songwriter wrote:Though if my levels are right it isn't causing any clipping and so doesn't fix the problem of my perception of my mixes. I do already do the filter on all tracks to remove unnecessary bass.
Keep in mind it doesn't need to clip to sound worse. I mean, if it clips, it will sound unusably
bad because you'll have digital noise on the recording, and there's no way to fix that other than cut it away.
But also by simply working hot , you end up constantly near the performance ceiling of the A/D converter. For concrete, physical reasons, like temperature changes with higher voltages etc this may cause a little less accuracy.
Converters are calibrated to a set of parameters which the designers of the surrounding analog circuit attempt to maintain (for example, a base reference voltage at a given temperature). These calibrations are based on an average expectation of usage.. which is not to keep the machine working at the ceiling of its operational envelope. Pretty much like cars engine aren't tested by running them near 6500 revs (or whatever's the max) all the time... it's just no way to drive a car - and it's certainly the way to make sure it doesn't last long :)
So the more you push the converter, the harder it'll be for the surrounding circuit to keep its environment stable - and you'll end up with a little less accuracy and a little more quantization errors (especially in lower-grade designs - even using the same A/D chip.. that's because less attention may have gone during design time to counter these changes and the stability of the involved electrical components may be lower, to allow for lower price)
These small errors in conversion will hopefully impact only the lowest significant bits of the sample (i.e. the ones which produce the smallest differences between the numbers they represent) , which in turn represent higher frequency info.. leading to a slight, barely perceivable loss of detail (and hence clarity) once the signal is D/A converted. 22 or 23 bits give such a large resolution that the effect is minimal, but when summed up over many tracks, it may lead to noticeable, perceivable effects.
I record in 48k 24bit for the extra headroom always.
Just to make sure - 24 bit is 24 bit regardless of sampling rate, there's no difference in headroom using 44.1, 48, or whatever. Sampling rate only drives the amount of information that you capture (and, in the real world, where the low and high pass filtering, that cuts out the frequency band to sample, occurs). 20-20Khz is the "human" audio band. but at 48 may that band will be "cut" a little better (and hence result in less aliasing).