BUG (?) Mapping to amount of another mapping

bibenu

I've come across this a couple of times, at first I wasn't confident there wasn't some bug in my patch .. now I'm fairly sure there's nothing wrong with what I'm doing.

Concrete example: I have Map6 routing Mod1 (LFO) to filter cutoff frequency. But it's a very intrusive effect that I want to control from the macro knobs and I choose chaos. So, starting from the matrix, I long-press 'chaos', choose to create a new mapping, and choose 'mapping 6', and then parameter 'amount'. Sure enough, I now find that with the chaos knob, I can now move the tick above Map6 'amount' to far right (chaos at max) to center (chaos at min). BUT THIS HAS NO EFFECT ON THE AUDIO .... the LFO modulation does not get through to the cutoff frequency unless I also change the fixed 'amount' controls (on Map6). Even then, 'chaos' has no effect on the modulation depth.

Surely this is not the intended behaviour ?

Any help would be appreciated. All I want to do is control the depth of the modulation from one of the macro knobs ... and I thought that routing a macro knob to 'amount' on the mapping (controlling that modulation) must be the way to do it.

join

It's not a bug, but you aren't doing anything wrong either. Let me explain.

The mapping system works in a sequential fashion. This means that MAP1 is computed before MAP2, which is computed before MAP3 (and that's the same with modulators, by the way).
The mapping you created by long-pressing has surely been put at the next free spot, say MAP7, but it acts on MAP6, which has already been computed! So you can see MAP6's amount parameter move, as when all mappings are computed, yes, MAP6's amount has moved because of MAP7. But by the time the synth arrives at MAP7 and applies its action, the action of MAP6 had already been applied onto the filter's cutoff, so even though MAP7 moves its amount just after that, it's too late.

So for this kind of thing to work, it's important to keep in mind that the synth treats things sequentially, meaning that if mapping X has an effect on mapping Y, then it should be ordered before mapping Y.
In your particular case, just swapping MAP6 and MAP7 would resolve the problem (yeah, easier said than done if you don't have the PC/Mac editor handy).

But there is a much simpler way for that kind of thing, using sidechain.
Set MAP6 amount to 0, but have its sidechain source from the CHAOS knob, and set the sidechain amount so that it covers the range you want.

So yeah, it's not that it's a bug, the engine works that way, but it certainly merits a mention in the manual, and maybe the "create mapping" screen should be smart enough to say "hey, we're modulating another mapping, so it'd be better to swap that mapping and the new one we're creating, just so that things could work" (it could be a bit more difficult than it sounds to implement this).

bibenu

@join said in BUG (?) Mapping to amount of another mapping:

The mapping you created by long-pressing has surely been put at the next free spot, say MAP7,

Correct,

In your particular case, just swapping MAP6 and MAP7 would resolve the problem

Okayyy ... will try out that and the sidechain thing ... Thank you.

but it certainly merits a mention in the manual,

Either that or perhaps something to prevent it happening. Like, disallowing the modulation where it can't possibly work.

I can think of three other ways to improve the user experience off the top of my head.

Since the scenario I'm thinking of must be pretty common (first set up a mapping, then map a macro knob to 'amount' on that mapping) maybe evaluate the mappings in the opposite order ?
Perhaps, determine the order of mappings evaluation according to how they're connected. So in the example I've given, because map7 controls the amount of map6, then evaluate map7 first, then map 6, then all the other mappings (assuming they are not interconnected)
or ... since it can be done with sidechain .... disallow the mapping to amount altogether ?

I haven't really thought this through ... I can see a flaw with (2) already, in that it could never work if there was a circular mapping (map 7 modulates amount of map6, map8 modulates amount of map7, map6 modulates amount of map8 .... does that possibility exist ?).

EDIT: Actually ... this constraint of a mapping only being able to modulate the amount (or min, max, smoothing, scale ...) of a higher numbered mapping is sure to trip people up. It's very hard to set things up right, given those constraints. Say I have mapping 1-10 already, and I want to set up modulation of my filter by lfo (mod1) with 'maximum value' determined by 'chaos'. I first have to create mapping 12 to control filter cutoff, then I have to create mapping 11 to map 'chaos' to the 'maximum value' of mapping 12, so that everything fits the evaluation order. Allowing mappings to other mappings has a lot of potential for confusion, surely ?

Sorry ... I'll shut up and let you think about it.

bibenu

@join said in BUG (?) Mapping to amount of another mapping:

But there is a much simpler way for that kind of thing, using sidechain.

Ah that works perfectly, doesn't involve another mapping. Thanks.

join

Yeah, these are good solutions too. Many times during design I've pondered whether to do solution 2, but I wouldn't because it is sort of complicated to pull off right… but now that we're seeing that mitigating the downsides of the simplest way (sequential evaluation) could be complicated too, maybe it's worth a shot after all.
I'll think about it. Don't hesitate to shoot ideas, I won't hesitate to steal them :).

edit: Evaluating mappings in the reverse order could maybe also work with minimal effort. For modulators, an evaluation in the correct order is better, but there will be cases where you want to add something in the middle of a series of transformations, and the problem will show up again.