digifish_music, bit depth (e.g. 16- vs. 24-bit) is a clear-cut, practical matter about which there is no great controversy or mystery. It corresponds directly to the dynamic range of a recording.
If you don't know in advance what the maximum sound pressure level at a performance will be, and you're there to record it, 24-bit recording offers a definite advantage since it allows you to set your levels very conservatively. Then once the concert is over and you know what the peak levels actually were, you can renormalize the recording and dither down to 16 bits. If the maximum levels were, say, -8 dBFS then you can have a nearly 8 dB improvement in the dynamic range of the finished product, as compared with recording the concert at 16-bit resolution using the same level settings.
I record classical concerts professionally, and am generally the only audio engineer at most of the events I record. To me this approach seems far more responsible (though less exciting, perhaps) than the old way of trying to guess and occasionally, guessing wrong which I did for 35+ years before I changed over to 24-bit recording for most events.
In my opinion 16 bits is usually more than enough for domestic audio, but occasionally some program material comes along (e.g. percussion ensembles) which can really exploit almost its full range, so I'm glad to have it. And for me, 24-bit recording is the low-stress way to get an optimal 16-bit recording--it's not an end in itself. As far as a concrete proof (or demonstration) of this is concerned, that is actually not difficult at all, and if you'd like to arrange something I'd be glad to help. But it's simply not correct to claim that there has been no proof of it up to now--it is an elementary principle which has been very well known for decades.
The question of sampling rate (sampling frequency) is rather a different sort of topic to discuss critically, and I don't want to try to go there in the same message. I'd rather try to establish some clarity about bit depth, since that really is pretty simple.
The only strong caveat with bit depth is to remember that there is always noise in any audio system, and the actual signal-to-noise ratio of any real-world recording (i.e. a rendering of acoustic events, rather than something algorithmically generated) will always be distinctly less than the maximum which the 24-bit format would allow. 24 bits = almost 143 dB dynamic range, which exceeds by far the dynamic range of the quietest physically possible microphones, mike preamps and A/D converters (let alone the quietest ones which are actually available). Thus by definition any 24-bit recording has less than 24-bit "actual" resolution (i.e. signal-to-noise ratio), and it's important not to confuse the container with its contents.
--best regards
Topic: Resample to 44.1 or record at 44.1 (midside Q) (Read 13372 times)