Hi guys,
I'm trying to build a patch that can do "clusters" of sounds.
I mean, I'm trying to do a big chord with 128 indipendent osc~ (like if I'd play the piano with arms on the keyboard... a cluster!)
But when I use multiple osc~, naturally the 128 sounds mix together into a unique sound.
Is there a way to avoid the osc~ mixing, to obtain a "polyphonic" sound instead of a mix?
Thanks

Don't play chords? I mean, chords are often perceived as a single entity, particularly when all the notes are played at the same time and have the same timbre. Clusters in particular have that effect. I remember reading somewhere (a David Cope book, I think) that with clusters, the listener usually hears the highest and lowest notes, and everything in between is just...well, clustered!

In addition, sine waves don't have additional harmonics to back up their fundamental frequency. So it's possible the ear is trying to make relationships between the sines and interpreting some as partials of another note instead of it's own frequency. Maybe creating a waveform in a table with the sinesum message using a few harmonics might help a little, but with 128 notes, I kind of doubt it. 128 notes is a bit much for the listener to perceive individually all at once, anyway.

Hi Maelstorm, you're right. But my problem is more basic:
When I use 2 osc~, one in R-inlet and the second in L-inlet of dac~, I can hear 2 distinguished sound, even if at similar freq - i.e. 130Hz and 131Hz. Instead, when I put both the osc~ in R+L-inlets of dac~, the 2 freqs are mixed together producing "beats". So I call "polyphonic" the first example, not the second.
Now, if I use a sample of sound, and I play it through a MIDI device (i.e. a keyboard, or a virtual device made with PureData) I can play many "notes" at the same time and the result will be "polyphonic" again.
So my question is: to obtain poliphonic effect, have I to use a sample through MIDI notes, or is there a way to use many osc~ objects in polyphonic effect?
A friend of mine said to me that in Max (sorry guys...) there's an object call poly~, and I read somewhere that in PureData there's something similar call nqpoly~... But I haven't found it in PureData-extended (Mac PPC version).

@adrjork said:

When I use 2 osc~, one in R-inlet and the second in L-inlet of dac~, I can hear 2 distinguished sound, even if at similar freq - i.e. 130Hz and 131Hz.

Well yes, that's just bad acoustics not polyphony. The sound produced by your speakers should also add up to a beating. Polyphony is just the ability to play different sounds triggered by different notes played (cf. here).

I think there is no way preventing more then two sinewaves from mixing at the DSP level. At least when working with stereo sound.

@adrjork said:

Now, if I use a sample of sound, and I play it through a MIDI device (i.e. a keyboard, or a virtual device made with PureData) I can play many "notes" at the same time and the result will be "polyphonic" again.

That's because [osc~] is just a static sine wave, whereas samples are typically complex, evolving sounds (otherwise people would only use one cycle of a sample instead of a few seconds or more). You might want to experiment with modulating the [osc~]s to make them a little more unique.

A friend of mine said to me that in Max (sorry guys...) there's an object call poly~, and I read somewhere that in PureData there's something similar call nqpoly~... But I haven't found it in PureData-extended (Mac PPC version).

[poly~] or [nqpoly4] probably wouldn't help you here. Those are just ways of dealing with polyphony using abstractions for the voice to help ease patching. They don't alter the sound in any way to make it sound more "polyphonic."

Thanks Maelstrom! You are really clear.