dodz Wrote:so, in your opinion, bottom line is that as long as you work in 24bit, and keep a good level on the track that you're bouncing, any rounding errors/unaccountable phenomenon will be negligable. (definately not audible, presumably) and repeated bouncing using the same approach will not degrade with only a few times through.
in my opinion, yes. the way i see it is that you can't do anything about it anyway other than keep your levels good, use good gain structuring and so on - ie use the same principles as in any mixing. of course minimal rebouncing is better, but it's soooo much better than fostex cassette tape 4-track etc :d it isn't worth worrying about really.
there will be some real proper audio snob who says 'it's really awful' etc etc cos they read it on a forum, but if you can hear it in a full mix then good luck to you. on, say, a solo acoustic guitar track then it is a bit different, but then you won't be doing many effects etc type-things on that anyway.
Quote:with the offline thing.....i just want to make sure i'm understanding correctly....so, if a lot of plugins upsample to 32 or 64 before returning to 24, will this have an audible impact?
all such processors that upsample do it whether it is real-time or offline processing occuring. it's all the same to them. in fact - it's all the same anyway. don't get confused into thinking that there's any difference - when you hit play on cubase or whatever, it's all still numbers - it isn't an analogue signal that becomes digital when you bounce it, or something. you apply a certain process (say, eq) online or offline, it's the same thing doing the same thing, giving the same thing.
anyway, the 'audible impact' of the upsampling is that there aren't such bad rounding errors, ie the process occurs at a much higher precision than 'necessary', meaning the final result is much better represented.
Quote:and if so, is there a way around it? or is it just the nature of teh beast!?!?!111
all of it is the nature of the beast man... you just have to work with it. luckily it's nothing like as bad as warbly tape, chronic noise floor etc etc that existed in the old days
anyyyyyway, all this is y-axis stuff and nothing to do with jitter. jitter is *kind of* analogous to rounding errors on the x-axis.
that is, because of natural inaccuracies of clocks, ie the internal clock in your soundcard, samples may not be taken at exactly the specified intervals.
example: say you are recording something at 44.1khz. that means your soundcard looks at the signal every 1/44100th of a second = 2.2675737 Ã 10-5 = 0.000022675737 seconds. then it sees the level is i dunno, -6db, and 'plots' it on the graph that amounts to your final waveform.
now cos of the inaccuracies of the clock, say cos of the nature of the crystal they use or power fluctuations blah blah (clock construction is a whoooole area i can't be fukt to go into and don't know that much about anyway... plus it is fucking boring), it might sample it at a slightly different time.
the quoted clock jitter for my emu 1212m is less than 1 nanosecond - one of if not the lowest of cheapskate-loser-level soundcards. let's call it one nanosecond for arguments' sake, meaning the error on a given sample could be half a nanosecond (!) either way.
so:
sample 1 taken
0.000022675737 seconds elapse
sample 2 taken
0.000022676237 seconds elapse (max of jitter)
sample 3 taken
0.000022675237 second elapse (min of jitter)
etc
so if you were sampling a sine wave, this sine wave wouldn't quite be a perfect sine wave - points aren't exactly where they should be. they are all 'believed' to be at the correct intervals, that is, every 1/44100th of a second, but they aren't (it's an error innit!).
similarly for playing back - all this is getting worked out by your sequencer (which 'listens to' the clock in your soundcard) and added together, then played back. if you had a perfect sine wave in your sequencer (if such a thing were possible in a digital system), then jitter means that it would play back imperfectly.
but
it's fucking nothing is it????!!!?!?? fucking half a fucking a nanopoxybollockingsecond?!?!?!?
so it isn't
exactly audible - that is, it's not like you don't trust your pc to play back something that sounds very very much like what it is supposed to sound like. it's not as if jitter is like pushing your thumb on the label of a bit of vinyl and the pitch is going all over the shop etc etc...
in my own experience, the only time i have ever noticed jitter was when i first put my emu 1212m in and hit play for the first time. i noticed straight away that there was some sense (not even audible exactly, just perceived more than heard
) of clarity about things like reverb. it just felt better somehow. it was sooooooooooo subtle, but so definite. it really blew me away. after the first time it was pretty much unnoticeable as i got accustomed so quickly to such a small effect.
whatever, it's not like peeps down at the club are going to be saying 'ohhh
.... sounds like that amen has some jitter problems'.