Limiter class or another solution

jamshark70 · February 8, 2021, 5:03am

Are your synths being freed? See references to doneAction in EnvGen help.

hjh

Davon666 · February 8, 2021, 5:03am

there are on doneAction:2 yes , no over loading in the plotTree

Davon666 · February 8, 2021, 5:04am

there is a lot happening but it all run fine wihtout companderD (just volume not flat as you know =D)

Davon666 · February 8, 2021, 5:06am

I tried running just one CompanderD by group of instrument and by instrument individually it all crashes the same. I use compander specifically for compressing before it works just fine but when I try to add another level for limiting volume i loose it all

Davon666 · February 8, 2021, 5:16am

initially once I compressed ,sidechained to kick, my sounds I want to draw a line of volume that cannot be passed. jamshark70 can I do compression + limiting with the Amplitude code you gave me ? i m gonna try anyway but I suspect that these compress using envelope but dosent limite volume

Davon666 · February 8, 2021, 5:51am

o i tried instead to boost volume of the first compressor that uses compander, sidechained to kick, and after put in the limiter. that seem to work with 2 instruments im gonna try with all project a bit later and update =p

Davon666 · February 8, 2021, 7:01am

well not a solution… jeeez how come this is so hard to do

jamshark70 · February 8, 2021, 7:40am

how come this is so hard to do

Going back to a comment from before: If the problem isn’t stated clearly, then the coding process is basically throwing code at the wall and hoping something sticks. The probability of this moving efficiently toward a good solution is extremely low.

TBH at this point I have only the vaguest idea what you’re trying to do. It looks like maybe you’re trying to do a sidechained pad technique, but there’s something about the sidechain input that isn’t working well.

You’ve thrown out a lot of very short, scattershot posts. Each of them has a little information, but the context for each bit of information is not clear. You haven’t posted any code examples.

The likelihood of getting a clear answer for an unclear question is pretty low.

hjh

Davon666 · February 8, 2021, 7:52am

all right i bring the codes, sorry this hasnt been taken seriously on my end

Davon666 · February 8, 2021, 9:40am

Here please haha

and there is a MouseY or X to apply cool gating

github.com

davon6/superCollider/blob/master/extract.scd


(
s.waitForBoot({


	var numSegs;

	"buff init".postln;

// tempo wanted

	t = TempoClock.new(151.703592/60).permanent_(true);


	s.meter;
	s.plotTree;

	//necessary for the buffer of the shaper to work
	~tf = Env([-0.8,0,0.8],[1,1],[8,-8]).asSignal(1025);
	~tf = ~tf.asWavetableNoWrap;

This file has been truncated. show original

Davon666 · February 8, 2021, 9:42am

haha the wink is actualy the ; ) at the end of the statement.

I d rather send the file if thats ok

jamshark70 · February 8, 2021, 1:46pm

You sent a link to the file, that’s good enough.

It’s… 500+ lines…

There’s a concept in programming called “minimal working example” – try to reproduce the issue with the least code possible. The reason why we do this is because it’s really hard to make sense of a big block of code, so it helps a lot to focus the code example on the specific issue.

Another testing methodology is to add elements in, one by one – presumably it won’t crash with just one element or two, but at some point, you’ll add in an element and it will “crash” (not clear what you mean by that btw) – so then it’s something about that element that causes the trouble. Or remove them one by one until it stops crashing – then the last one to be removed is suspicious. That’s a form of “differential testing.”

At this point, I personally don’t have time to pore through 500 lines. If you could reduce it or at least point to specific areas, that would help.

hjh

Davon666 · February 8, 2021, 1:51pm

all right i keep digging thanks it was over 2000 initially i only kept whats mandatory for the all thing to run also i am programmer and am used to the troubleshooting debugging technique you mentioned or piece per piece.

I understand u dont want to trouble yourself with this file simply to say it runs with one cmd + enter or click haha

jamshark70 · February 8, 2021, 2:06pm

Maybe tomorrow I could try just running it, see if the crash reproduces on my machine.

Oh, and code bits should go inside backticks: inline, use backticks like single quotes, or for longer blocks:

```
write your code here
```

… same as github. I’m sure there’s markup documentation somewhere on the forum.

As you’ve recognized, the copy/paste of the code above is pretty well disastrous. Might as well edit and delete – nobody will try to clean that up. But it’s not a big deal because we can get the code from the link.

hjh

Davon666 · February 8, 2021, 2:10pm

thanks I say it runs as it is simply volumes wont settle to one value (even if total flat isnt possible) there is too much gaps and yes i dont know why but companderD crashes once I play few instruments together, but its not implemented in this version

jamshark70 · February 8, 2021, 2:40pm

So there are two issues… One, level “won’t settle” where I don’t know what it means. Two, crashing with more instruments. (By “crash,” do you mean that a process quits unexpectedly or something else…?)

There’s a reproducer for #1 but not for #2.

I’m not sure I can be of much more help. A technical issue is usually “I expected/wanted X but I’m getting Y.” X, at least, is not at all clear. I’m still not even clear about the context. Are you talking about sidechaining? Or volume limiting the final mix (“mastering”)? Or something else entirely? I keep asking about details because the correct solution depends on the context. Trying to guess what you mean is a bad way to solve problems.

Anyway it’s almost 11 pm here, so good night.

hjh

Davon666 · February 9, 2021, 12:56am

i ll reupdate in a bit but it might be fixed.

Thanks jamshark

Davon666 · February 9, 2021, 4:13am

hey thanks heaps for your help

I figured the plop silent / crash happen less after using a bunch of thoses
(
Server.killAll;
Pdef.all.do( .clear);
Pbindef.all.do( .clear);
Pbind.free;
Pbindf.free;
)
i am taker if there are other ways to clean the server from left overs

also one of my instrument seems to be faulty im working on it now =)

thanks a lot

jamshark70 · February 9, 2021, 11:49am

Hi, please remember to use backticks for code.

Right, after you start some processes, it’s your responsibility to stop them / clean them up. Exactly how to do that is a big topic. My own strategy is to package everything for one “process” into an object that allocates everything for that process (buses, buffers, groups, infrastructure synths), and releases them all upon .free. So there’s a single point for initialization and cleanup, and I can start/stop/recreate objects at will, for several hours, without reboots. It took a long time to evolve this workflow but it helps a lot.

Saving things in Pdef etc and clearing them all out is a good starting place.

Glad to hear that the situation is improving

hjh

Davon666 · February 9, 2021, 11:00pm

i will keep the backticks in mind, also for everyone I got some noise artefact or the “plop” noise that crashes CompanderD if a modulation faster than 1hz is done on the rq of BHiPass4

bit specific but it does not happen anymore after correcting