On 8/31/07, Martin Ahnelöv <operagasten@gmail.com> wrote:
ons 2007-08-29 klockan 23:11 +0200 skrev Kassen:

>         dur => blackhole would be exactly like SinOsc => blackhole,
>         but with
>         time. When you, for example, chuck second into blackhole, the
>         shred will
>         forward 1 second in it's own timeline (A bit like the -s
>         otion, if I
>         recall).
>
> I'm still not sure exactly how this would be different. What would
> happen to playing Ugens that would be in the shred? Would those keep
> computing?

The thing is... Hm, if we have something like this:

SinOsc 1 => dac;

while (true) {
        100::ms => now;
        3::samp => blackhole;
}

The shred would jump 3 samples forward every 100 ms. If you change the
duration to 10::ms, you might get some glitching going on every 100::ms.


Ok, I get it now, you'd like to jump into the future... That should be possible but I think it will often lead to a big strain on the cpu. For simple oscilators or a envelope it could be quite easy but if the shred involves -say- three oscilators in a fm feedback loop there would be no way around calculating all the samples inbetween to know what value the final one is at after the time we skip..

It's a interesting idea but I'm not sure the benefits outweigh the dangers and problems.

At least you could record your sounds like that, for example recording them to a LiSa and turning record off at a certain point, advancing time, then resuming recording at the spot where you left off.


Kas.