The Official Sequential/DSI Forum

Differences in filter tracking

LPF83

  • ***
  • 526
Re: Differences in filter tracking
« Reply #20 on: June 05, 2021, 12:44:02 PM »
Experimenting with mine, I have seen some things I cannot explain but overall have to conclude the filter is tracking properly.

In one test, using the built-in Cubase Tuner plug-in, when I started out with the settings (env amt=0, resonance=max, keyboard=full, cutoff=approx 3 (tuned to A by ear using the reference tone), I found it tracked fine but could be off by maybe plus or minus 7 or 8 cents in either direction, which to me is a musically useful (if not pleasant) amount of detune and would be typical for the oscillators a VCO synth...  Then with some minor knob tweaks, suddenly I found that even at the same settings, with a new init patch and repeating the adjustments, everything was off by a much higher range, plus or minus 40 cents which is very noticable.  Then I pressed the Tune button to auto calibrate, and then repeating the test, the detune amount was approx 2 or 3 cents in either direction (about as tight as it can get with an analog synth).
Switching between filter types 1/2 or 3 didn't seem to matter.

So I have to conclude for now there is no problem with my particular unit here -- but it does seem to suggest that when calibrated, the filter tuning should track reasonably closely, at least not a half-semitone off or anything. 
Prophet 10, Prophet 6, OB-6, Prophet 12m, Prophet Rev2-16, Moog SlimPhatty, Hydrasynth desktop, Korg Minilogue XDm, Roland SPD-30, Roland SPD-SX Special Edition, Roland KT-10, Maschine, Focusrite Scarlett 18i20 3rd Gen + Octopre, Strymon Pedals, Cubase Pro 11.

Re: Differences in filter tracking
« Reply #21 on: June 05, 2021, 04:06:39 PM »
I have a support ticket open but seem to being told itís normal
I don't see how they sit there with a straight face and say tracking so badly out of tune is normal. And if it is normal, then why doesn't rev1/2 filters do it ?

The end all of end all to this would be simple - Find someone with a vintage rev.3 and see if indeed at high registers the filter tracks out of tune. If it does, then hell, this is normal and oh well. Don't use rev3 filters ! lol Cause they suck.

If it is not the way a vintage rev3 tracks, (which I am betting this is the case) then we have proved to them that this is *not* normal.

Re: Differences in filter tracking
« Reply #22 on: June 05, 2021, 04:20:21 PM »
Experimenting with mine, I have seen some things I cannot explain but overall have to conclude the filter is tracking properly.

In one test, using the built-in Cubase Tuner plug-in, when I started out with the settings (env amt=0, resonance=max, keyboard=full, cutoff=approx 3 (tuned to A by ear using the reference tone), I found it tracked fine but could be off by maybe plus or minus 7 or 8 cents in either direction, which to me is a musically useful (if not pleasant) amount of detune and would be typical for the oscillators a VCO synth...  Then with some minor knob tweaks, suddenly I found that even at the same settings, with a new init patch and repeating the adjustments, everything was off by a much higher range, plus or minus 40 cents which is very noticable.  Then I pressed the Tune button to auto calibrate, and then repeating the test, the detune amount was approx 2 or 3 cents in either direction (about as tight as it can get with an analog synth).
Switching between filter types 1/2 or 3 didn't seem to matter.

So I have to conclude for now there is no problem with my particular unit here -- but it does seem to suggest that when calibrated, the filter tuning should track reasonably closely, at least not a half-semitone off or anything.

Thanks for reporting back and even taking the time to do all this. I agree 7to8 cents in either direction or less is fine. But you can clearly see in mine it was way further off. Also, just mix a osc with it and its clearly ugly. So..hmmm

I was wondering, maybe its not a tracking issue when in lower registers tone wise ? So just curious, did you tune your filter's "A" to the same pitch register wise as I had mine ? (in other words, high pitch.) - If you're cutoff was around 3, that suggests to me maybe you tuned your "A" a octave lower ?

As far as the calibration - I will try this too. But again, 2 things - #1, I tune regularly and have it pretty set up as far as the memory bank of tuning thing. But I did retune before when I first noticed this as of course I thought that was the issue too, and it did not help the rev.3 filter. And #2 - if it was just a tuning thing, apparently it seems both filters would be out of tracking tune. In which yes, a tuning would help.

It would be weird if only I and Quatschmacher had these issues. But I guess it is possible ? Which really sucks if so :-(
Then again, I will have to be honest, I barely use rev3 filters+ it takes certain settings to make this happen, of which settings I rarely would use normally. So its no disastrous for me, but it is a concern, nor should it be like this regardless.

LPF83

  • ***
  • 526
Re: Differences in filter tracking
« Reply #23 on: June 05, 2021, 05:39:26 PM »
I was wondering, maybe its not a tracking issue when in lower registers tone wise ? So just curious, did you tune your filter's "A" to the same pitch register wise as I had mine ? (in other words, high pitch.) - If you're cutoff was around 3, that suggests to me maybe you tuned your "A" a octave lower ?

It would be in the same octave if you took OSC1 up one octave from init patch to C3.  I guess 3 isn't really accurate, its between 3 and 4... I look at the controls from an angle rather than straight down..  10 oclock is a better description.

But as I verified for purposes of this post, I noticed it couldn't hit a solid A again, it was about 40 cents flat.   No idea why, I reinit the patch and then its back to only a couple of cents difference.   Seems like something is going on that I can't quite explain.  None of this is really an issue for me though, because the rare times I use the filter to depend on pitch, it is mostly only to add a certain organ-like quality to the timbre.

My understanding is that self-oscillating filters behave a bit differently than an oscillator anyway -- I don't fully understand the science but it is effectively tuned feedback and isn't necessarily guaranteed to produce a pure sine at all ranges for all filter types, its more of a serendipitous side effect (or that's the way I've always viewed it).



Prophet 10, Prophet 6, OB-6, Prophet 12m, Prophet Rev2-16, Moog SlimPhatty, Hydrasynth desktop, Korg Minilogue XDm, Roland SPD-30, Roland SPD-SX Special Edition, Roland KT-10, Maschine, Focusrite Scarlett 18i20 3rd Gen + Octopre, Strymon Pedals, Cubase Pro 11.

LPF83

  • ***
  • 526
Re: Differences in filter tracking
« Reply #24 on: June 05, 2021, 06:07:48 PM »
I just tried this experiment on the Prophet 6, and results were similar -- the filter could easily be 45 cents out of tune.
I think it's just the nature of how a self oscillating filter behaves, probably nothing wrong with your P10.

The OB-6 has a 2-pole filter and cannot do this at all.   The Prophet 12 can in 4-pole mode, but because the key tracking has such a wide range (its not just half or full), if you find A in one key, good luck getting the key up an octave to even be in the same key, forget about cents out of tune.

So basically I'd say what support said is correct.  And on the Rev4 you might actually be getting the closest thing available (at least in the sequential line) to having a self oscillating filter that can track across so many octaves.

Maybe someone with a vintage P5 could confirm the behavior there?
« Last Edit: June 05, 2021, 06:15:52 PM by LPF83 »
Prophet 10, Prophet 6, OB-6, Prophet 12m, Prophet Rev2-16, Moog SlimPhatty, Hydrasynth desktop, Korg Minilogue XDm, Roland SPD-30, Roland SPD-SX Special Edition, Roland KT-10, Maschine, Focusrite Scarlett 18i20 3rd Gen + Octopre, Strymon Pedals, Cubase Pro 11.

Re: Differences in filter tracking
« Reply #25 on: June 05, 2021, 08:31:52 PM »
So then the rev1/2 filters just somehow magically track better ? A older filter design performs better ? Oh wait -- yea that does make sense haha !  ;D :P

Thanks for trying it on the 6. I'd say its how it is then. I admit, the only P5/10 I ever owned before or extensively used was when I got my P5 rev.2 (original version) about 15 years ago. So I really had no basis to compare, other than my rev.2 does not do this. I've never owned a rev.3.

Yes I agree, finding someone with a rev3 to verify this would be great.

LPF83

  • ***
  • 526
Re: Differences in filter tracking
« Reply #26 on: June 06, 2021, 05:12:09 AM »
So then the rev1/2 filters just somehow magically track better ? A older filter design performs better ? Oh wait -- yea that does make sense haha !  ;D :P

Thanks for trying it on the 6. I'd say its how it is then. I admit, the only P5/10 I ever owned before or extensively used was when I got my P5 rev.2 (original version) about 15 years ago. So I really had no basis to compare, other than my rev.2 does not do this. I've never owned a rev.3.

Yes I agree, finding someone with a rev3 to verify this would be great.

I guess somewhere along the line confused between inconsistencies I found during my test (which focused on Rev3 filter type) and the actual topic of the thread (why Rev3 and 1/2 would behave differently).  I did not see much difference when I switched filter types.  I just came to the conclusion that because of the nature of what a self-oscillating filter is and how keyboard tracking is implemented on various synths, it cannot really be expected to behave exactly like an oscillator when it comes to keyboard tracking -- and my conclusion may be off.  In terms of why the difference between filter options on one synth versus another, I have no real insight there.  But you say filter tracks perfectly (on both your P5 rev2 and your P10)?
Prophet 10, Prophet 6, OB-6, Prophet 12m, Prophet Rev2-16, Moog SlimPhatty, Hydrasynth desktop, Korg Minilogue XDm, Roland SPD-30, Roland SPD-SX Special Edition, Roland KT-10, Maschine, Focusrite Scarlett 18i20 3rd Gen + Octopre, Strymon Pedals, Cubase Pro 11.

Re: Differences in filter tracking
« Reply #27 on: June 06, 2021, 04:59:01 PM »
So then the rev1/2 filters just somehow magically track better ? A older filter design performs better ? Oh wait -- yea that does make sense haha !  ;D :P

Thanks for trying it on the 6. I'd say its how it is then. I admit, the only P5/10 I ever owned before or extensively used was when I got my P5 rev.2 (original version) about 15 years ago. So I really had no basis to compare, other than my rev.2 does not do this. I've never owned a rev.3.

Yes I agree, finding someone with a rev3 to verify this would be great.

I guess somewhere along the line confused between inconsistencies I found during my test (which focused on Rev3 filter type) and the actual topic of the thread (why Rev3 and 1/2 would behave differently).  I did not see much difference when I switched filter types.  I just came to the conclusion that because of the nature of what a self-oscillating filter is and how keyboard tracking is implemented on various synths, it cannot really be expected to behave exactly like an oscillator when it comes to keyboard tracking -- and my conclusion may be off.  In terms of why the difference between filter options on one synth versus another, I have no real insight there.  But you say filter tracks perfectly (on both your P5 rev2 and your P10)?
Yes, vintage P5 rev.2 filter tracks perfectly under same test, and yes P10 .rev1/2 filter tracks perfectly as shows in my videos. (perfectly =within musical reason). But P10 rev.3 filter has the issues shown in video.

My video was showing differences between filters, as that is what I was doing, swapping 1/2 filter for 3 filter and not changing any settings. Proving, normal or not, that 1/2 filter tracks perfect and 3 filter is horribly off in upper high registers.

Re: Differences in filter tracking
« Reply #28 on: June 07, 2021, 12:25:45 AM »

[/quote]
Yes, vintage P5 rev.2 filter tracks perfectly under same test, and yes P10 .rev1/2 filter tracks perfectly as shows in my videos. (perfectly =within musical reason). But P10 rev.3 filter has the issues shown in video.

My video was showing differences between filters, as that is what I was doing, swapping 1/2 filter for 3 filter and not changing any settings. Proving, normal or not, that 1/2 filter tracks perfect and 3 filter is horribly off in upper high registers.
[/quote]



Conclusion :  Use the 1 / 2 filter if you want a sound with oscillating feedback in perfect pitch   ;)

Case closed ?


Cheers !
1976 MiniKORG700s // 1978 Prophet-5 rev.2 // 1981 KORG CX-3 // 1984 DX7 // 2020 Prophet-10 rev.4 // MoPho Box // 2 Creamware MiniMax // Creamware Pro-12 // 2 EMU-Proteus 2000 // EMU-Vintage Keys  // Casio VZ-10M // Roland VK-8M // Fatar SL 880 //

Re: Differences in filter tracking
« Reply #29 on: June 07, 2021, 01:50:20 AM »
Conclusion :  Use the 1 / 2 filter if you want a sound with oscillating feedback in perfect pitch   ;)

Case closed ?

Cheers !
Well no, and yes. Like I said, for me, yes that is what I will do. And frankly, it doesn't effect me much. But that's not what these forums are about. This is about reporting issues, whether they effect you or not. To have no compassion towards the fact someone else might be effected by it drastically is not a right attitude. We just want to report everything and get things working right to help Sequential too. No matter how small.

Check it out, that would be like me popping into your thread and saying: "Conclusion, don't use a expression pedal ! Case closed."

I would never do that. Even though me personally do not use expression pedals, so it would be very easy for me to look at it that way. It don't effect me, I will never need that or worry about it, so why should Sequential fix that ? Well to be the bigger man here, I will say they should fix it and everything possible that we report or ask, because then, it will be a super kick ass product ! It already is, yes, but everything could always be better. Especially when working directly with users. And even if something doesn't effect me much, it would be great to help everything get fixed/addressed.

I imagine someone might want to use filter rev3 for some characteristic is has, plus have it track in good pitch. Using filters 1/2 might ruin their vibe. And that would suck.