another repost from Peanut Road, with a few added comments:
(originally posted Oct 29, 2007)
For those who don't know, 16/44.1 is the 'standard' for recorded sound, as it comes on CDs. The 16 being the bit depth, the 44.1 being the sample rate -- and although I'm somewhat knowledgeable, I'm not exactly qualified to explain all the ramifications and nuances of this subject. However, in rather simplistic terms, it could be said that the bit depth impacts the dynamic range (i.e. from quiet to loud) and the sample rate effects the frequency range (i.e. how high can you go?). Oh, the subject is far more complicated than that, of course, but for recording, that is probably (well, perhaps...) all one really needs to know. If one even needs to know that -- good recordings have a lot more to do with ones ears than ones equipment.
Now, 44.1 is more than high enough to reproduce anything human ears can hear (i.e. up to around 22 kHz), yet many engineers feel the need to go to higher sampling rates. Many of the reasons given are quite unscientific and won't hold up at all; the main one is that it simply 'sounds better.' And maybe it does or maybe it's in their imaginations. The fact is, one digital converter that puts out, say, 96, may simply be better designed than another that outputs 44.1; it's just not a valid comparison. Even the same converter may be better at one setting than another.
The one valid argument (at least for me) is that cutting off at 44.1 (or even 48, which is the video standard) doesn't leave enough room for a clean roll-off at the top. There may be distortion in the higher frequencies. I'm sure I stated that poorly and quite unscientifically! But I do understand there are filters involved that can impact the audible sound. I also understand that with most decent present-day converters, this is not much of a problem. There is no reason not to use a higher rate, of course, other than the fact that it takes up extra storage space. Quite a bit more storage space!
And then there is the 16 bit thing...okay, 16 is quite acceptable, but the majority now probably uses 24 (18 and 20 fell by the roadside sometime back, though they would seem good compromises). As far as I can see, the only real advantage of using the higher setting is that it can result in a lower noise floor, i.e. quieter backgrounds. Good in theory but in practice, mostly unnoticeable. However, there is a nice side effect of using 24, and that is that one can record at lower levels and have less chance of overloading something, creating distortion -- digital distortion is truly nasty and irreparable, unlike the smooth distortion we came to know and love on analog tape.
So, what does Steve use? Well, my two Fostex recorders only record at 16/44.1, so I use that quite a bit. They do it well and I have no problems at all with the results. And, after all, that's what's going to go on to a CD in the end...or it may even end up as a low quality mp3. The Alesis Masterlink, though, will go higher. I have used it at 16/44.1 but I also have no objections to field recording at 24/48, just for a little safety margin.
I'd be unlikely to go higher in the studio, using the computer. It's faster and easier to avoid 88.2, 96, or even (God forbid) 192. Plus, my system now is somewhat set up around ADAT connections which are 24/48 at their highest. And I'm happy with it!
(Btw, if your converter does both 48 and 44.1, there is a good chance that it is working at 48 and dithering it down to the 44.1 standard for output -- I consider this a modestly strong argument for keeping it at 48 until final mixdown.)