gensyn~.pd
gensyn~-help.pd

gensyn~: generic, multi-modulator synthesizer abstraction

REQUIRES: zexy

what makes this synthesizer abstraction different is:

In addition to the common access to adsr(filter and amp) and vcf~(cutoff and frequency)

It also includes

• 5 available modulation types for two wavetables
  • additive and frequency, phase, pulse, and amplitude modulation
• 8 possible types for each wavetable
  • sine, square, triangle, saw, random, duty, pink noise, noise
    • the random is by a settable number of partials
    • the duty is a square wave with settable dutycycle
• access to two unique parameters per modulation type

Note: The -help file has a substantial number of examples. Especially, interesting are those changing modulation type, wavetype, parameters (per mod), partials, and dutycycle on a per-voice basis by both formula and over time.

I hope you enjoy it, get some use out of it, or it just makes you smile or giggle.

Sincerely,
Scott

p.s. no picture necessary. And personally, I find this Very funny. -Peace.

gensyn_b.zip

Via @Valuska-Yoshimitsu's comments and some work I have been doing to make Gemini/Jass~ a raspberry+arduino piece of hardware, I have made some improvements on gensyn, including but not limited to the following:

1. The code is now better encapsulated to separate the controls from the audio processing;
2. The cpuload has been improved (dropping on my computer the original 6% cpu load to 1-2%, putting it back in range of a Pi 3b+)
   (this was done by turning each voice's (audio processing) switch~ on when velocity >=0, and off when the outgoing env~=0)

As to your comments @Valuska-Yoshimitsu:

'First, the $1 creation arguments on the wavetables are not assigning the $1 float to the table id, instead, the wavetables are named "$1-wavetable" If I rename the wavetable to, for instance, 2-wavetable, then everything writes correctly.'

This occurs because the gensyn is designed to be used as a clone, in which case the $1 is found by the clone object and then created.

The implication is that each voice (i.e. in terms of poly) can have its own set of wavetables. Check out the example: "CHANGING VALUES OVER TIME > randomize wav1/voice every 2secs" to see what that sounds like. Sort of like Chaos , but personally I like that sort of thing and since all of the other parameters are also /voice based they could be tuned to make the differences more smooth.

'...switching to random partial generation for all gensyn clones in the help patch causes a huge lag, after which everything works. (this one could totally be a system issue, but I thought I'd mention it as long as I was commenting here.)'

I think this was probably do to the "pd-float:round to 2048" warnings it had been throwing. I have resolved those errors in the latest gensyn_b version(above). Thank you for inspiring me to resolve that long standing issue.

Please, Do let me know if you come upon anything else or just want to discuss the work.

Peace through sharing,
s

@svanya this is rad! thanks so much.

I wonder, though, if there's a way to re-write those tables so that they take creation arguments from the abstraction so that I can call a bunch of abstraction separately and have them still work I've been digging through documentation but just can't seem to find out how to generate the abstraction ID.

Again, thanks, I love your work here.

oof, pardon my lack of periods and commas above

@Valuska-Yoshimitsu

Given the way clones work, you can send each table/abstraction their own values by send the [r msgs] in the form [abstractionId parameterName value(.

So, for example, if you want to change the first clone's mod-type to the 3rd type you send a list to [msgs] as [0 t1 2(.

If you open the examples abstractions, they will point you to how to write the list for each abstraction/parameter pair.

Having each table have its own ID is handy, because then you can change the tables on a per (clone) voice basis.

p.s. if you want to change all of the voices/clones to a certain value, you send to [msgs] [all parameterName value (.

p.p.s. Thank you, for your support. I am glad you like it/them. Keep sending the questions, as helping to clarify for you may also do the same for others.

p.p.p.s. am currently working on a version that further abstracts the abstraction to allow for multiple [clone] objects simultaneously. will share it...prob here...once it's stable.

@valuska-yoshimitsu

Rereading yr comment...

You could also write independent, non-clone abstractions as

[gensyn 0] [gensyn 1] [gensyn 2] etc. and then all the tables would take those $1's into account and you would send in msgs to each one (on their left inlet) as [t1 value(, etc.

That is, if I understand your desire, should work, too.