I take on the notion that you do not need audio at a quality higher than 16 bits, 44KHz
We’ve had various articles from audio experts talking about why you don’t really need 24 bit, 96 KHz audio for playback (you need it for recording, and I think no one argues against that). They argue that CDs actually have enough quality for anyone. I’d like to make a simple counterargument to that.
Firstly, let’s get the immediate counterargument out of the way: Audio you need for “recording” vs “playback” is part of the old hegemony of “recording industry” and “consumers”. In the modern, digital era, we want to be in a world where everyone is a creator, and there’s no reason to say that a “listener” is an inherently different individual than a “creator”. You’re being unfair when you give someone the kind of audio that they cannot mess with.
OK, let’s get to the argument. To start with, let’s look at the sampling rate. While it’s true that as you get older, you can no longer hear the highest frequencies, we would like to record audio for all ears, especially ears that have been well looked after. For this reason, we would like to keep all the audio to 22 KHz, and maybe a little extra. The common sampling rate was actually 48KHz which would give you up to frequencies upto 24KHz. “CD Quality” audio is actually slightly cheapskatey. However, this isn’t the real reason that you want to go up to 96KHz. The real reason is that you want to push the extra “processing” onto the digital circuitry.
See, in order to get good playback at 44.1 KHz, you need really nice low pass filters. These are theoretically possible, but expensive. If you don’t, you will either chop off some of the high frequencies that are present in the digital audio, or you will introduce aberrations. One way of fixing this is to upsample the audio using digital filters (to, say, 96KHz), then do a cheapy lowpass filter. The idea is to push the cost onto the digital electronics which is constantly getting cheaper, as opposed to the analog electronics which is always going to be expensive (for the same quality).
But the thing is, why bother with the upsampling? Your 96KHz audio only goes upto 22KHz anyway, so it compresses really well. Just record and playback 96KHz audio! Your low pass filters are cheap and don’t introduce any funny noise and you have plenty of headroom. The question is less “why” and more “why not”?
The other side is the bit-depth. 16 bits gives you 96Db, which means a CD has the dynamic range to produce audio volumes from a whisper to a lawn mower. This is actually really great, but some very high end amplifiers are better. They can get to around 100Db. You want your recording to be of a higher quality than any sound system could possibly produce right? If you want some headroom, you want to add a couple of bits. Instead of 16 bits, maybe 18 bits. So why 24?
Because when has anyone ever played back audio completely flat? People tend to boost the bass or otherwise futz with their music before playback. When we look at the “war on loudness”, the other side of that coin is that not only can you not “uncompress” a “loud” CD without losing quality, you also cannot “compress” a “quiet” CD without losing quality. The fact is, you want to hear your music in a variety of contexts. Sometimes you’re at home and everything is quiet and you really want to enjoy your music at its fullest. Other times you’re on a train and you need to process the shit out of your songs just to hear them. In a world like that you really want some headroom. 24 bits is more than you’ll ever need, but it’s a convenient 8-bit boundary, and far more comfortable than 16 bits.
In conclusion, it’s theoretically correct that you only need 16 bits “worth” of dynamic range, and 44KHz “worth” of sampling rate, but for practical playback of audio, the amount of preprocessing most people do means 24/96 audio is far more appealing, and a far more flexible audio format. So the next time someone pooh-poohs 24/96 audio, send them here.