Filter causes oscillator's phase to change, multiple osc outputs

[LBH]

It's great you bring up things.

I'm just saying it's a sync behavior, similar what you can read about in the Modulars manual page 93.

[Tausendberg]

but it's a sync behavior that shouldn't be happening, I mean, there's nothing about analog technology that would cause a signal when passed through a low pass filter to be delayed by some amount.

[LBH]

As you can see in my posts, then i wonder about why it sync in some routings, but not in others, just like i have explained.
I can't say for sure, if it's a exspected behavior or not. But i wonder about the difference.

I can add - If i recall correct, then you for instance can take two outputs from the same waveform, and pass each of them thru a filter, then combine them in the mixer, and you don't get the sync. I guess that could be similar to the Reaktor example you mention.
EDIT: How come you can do this, but you can't do ithe same, when only sending one output thru a filter, without getting this sync behavior? To me that make no sense at this point.
Perhaps it has to do with using direct waveform outputs  not using a filter.
But then why is'nt there the same sync behavior when combining 2 direct outputs from the same OSC waveform?
Somehow i think the filter among other things alter the pitch, like i suspect it do  in MINI V and i guess in other V5 applications. And perhaps when using the same OSC waveform output twice that produce a sync behavior, when one output is direct and another use a filter. But i still wonder. And as i wrote, then i don't find the normal soft/ hard sync behavior to good, and in addition to that i also find there is a problem with having a sync behavior added to the PW for the SAW that even have a different octave at minimum and maximum. That make it impossible to use a normal SAW and another waveform from the same OSC together using the same octave and/ or without having some kind of synced SAW. EDIT END.

[Tausendberg]

I mean, really, I think every module that the signal goes through delays the signal ever so slightly and that's what causes the out of sync effect.

If I take an oscillator output and split it with one osc going straight to the mixer and then another signal going through two more filters then I can see that it's been delayed even a little bit more than it would be if it went through one filter.

Let me demonstrate...

[Tausendberg]

Ok, so, as an extreme example, I split up the signal at three different points and then mix all of the signals back into the linked mixer channels and you can clearly see how there are now four oscillators out of sync relative to another because each time a signal passes through a filter, it gets delayed just a little bit.

From what I've been able to research, this essentially shouldn't be happening, signals, in an analog synthesizer, that would be passed through multiples, wouldn't be delayed like this.

[LBH]

Beside the below, then please refer to my other posts. I can only repeat my self, including what i wonder about regarding the behavior.
I think perhaps a issue can have to do with how pitch is handled somehow throughout some of the applications. This i have reported before in the forums,that  i think needs Arturias full attention.

1. The results you called out of phase is a sync behavior, and not about something being out of sync or delayed as such, except perhaps from what for pitch differences may introduce. The sync we talk about is creating complex waveforms, but how this is achieved i don't have the technical knowleadge to tell.
Have you seen Modular V3s manual page 93, that i have refered to some times? Otherwise please do so.
(BTW - If you check Matrix 12 sync, then it's a bit different because it don't get thin, and don't change waveform before you change pitch of VCO 2. Change pitch upwards, and you'll see what you call delay, but is a what's called a complex waveform.)

A sync behavior is not a delay bug.
But i have never said, that there is'nt something odd or wrong going on. Actually quite opposite like you can see in my wondering in my posts.
I just wanted to find out, if what you reported actually was about  some sync behavior going on, or was about phase like you was talking about. It is about a sync behavior and not about phase - (except - as i said - their perhaps are unwanted phase issues involved in pitch and sync behaviors in some V-colection synths causing thin sound). Then it's easier for Arturia to understand, what this is about.

2. This is just to comment your anolog comparison, and nothing else.
Analog synth signals are allmost intant i think. It's fast, but it's not at lightspeed, so there must be some kind of delay. We talk about digital computer stuff processed in series. There will be some kind of delay.
If you for instance in Studio One insert a effect on a track, then below the performance meter you'll see how much the signal is delayed. Not all signal delay is audioble - You don't hear it.
(Perhaps some day signal delays will be gone (or nearly gone), when we get quante computers where processing is'nt made in series.)

[LBH]

I can add this too:
I do believe there is some kind of unwanted behavior - some kind of bug.
I think it might be about the same as for instance about this (Even if i posted this in MINI V3 forum):

TOPIC: STILL A NOT WANTED BEHAVIOR WHEN OSC'S ARE TUNED EQUAL OR SYNC IS USED.
http://forum.arturia.com/index.php?topic=87770.0

TOPIC: PITCH DRIFT AND OSC DETUNING ISSUE.
http://forum.arturia.com/index.php?topic=89175.0

Something unwanted is going on, that seems to be pitch related issues, and that might cause other issues like phase issues.
To me it's a bug of some kind, or more bugs or a vital construction issue, like i can suspect it is. As said, then i think Arturia should pay full attention to this, as it has to do with some of the most important things in the applications, namely the sound and the sound stability.

In the case of Modular and the experiment you show one can say, that in a analog modular you don't take more than one output from a single output, and that may not be intended to do in modular V, even if it's possible in some routings, and that difference also make me wonder, as i have written in my posts.
But you also get similar unexspected results. when doing things like you should be able to do.

I have explained in my posts. (Also in  other threads).

[LBH]

Just to report, as i have mentioned Matrix using Sync in this thread.

Sometimes when just tuning Matrix OSCs equal - (not using sync), then one get a sync behavior. It's not allways. This confirm it's something Arturia should give full attention.

[Tausendberg]

Yeah, I really think they should.

[LBH]

I have been thinking a bit more about all this.
I think there actually are some kind of phase shift involved in all this. The sound it self is not out of phase, but the waveforms start at different points - and create a complex waveform that i think have a phase relationshift.

So something introduce this kind of phaseshifting.
I guess something like pitch/ pitchdrift and/ or some kind of delay like you talk about can do so. And as you say in this topics header, then it looks like the filter introduce this behavior somehow. And because of some odd behaviors i have exsperienced i think perhaps envelopes for some reason introduce phaseshiting relationship as well.
If this is the case, then i don't know if this actually can be seen as a bug.

In your experiment 2 outputs from the same waveform directly combined in the mixer do not introduce any phaseshift relationship.
However - Then this none phaseshift relationship behavior should also be seen when hardsyncing two OSC's to eachother, but that do introduce the phaseshift relationship even when the OSC's is tuned equal. This should then be wrong. And then this is the bug.
When tuning 2 OSC's equal then a pitch or phase drift may introduce the phaseshift relationship.

Having an option to not have the osc's be "freerunning" might solve the issues - at least what the OSC's are concerned. Perhaps one need the same "Freerun" option for filters and envelopes, in the cases where one use multiple outputs from a single oscilator.

Looks like Matrix is closest to be correct - if not fully correct taking the analog simulation in account - as Matrix do have the phaseshift relationship when having two osc's tunned equal, but not when the OSC's is both using Sync and is tuned equal.
But still an option to not have the OSC's to "Freerun" would be nice to have.
So looking at how Matrix do it might also fix things, if my thoughts in this post is correct.

There do seem some odd behaviors concerning pitch and other things as you can see in the threads i have had linked too in the thread. But it look like Arturia should be able to fix it all.

[Tausendberg]

https://www.arturia.com/mini-v/behindthemini-v?&utm_source=elasticemail&utm_medium=email&utm_campaign=mini-v_com_2016&utm_term=nl-halloween&utm_content=stefano-filter_picture2

I just looked at this article and in there they are talking about employing a zero delay filter...

Do you think that might have something to do with it?

[Tausendberg]

I did some experimenting tonight with oscillators syncing and I agree with what you say, the simplest solution would be for Arturia to just implement the Matrix's sync algorithm when on the Modular V the user chooses 'hard sync' and maybe to make the programming simpler, that could also turn off free running.

One thing I noticed today was that I took the triangle output of OSC 1 and connected it to Mixer 1 and took the square wave output, also of OSC 1, and connected it to Mixer 2, and it worked perfectly, the waves were completely in sync, like what I would expect from another OSC 'hard syncing' to another Osc even when it's perfectly in tune.

Right now when I use hard sync with two oscillators perfectly in tune with one another, what appears to happen is that at first it behaves like soft sync for a short amount of time and then it "snaps" into sync. What I think it should be doing with hard sync when the oscillators are in tune to one another is that they "snap" into sync as soon as a key is pressed or perhaps are already in sync before the key is pressed.

I know it's a crazy idea but some of these problems could be solved just by giving the user an option to reset the oscillators every time the key is pressed (I don't know if something like that ever existed in the analog world though).

[Tausendberg]

To put it another way, I can tell that Modular V's "hard sync" doesn't work like it should because I've watched oscillator sync being demonstrated on analog instruments on youtube and it's instantaneous and perfect, it doesn't 'drift and snap' like Modular V's "hard sync".

[LBH]

https://www.arturia.com/mini-v/behindthemini-v?&utm_source=elasticemail&utm_medium=email&utm_campaign=mini-v_com_2016&utm_term=nl-halloween&utm_content=stefano-filter_picture2

I just looked at this article and in there they are talking about employing a zero delay filter...

Do you think that might have something to do with it?

I don't have the knowleadge to say yes or no.
I think in your exsperiment that show that the filter somehow introduce the phaseshift relationship, and regarding the envelope behavior i have noticed, that the fact it's not zero delay might be the - or a cause for the phaseshift relationship behavior.
I guess if a filter or a envelope changes the pitch a little, then that too can be a cause.

I really don't have enough knowleadge about this, and about how all those things works, to say for sure.
But something could and should be done better. Perhaps the things we have suggested.

I know it's a crazy idea but some of these problems could be solved just by giving the user an option to reset the oscillators every time the key is pressed (I don't know if something like that ever existed in the analog world though).

I don't think it's a crazy idea.
I don't know either regarding the analog world. But sync is possible in the analog world.
To me no matter what,  then as long the original behavior is availble as good as possible, then i would say options can be added, especially if that option repair things or ass functionality.

As said - then something could and should be done better. Perhaps the things we have suggested could help somehow.
Also observations can help finding solutions.

[Tausendberg]

One more thought, something that just occurred to me, the oscillators should be synchronized, when set to hard sync, as soon as the connection is made, even before a key is pressed, because in analog the oscillators are constantly running even when the VCA is closed and so they're being forced into sync even when the VCA is closed.

That means for Arturia to depict hard sync authentically, when a key is pressed there shouldn't be any perceptible time that the oscillators aren't synchronized.