pouët.net

MeteoriK: I have no idea what the musicians did to make drums. You'd have to ask lug00ber and Optimize :-)

trc_wm: That's a bit strange, perhaps; I know I got it tested on an Intel Mac quite recently, and it ran just fine in PPC compatibility mode. Perhaps 10.6 breaks it, though.

Basically,., when I was doing that quiz, there was that distinctive smell of piss.

As I got closer to maali, it became quite succinct,. Mayby this guy doesn't go to the toilet. The stain down his pants told the story.

He's an ugly piss-riddled long-haired bastard, and he masturbates when people call him that! l

This thingy might be related to the topic: http://es.0rel.com/. It is not finished yet though...

very neat, 0rel :)

Great sounding stuff, 0rel. Can't wait to get my hands on this!

For those who are not codingly inclined, there's always picatune for exec music compos.

@0rel: 0rel, going to contact you :)
@linde: picawhut?

0rel: How small is that? My first attempt at a synth was along the lines of "csound with a gui", but it just got too big, so I lowered my expectations of myself and made it simple.

@knl: modular synth/tracker thingie that outputs small exe files

Will google about it.

Whatever you do, remember to put a DONK on it..

http://www.youtube.com/watch?v=ckMvj1piK58

Thanks for the motivating words! - Just found out about this a couple of days ago. It seems to have almost the same purpose as ES(lib)... Oh well...

0rel: modular synths sure are fun! ;)

I've recorded some "noise" this evening. and a screenshot. hurray!. .)

st2-grainy_days.ogg

and that's how the audio routing for this experiment looks like

0rel, yep, I immediately thought about msp or pure data when I saw your screenshots :)

0rel: you've got mail :(

Is anyone, besides Meteorik, using alias-free oscillators?

Yes, in two different ways: BLEPs and wavetable resampling. BLEPs are really not that hard for simple saw/square/osc sync stuff, but the code gets big fast, maybe too big for 4k.

BTW. not alias *free* exactly, that's hard.

trc_wm: Most softsynths probably have at least some way of doing oscillator antialiasing. The differences are in flexibility, CPU usage, code size, and how well they actually work once you start getting into higher frequencies.

Sesse: many softsynths, including commercial ones, do a poor job of anti-aliasing. I was wondering how many coders do "proper" oscillators in 64k engines.

Speaking of Pure Data: http://www.rjdj.me/. Basically it's Pure Data Vanilla on iPod/Phone/Pad, + the sensors (accelerometer, GPS, microphone, ...).

erm, wass?

trc_wm: OK, well, it defines again on where you draw the line. SaneStation (the synth used in Nemesis) uses wavetables that are certainly anti-aliased (including correction for the Gibbs phenomenon), but once you start getting into things like FM (or even pitch bend) we probably start aliasing. (Even proper _analysis_ of FM's frequency content is not all that simple, much less actually implementing it right from the start.)

Another example: The V2 antialiasing is essentially the raw waveform convolved with a (short) rectangular window; it aliases heck of a lot less than just using the raw waveform, but is far from theoretically ideal — is that “proper”?

I draw the line at an "alias-free" static tone, regardless of the frequency. Is implementing generic "alias-free" FM even possible in real time?

I don't consider a rectangular window sufficient for "alias free" oscillators as it's first sidelobe is just 13 dB below the main lobe. Granted, it does reduce the audible aliasing quite a bit.

What is mostly neglected is that post processing the synth audio by, for instance, a multi-band compressor, can make the aliasing much more audible, especially if the filter is wide open.

Yes, you can do alias-free FM using IFFT. kb mentioned a Yamaha paper in some talk; it's supposedly a lot of work to set up and get right.

That doesn't sound like fun!