More bit-depth equals more headroom?

[jgiannis]

I've read it over and over again, that more bit-depth equates to more headroom. But I don't understand it; at least not in the practical sense.

As I've read/heard, the traditional method of recording 16-bit audio at 44.1k, one would try to record as loud as possible, but without clipping. This would produce the greatest signal-to-noise ratio. However, the risk would be that your performer could clip at any given moment because you are sacrificing headroom for greater signal-to-noise.

First of all, what "noise" does this scenario refer to?

1. Quantization noise?
2. Aliasing noise?
3. Internal hum/buzz from the recording equipment?
4. Background noise from your environment?
5. Other?

Secondly, how does recording louder produce a higher signal-to-noise ratio in each of these 5 types of noise? (Although #3 and #4 are obvious.)

In more modern times, I keep hearing about recording methods and recording devices that have huge amounts of headroom due to increased bit-depth. For example, when recording in Pro Tools, setting the parameters to 24-bit, 48k supposedly produces greater headroom. I also hear of portable recording devices that, as the box will say, has over 100dB of headroom. In the practical sense, I don't understand this because wouldn't you still want to record your signal as loud as possible while avoiding clipping? If so, then where's this "greater" headroom?

I think I have some answers to some of my own questions here, but I don't feel like I have a firm understanding of it, so I'd rather hear some answers from others.