Would it ever be possible for pianoteq to use memory as well as CPU to 'share' the load? I understand sampled virtual instruments use memory and wonder whether it would be possible for pianoteq to use both memory and CPU to enable less CPU load?
I think this is a bit like me asking you if it would be possible for you to drive faster in your car by also using the boot as well as the engine...
Nope. Pianoteq doesn't use as much memory because it has nothing much to "read" really. All the sound is created by computing the mathematical model in realtime, so that's why a fast CPU is prerequisite. This is the main and obvious difference between sampled and modelled instruments.
BTW, I saw Pianoteq run on i7, full 256 voices (heavy bass notes while holding the sustain pedal), 4 ms latency, 10% CPU load across all cores. Amazing
Are you suggesting some sort of pre-calculating and caching the results for rapid access "soon" ?
I can see how it COULD be done for "recording" in a delayed sort of a way, but I can't see a PRACTICAL way for it to anticipate my key strokes ahead of time.
I'm predictable and have a very limited repertoire, but not THAT limited (-:
BTW, if you can wait just a FEW months - faster processors will be available.
EVERY few months from whenever you ask the question.
or
"The problem goes away over time."
Yeah, soon there are gonna be some cheap 6-core solutions from AMD.
And in 2011/2012, 8- and 12-core by Intel.
Why cutting down calculations, then? INCREASE THEM, MODARTT