Just FYI, “high res audio” is a fraud, a marketing gimmick based on pseudoscience. It is not physically possible to hear frequencies that high. A standard CD already captures everything within human hearing range and then some, after you’ve reached the age you’re likely to care about audio quality your hearing won’t even be able to reach that, and the vast majority of music does not make use of anywhere near the dynamic range supposedly offered by this “higher resolution.” All it does is waste disk space. It’s useful only in studios for purely technical reasons, utterly useless for the end consumer.
Unfortunately, this is misinformed and based on misunderstandings about sound, among other things. There’s nothing “pseudo” about the science, it is what is. While it’s true that humans can’t hear fundamental tones above a certain threshold (20K if you’re a kid, for instance), high frequency response is only part of what’s happening with high res audio. I won’t go into a lengthy discussion about it, but, as an example if you are trying to recreate a waveform, the more plot points you have, the more accurate that waveform is going to be. Higher sample rate = more accurate waveform.
The time resolution of Redbook audio is not as high as the time resolution of “hi-res audio”, but it’s still more than enough for any of your playback and production needs: http://forums.stevehoffman.tv/threads/time-resolution-of-red-book-45ns.85436/ Sampling does not work as you’re describing it. The reconstruction filter reconstructs the original analog signal up to a certain time resolution, which is very high in Redbook audio. I also think the time resolution only depends on the amount of bits per sample, rather than the sampling rate, but I don’t remember where I read that.
Secondly, all music is extremely dynamic, and while most popular music is compressed and doesn’t make use of 120+ db of dynamic range (nor would you want it to), dynamic range isn’t the entire point. It’s the fact that going from 16bit to 24 bit gives you way, way more info: 16 bit = 65,536 possible volume levels, and with every bit, that number doubles. So at 24 bit, we now have 16,777,216 different volume levels. Further, the digital noise present in every digital recording (the “noise floor”) is moved even further into the background.
24 bit audio has more dynamic range, since the noise floor is lower, but 16 bit audio has more than 96 db of dynamic range. The article Density has posted already gives an explanation of why that commonly cited number is misleading. When I convert from 24 bit to 16 bit, I use the gesemann noise-sampling curve from sox (it’s supposed to be much more transparent to the human ear than regulat TPDF and other noise-sampling curves), and the resulting audio has more than 96 db of dynamic range.
Hi-Res audio is a scam.