are not true.
For example, saying that a stairstepped sine wave is mathematically the same is wrong --- stairsteps are impulses convolved with a square wave "impulse". This creates a roll-off at high frequencies. Basic signal processing. If you don't understand it, don't worry. Many don't. But the resulting sine wave will be the wrong amplitude. Sampling theory is based on infinitely fast impulses at each sample point, not stairsteps. A subtle point, but he misses many subtle points.
As for 192K 24Bit, there are reasons it is useful as opposed to 48Khz or especially 44.1KHz 16Bit:
1. Dynamic range. 20 bits gives 120dB +_ a few. But 16 bits (96dB) is not enough. 24 bits is way overkill, but doesn't hurt anything except storage space. Home theatre systems with 16 bits make audible noise when you turn them up. Put you ear next to the speaker when it is quiet and you will hear hiss. It may hurt your ears if they are there when when some sound comes through, but that depth is audible. His assertion that 16 bits is enough is not science, it's his opinion. (maybe even 18 bits is enough, but 18 is on the edge)
2. Simplicity of DAC - 192KHz means that dac filters can easily remove images. 96Khz is high enough to make the filter job simple, but 192Khz is simpler yet. His assertion that doing steep filters in digital is no issue means he doesn't really understand digital filters. Steeper slopes means higher lobes and more passband ripple.
3. All his talk about ultra sonics is laughable. Design a bad amp and it will sound bad. So what? Oh --- put in a bad signal so it will sound better?
4. My only point of full agreement is that you need good equipment first, 192/24 second. And I partially agree in that 192/24 is overkill.
