what sample rate do you use?

i have a question for anyone with some knowledge on this subject and on pro tools, Pro Tools 7.4 is introducing something called elastic time from what ive gathered basically it makes it possible to stretch out audio if this is true then would we not then need to use a much higher sample rate as the lengths between the samples will be stretched along with the audio?
 
the elastic time feature doesn't really have anything to do with sample rates - it's the new and upgraded version of the time compression/expansion tool

also, i would avoid upsampling whenever possible. i had to upsample some tracks from 48k to 96k the other day, and there were a lot of noticeable artifacts from the sample rate conversion....good thing they were just scratch/reference tracks.

think about it - if you go from 48 to 96, or 44.1 to 88.2, you're inserting an extra 1 or 0 behind every one that exists already, but your DAW can't be 100% sure which is supposed to go in there, which in the end will degrade the audio. the more sample rate conversions you go through, the more artifacts you're going to end up with.
 
...but your DAW can't be 100% sure which is supposed to go in there, which in the end will degrade the audio...

Nonsense - unless your computer is incapable of simple mathematics, you can add data and not change the interpolation. Say you're plotting a line, and you want to represent it with two points - you pick 0,0 and 1,1. Now, all of a sudden, some jackass comes along and wants four points on your line - what do you do? You figure out two more points that would be on the exact same line - add 2,2 and 3,3 - and you've 'upsampled' with no change in the function you're interpolating.

Jeff
 
of course it seems that simple...but i can absolutely guarantee that there was some weird jittering and such on my upsampled tracks that didn't exist at 48k
 
i was also upsampling at a factor of 2

i recorded some DI guitars in cubase @ 48k, then upsampled them to 96k

then i took the upsampled tracks to a PT HD rig, imported them, and tracked drums @ 96k. while listening back, i noticed more than a few anomalies in the 96k guitar tracks that don't exist in cubase @ 48k. maybe it wasn't a result of the increase in sample rate...but i don't know what else could have caused it
 
I use 44.1.

As for the bitrate, it's something I've never been sure about - my Behringer desk (not very glamourous I know) has options for 16, 20, 24, or off altogether. I leave it set to off, as I'm guessing it applies more to if you're tracking directly to a tape deck, HD recorder or whatever? Then I set the sample rate in Cubase instead - currently at 32 bit float. Is this a good idea, or should I be doing something different?