On Fri, 2004-10-01 at 06:38, Bill Schottstaedt wrote: > > this is really not currently feasible in RAM. > > But that's irrelevant from Snd's point of view -- it normally only > keeps 8K samples (or is it bytes?) in RAM for each open channel > or mix -- are you thinking of some other editor perhaps? One > of my standard "regression tests" operates on a 6 hour stereo > 44 khz sound. strange.... my RAM usage goes up by an amount that appears to be equal to the wav file size, when loading a file into SND. of course, if the file size is large enough to slop into swap space, things get very slow. furthermore, and this is really disturbing, that memory is not released either when that file is closed, or when i quit SND. > > Anyway, I looked at the spreadsheet-for-sound business, and I > think I can make it work in Snd without great pain -- you could > use it that way now, but not easily. I'll poke at it further. > awesome! -p
