multiple sines from a single sine osc?

[Zolzer]

So i'm working on a synth that ultimately begins with a sine, generating overtones based on user-defined ratios (multiples of the fundamental frequency). I want to build a modifiable float array that will store frequency and phase information for each consecutive sine, but I haven't gotten to this yet. I first need to know if it is possible to create multiple sine outputs using one sine oscillator, since I have found that sending multiple frequency links to the input results in the latest link overriding it. any clues on how to create a loop in which one sine can have many harmonics? perhaps drawing info from the float array post-output of the sine an then retriggering the oscillator somehow? any help is appreciated lol i'm relatively new to this

[MyCo]

You can do it in code or assembler. But this method doesn't make sense for huge amount of sines. For a huge amount of sines, you should take a look at the "additive oscillator" in the toolbox. But this has the limit, that you can't modulate each individual sine.

[Zolzer]

I will have a look into this and I will do my best to set up what I think I can manage so that I can share screenshots. I will likely have more questions, but I will take a few days to look through the manual once more and I will try to write the memory function of the code for this synth (again, i'm newwwwww). i am still wondering, however, if anybody can tell me if there is an easy way to code in a new sine generated at a particular frequency and phase using DSP? i get the feeling that this is something nobody regularly asks since you normally either perform subtractive synthesis on saw waves or simply program/reverse engineer clip distorts. all i need to know is, if i have a sine A (440hz), do i have to use this to generate overtones in DSP or is there an easy way to make a new sine materialize with its own definable frequency and phase? thanks for the help guys, much appreciated. i will let you guys know the progress later on the sin-number/phase arrays. if it's any incentive, this synth will be FREEWARE lol

[MyCo]

Right now, you can't pass arrays into code, so you can't generate multiple sines with different parameters in an arbitrary number automatically. However, I've attached some possibilities that FS has to generate 4 sines.

If your parameters for your sines are generated using a formula, you can use code to loop and accumulate the sines. You would have to implement the formula for the phase, frequency and volume in code.

[tester]

Hmm... Looks interesting for one of my projects too. Thanks!

[Zolzer]

Sounds like an interesting assignment

Thanks for the replies, big help and I will look at this later tonight when I get back from work!!

[tester]

A quick question MyCo. Will be there a noticable difference in performance (CPU requirements) between using separate sine modules and multiple voices on one sine module (example 2 in your schematic)?

[MyCo]

There is CPU difference but it is not so easy to calculate. This version uses less CPU, when there is only "per Oscillator processing" in the poly section. This is a very special case. Normally you add several oscillators together in a poly section, then pass the sum signal through envelopes and filters.
With the multivoice setup you pass every sub oscillator through its own filter and through its own envelopes. This means there is much more processing involved. But if you want this kind of processing, then this is the way to go, because it uses less CPU than in a setup where you duplicate each and every module in the poly section.
eg: You want to have a setup where each oscillator has its own filter and it's own envelope and everything is running separated til the end of the synth voice processing, where it's finally mixed together with the "poly to mono".

[tester]

Thanks for explaining.

In one of my sub-projects, right now I have a section where multiple independent sines are in use (stream flow, not poly, c.a. 20 sines per clone), and I was wondering if changing that segment into poly processing would save me some noticable CPU. That why this topic evoled my attentnion. I generally work on streamy things (different purpose, not synths), so I have no experience with poly in practice.