We read about digital audio devices boasting bit depths up to 32 bits and understand more is better. But better at what?
Bit depth is a measure of the signal to noise ratio (how much louder the music can be than the background noise) and also tells us dynamic range (the softest to the loudest levels possible). The measurement used is dB (decibels) and just to give you a reference, every 6dB is twice as loud as the last 6dB. A 16 bit system has 96dB of range, a 24 bit system 144dB.
I am going to give you one more figure and then onto the point of the article: the human ear can discern (on a good day) 124dB which is a mere 21 bits. So why do we care about higher bit rates like 32 and beyond? Does it really matter?
The answer is, as usual, not so simple. On the one hand I would tell you that right now, with 99.99999% of recordings and audio equipment out there, no. Could it someday matter? Yes.
Here’s the deal. As DAC designers have pushed for more bits to extend the dynamics and increase the S/N ratio, recording engineers and amp/preamp designers are still living in the past because their maximum loudness levels are identical and fixed. What’s actually improved is the other side of the recording: how quiet it can get.
Think of it like this. Imagine you’re in a noisy cafe listening to a musician playing his guitar in the corner of the cafe. When he sings loudly you hear him rise above the noise of the crowd and that sets the maximum level he can sing at. Now imagine the crowd leaving and the background noise dropping to a very quiet level. The singer still hits the same loud notes but now you can hear all the small nuances you missed because of the background noise. That’s what happens when we go from 16 bits to 24 bits.
In this analogy you’ve gained the ability to hear many small nuances from the singer – but not all. You can’t, for example, hear his heartbeat or even the slight movements he makes in his seat. These are just too quiet to hear – but in fact – going from 24 bits to 32 bits buys you enough quietness to where you should be able to hear them – but that extra quietness is wasted because you can’t hear it even if it was on the recording (which it is not).
If, on the other hand, system engineers and recordists went the other way – adding dB’s of maximum loudness instead of quietness – then when a waiter in the cafe knocked over a pile of plates, the dynamic range you got would be so realistic and perceivable that you’d be startled out of your chair; the sound so much louder than the singer.
Technically, we already have the dynamic range needed, we’re just not taking advantage of it and it is wasted.
Tomorrow we’ll examine why vinyl, which has even less dynamic range than a 16 bit CD, seems to have greater dynamics.