Split mode not working properly for FX

[jhlz]

Hi,

I searched already the board but couldn't find a topic about this.
In the manual it is stated that you can have 2 inependent sounds.
Given split mode is on
and A uses no FX
and B uses reverb.
Whenever I turn the morph to B then the sound of A also gets the FX of B like whatever delay, reverb is set on B.

Is this by design or a bug because seeing the different FX values for A and B on the display while morphing wouldn't make sense then as the resulting sound would be always affected by the FX settings of B.

(using Firmware 2.0)

 Thanks!

[modalmixture]

On p 78 in the manual, there is a list of parameters which cannot morph, which includes the effect type.

[jhlz]

Wow, what a pity!

I missed that..thank you 

But actually it morphes back to no FX when I turn back to A so it seems it is more a limitation to the split mode that I cannot have 2 settings for the same FX simultaneously?

[larioso]

I think you found a bug.

The morph is kind of creating an invisible routing between A and B.
And it does not keep track of lower zone voices are all to be A, it seems.

I worked around it though like this.
- in this case I did not use MOD ENV in preset so used that as something that is morphed anyway, indirectly reverb amount
- all sliders down on A side(EDIT A)
- MOD ENV sustain up for side B (EDIT B) it will drive reverb amount later
- A side : I create a destination for reverb amount and a node from MOD ENV and 100% amount
- B side: same as for A-side above

So a morph later from A to B will vary sustain slider on MOD ENV

This works perfectly, having split zones.
- lower zone will not have any reverb as being strictly A side and I can play dry
- upper zone I can vary between A and B and get reverb without affecting lower zone sound

Since this is supposed to do the same with internal invisible morph routing and A side on lower should be untouched by morph movement, as I see it and also confirmed as my special routing works. I see no reason why amount reverb should be morphed on lower zone since stated being all A.

So think this is a bug/oversight for lower zone voices.

In presets using MOD ENV this is not free to use as I did above, but the principle could be made with some other source.

If you want this to be regarded as bug and fixed, contact support reporting it. Forum has no effect.

At least a workaround for you. Have fun....

[jhlz]

Thanks man!

unfortunately I still can hear reverb on A

[larioso]

Then I did not describe it that well.

I did it twice to be sure, and second time without using EDIT A or EDIT B, just turning Morph knob to ends.
- in my test, zero level reverb on A, and zero on B too, by the knob.
- ensure all sliders on MOD ENV are down, so no release which can give reverb as a key is released
- it's only this MOD ENV sustain that indirectly apply reverb through that matrix routing
- I use send fx position
- I do not use PB Connect either, no software if that intervenes, just PB
- I did not use any aftertouch which I often use to increase FX for effect, so no hitting a key affect fx
- I set split point hold timbrality and press C key to have two octaves for left hand

"..I still can hear on A"?
- does that mean same amount of reverb as B or less?
- if some leakage of some sort

So we understand each other.

I get completely dry left hand even turning morph knob all the way to right.
Seems to me morph for matrix routing is working

I will test later and see if I can use Morphé axis X, Y or Z just as well as MOD ENV, or if that is realtime only.

[larioso]

I did yet a couple of times, starting with init preset - and it works just fine.

I describe my process
- init preset
- ensure morph knob all left
- move all mod env slider up and then down again to be sure zero everywhere
- do same with Reverb Level if the used preset had that up
- enter mod mode button
- hold assign 4 and move reverb level for slot 4
- create a node mod env to slot 4 and set amount to 100%
- move morph knob all right
- move all mod env slider up and then down again to be sure zero everywhere
- do Reverb Level up and down so absolute zero on that for B too, ensures no morph is done on that
- move sustain slider all up on mod env
- press mod button and check node is 100% for B side too
- play a bit and move morph knob to see that dry on A and reverb towards B
- stop with morph all on A
- press timbrality over to split
- hold timbrality and set split key where convenient
- play on left zone to hear dry
- play on right zone and move morph to all B hearing extensive reverb
- play on left zone again and it is all dry

It works fine. Nothing played on left zone has reverb.
Hope I did not forgot to describe anything.

I run the v2 firmare release march 1st, if they possibly did anything more to that than release notes says.

But using say Reverb Level to zero on A and something else on B does move over to lower zone too. Voices are not handled correctly there, as far as I can see. But this indirect modulation through matrix works.

If any difference on Reverb Level, in this case, on A and B it will do morph on that too - which will leak into lower zone. So vital that PB does not encounter a difference in that on A and B side.

I will have a look at if any other good source to use than mod env.

[jhlz]

Hey larioso,

thank you for your efforts very appreciated!
I did it exactly as you described with my preset and with an init one and I got the idea.
If I only play the lower notes it works and it is dry but a soon as I play both split zones together, A (lower notes) also sends to the reverb of B as long as I hold a higher note. If I stop playing higher notes A is dry again with the next note I play there.
So unfortunately still reverb on A.
I wanted to have a sequence running on the lower split zone without reverb and using the higher split zone to play a melody along with it and using reverb but
as soon as you start playing highr notes the sequence also gets reverb.

My FW: 2.0.0.747

[larioso]

I'm not getting that, the lower zone remain dry, playing simultaneously on upper zone with reverb.
If using the default Hall, it's pretty obvious if there.
And for this test also max reverb make it even more noticable if on lower zone notes.
It stays in your face on lower without any tails or anything.

You hear reverb clearly on upper notes, so don't fool your ears playing on lower thinking they get reverb. I really listened for that.

Real world preset you would of course lower sustain level, or amount on modulation node to regulate how much on upper zone.

If possible make a test with PB Connect out of the picture.

I did 2-3 note chords on lower and similar on upper, was thinking if any voice stealing is happening. But could not see that either. But don't think I did 7-8 notes total if that might be different. And don't know if doing unison on either side either, did not test that. If lower zone gets a voice that previously was on upper, if that is possible.

I never changed any voice poly allocation, the default cycle or whatever it says.

And if doing notes with long releases, don't know how voice stealing might act then. More chance that is happening then, then on my init preset.

Either way report it to Arturia. I see no reason two different levels on reverb on either side should not morph properly without leaking into lower zone. I think it is just overlooked.

[DrJustice]

...
Whenever I turn the morph to B then the sound of A also gets the FX of B like whatever delay, reverb is set on B.
...

This is the correct operation, not a bug.

Think about it like this: the PB has two parameters sets for the A and B sounds, but both go through the one global FX unit. Therefore the last action that affects the FX unit will take effect for the FX unit and heard on both the A an B sounds (that action is morphing FX parameters in your case (the Type can't be morphed)).

Edit: Similarly, if you use a polyphonic modulation source, say, Mod Envelope to Reverb level, all voices will "compete" for control of the single Reverb Level. This is resolved by an algorithm that applies a sensible modulation. In the case of A with Mod Env set to all 0 and B set to S and R = max, you will not hear the reverb when morphed to A, and you will hear it with the morph at B. Since the lower part of a Split is always sound A, playing there morph at A will not give reverb. Then if you morph to B and only play the lower part there will still be no reverb since lower is always sound A. Now, when you hit a key in the upper part, the Mod Env of the B sound takes effect and you then get reverb on both the upper and lower parts. If you then morph back to A and play in the lower part, the Reverb is gone - that is, until the next time a B sound (upper part when morphed to B) is played again.

[larioso]

...
Whenever I turn the morph to B then the sound of A also gets the FX of B like whatever delay, reverb is set on B.
...

This is the correct operation, not a bug.

Thinks about it like this: the PB has two parameters sets for the A and B sounds, but both go through the one global FX unit. Therefore the last action that affects the FX unit will take effect for the FX unit and heard on both the A an B sounds (that action is morphing FX parameters in your case (the Type can't be morphed)).

I think he means the A sound of lower zone - which is not to be morphed at all.
If it was zero reverb level in A
- that means A will not have reverb  on lower zone ever.
- but follows morph position on upper zone
And my illustrations suggest
- this is overlooked and voices morphed end up in lower zone too, stuff not going through matrix
- but my workaround works since mod matrix handling works, and no morph is applied to lower zone voices

That is my conclusion, anyway. I think he found a bug or something overlooked.

[DrJustice]

@lariso : I just edited my posting above to explain what is happening. No bugs, and nothing overlooked there. Hopefully it's clear now

Edit: Short story - I reckon that having a 6 stereo input freely routeable multi (3) stereo FX unit is a bit heavy on complexity, parts and cost in an analogue synth like the PB...

[larioso]

@lariso : I just edited my posting above to explain what is happening. No bugs, and nothing overlooked there. Hopefully it's clear now

Edit: Short story - I reckon that having a 6 stereo input freely routeable multi (3) stereo FX unit is a bit heavy on complexity, parts and cost in an analogue synth like the PB...

I understand the concept of global fx. I have hardware rack effect units from daw
- I also understand the concept of multiple sends to fx with individual amounts

And all needed is two sends to each fx, one for lower and one for upper zone.
- each can have it's own level any moment since
- lower zone voices can be all dry at the same time as upper zone is wet to a degree

Especially since FX is digital domain I see no problem doing this.
- the entry to digital domain can have parameters as input representing lower and upper zone
- it's even possible to run separate instance of fx for lower zone, but probably a waste of resources

I will look further if there is a flaw in how I tested through mod env since OP still could not reproduce.
- but it that holds up working for mod env through matrix there is something overlooked in how morph is done in the invisible connections from A to B morphing it's way through.

This is nothing primary for me, not attempting to use this myself. Just interesting to know how things work and what limitation there are for lower zone if using it in future.

I'll see if I can upload an audio example. I tested as running a chord with extensive reverb on upper zone and hammering notes and chords in lower zone, and this would alter how upper zone sounded if just one handling.
- It does using morph, but not using mod env in matrix for some reason.
- if A side has zero send level it would silence all that upper zone on B had going, as I see it
- but I heard no difference in voices sounding in upper zone, and lower zone was all dry
- continued to play upper zone sounding wet while holding lower zone dry

That you get a clearly different result using morph with different send level to reverb on upper and lower zone and my workaround going through matrix alone with send zero on both A and B is interesting in itself.

[DrJustice]

...
And all needed is two sends to each fx, one for lower and one for upper zone.
- each can have it's own level any moment since
- lower zone voices can be all dry at the same time as upper zone is wet to a degree
...

Technically, inside the synth, there are no lower and upper zones, only the 6 voices that are dynamically allocated at all times.
To provide separate dry/wet level for each of the 6 stereo voices for each of the 3 FX blocks would require several tens of VCAs and corresponding CVs to control them, plus some mixers and buffers, to keep the FX routing in the analogue domain.

Especially since FX is digital domain I see no problem doing this.
- the entry to digital domain can have parameters as input representing lower and upper zone
- it's even possible to run separate instance of fx for lower zone, but probably a waste of resources

That would require 12 AD converters (6 stereo voices) with more of it ending up in the digital domain, and there would still have to be VCAs and CVs for the dry analogue path.

It can't be done without adding quite a bit of complexity and parts (which results in higher price, more heat etc.).

[larioso]

...
And all needed is two sends to each fx, one for lower and one for upper zone.
- each can have it's own level any moment since
- lower zone voices can be all dry at the same time as upper zone is wet to a degree
...

Technically, inside the synth, there are no lower and upper zones, only the 6 voices that are dynamically allocated at all times.
To provide separate dry/wet level for each of the 6 stereo voices for each of the 3 FX blocks would require several tens of VCAs and corresponding CVs to control them, plus some mixers and buffers, to keep the FX routing in the analogue domain.

Somehow PB differs in handling of voices allocated by lower zone in that they are all A side.

It's all digitally controlled.

Especially since FX is digital domain I see no problem doing this.
- the entry to digital domain can have parameters as input representing lower and upper zone
- it's even possible to run separate instance of fx for lower zone, but probably a waste of resources

That would require 12 AD converters (6 stereo voices) with more of it ending up in the digital domain, and there would still have to be VCAs and CVs for the dry analogue path.

It can't be done without adding quite a bit of complexity and parts (which results in higher price, more heat etc.).

That is what Sequential Take 5 do after the filter even, and doing amp and the rest digitally.
There is no analog path on Take 5.

All parameters stored in a preset are digital info, through a DA over to a control voltage.
This goes for all modern analog synths, that take presets.
It's not a biggy with that stuff.

If PB mix all voices down to a stereo mix and then enter digital domain audiowise still no problem.
So many things are governed by processor anyway, digital control envelopes, LFO's etc.
Processor handles that a key is pressed and allocated a voice etc.
All modulation every millisecond is controlled by processor.
Analog domain stuff is all done through DA converters over to control voltages etc.

To ship this into FX taking parameters is no problem.
You have all the time that lfo's and velocity is changing parameters and how things sound.
Nothing extra is needed hardware wise.

Will see what further tests show if I made a mistakes in how I confirmed my result getting lower zone having it's own send level, all dry in this case as a test. Later in the week maybe.