Voice to voice filter variation on high resonance

Voice to voice filter variation on high resonance
« on: July 15, 2018, 01:28:23 PM »
On my Prophet X I'm seeing reasonably large variation between the 16 analog filters when on high resonance:
- The point they each self oscillate varies - some starting at 230/255, others needing more like 240/255
- The frequency of self oscillation varies by a good fraction of a semitone - this makes it impossible to "play the filter" as an extra sinewave voice
Repeatedly stabbing a note (cycling through the voices) makes this obvious.

Is this considered normal? On (say) my Modal 008 I can play the filters up and down the keyboard with no obvious inter-voice variation.

dslsynth

  • ***
  • 1040
Re: Voice to voice filter variation on high resonance
« Reply #1 on: July 16, 2018, 10:43:53 AM »
Any official word from DSI on how the filters on Prophet X are calibrated and if they are calibrated at all?

Razmo once posted a potentially useful hint on another DSI instrument: let the instrument heat up for some time and then recalibrate the filters. Likewise let the instrument heat up before playing the filter. Consider giving that a try.
#!/bin/sh
cp -f $0 $HOME/.signature

Razmo

  • ***
  • 2168
  • I am shadow...
    • Kaleidoscopic Artworks
Re: Voice to voice filter variation on high resonance
« Reply #2 on: July 17, 2018, 12:36:09 AM »
Yep... I let my REV2 warm up for half an hour before calibrating the filters (and also using it)... that insures that the filter can self oscillate rather well, actually over 8 octaves which is pretty cool... if I calibrate when cold, half an hour later it will be horribly out of tune... if I calibrate after half on hour it'll be horribly out of tune for about half an hour after turning it on... I choose the last scenario ;)

But the REV2 use a different filter chip... the X has a new one, so I'm not aware of how those work regarding calibration and heating up... but try it, it may very well work.
If you need me, follow the shadows...

Re: Voice to voice filter variation on high resonance
« Reply #3 on: July 17, 2018, 04:53:10 PM »
My Prophet X was a half step sharp on all patches. Noticed after a day when I started combining it with my Jupiter 6 and Rhodes. I ran the filter calibration and everything was perfect after that. Manual says it is done before shipping and should be perfect and not need to be touched but for me that wasn't the case. I let my X warm up for a good half hour or so before I ran it.

Razmo

  • ***
  • 2168
  • I am shadow...
    • Kaleidoscopic Artworks
Re: Voice to voice filter variation on high resonance
« Reply #4 on: July 19, 2018, 03:21:22 AM »
My Prophet X was a half step sharp on all patches. Noticed after a day when I started combining it with my Jupiter 6 and Rhodes. I ran the filter calibration and everything was perfect after that. Manual says it is done before shipping and should be perfect and not need to be touched but for me that wasn't the case. I let my X warm up for a good half hour or so before I ran it.

Problem is, that none of DSI's synths can avoid those calibrations... I've had to do them on EVERY synth I've had from them (Mopho, Tetra, P08 Desktop, Desktop Evolver, Mono Evolver, Poly Evolver Rack, Tempest, Prophet 12 Module)... if you want to use the filters for playing in self oscillating mode, then you WILL need to calibrate your machine, and you may even have to do it often depending on your ambient temperature in your working environment.

The Curtis oscillators are usually stable, but the filters are rarely, and need this calibration... and you'd have to let the synth warm up before doing it, otherwise you'll end up with filters out of tune when you play them...

If users do not use the filter for playing in self oscillating mode, then DSI's claim about not needing to do calibration may be true... in that case you probably will not hear them being a bit off... it only become evident when the filters are extremely resonant to the point where it seriously start to enhance certain harmonics at high settings of resonance.

Unfortunately, many cool sound features are only possible on the REV2 when using self resonating filters, so to me it's rather important to have a good working calibration... half my own presets rely on this... any sound that use the Audio Mod parameter in conjunction with a highly resonant or self oscillating filter, to create bell like FM tones will require the filters be in tune, or they would sound like a sick orchestra playing.
« Last Edit: July 19, 2018, 03:25:22 AM by Razmo »
If you need me, follow the shadows...

Re: Voice to voice filter variation on high resonance
« Reply #5 on: July 19, 2018, 10:01:55 AM »
My Prophet X was a half step sharp on all patches. Noticed after a day when I started combining it with my Jupiter 6 and Rhodes. I ran the filter calibration and everything was perfect after that. Manual says it is done before shipping and should be perfect and not need to be touched but for me that wasn't the case. I let my X warm up for a good half hour or so before I ran it.

Problem is, that none of DSI's synths can avoid those calibrations... I've had to do them on EVERY synth I've had from them (Mopho, Tetra, P08 Desktop, Desktop Evolver, Mono Evolver, Poly Evolver Rack, Tempest, Prophet 12 Module)... if you want to use the filters for playing in self oscillating mode, then you WILL need to calibrate your machine, and you may even have to do it often depending on your ambient temperature in your working environment.

The Curtis oscillators are usually stable, but the filters are rarely, and need this calibration... and you'd have to let the synth warm up before doing it, otherwise you'll end up with filters out of tune when you play them...

If users do not use the filter for playing in self oscillating mode, then DSI's claim about not needing to do calibration may be true... in that case you probably will not hear them being a bit off... it only become evident when the filters are extremely resonant to the point where it seriously start to enhance certain harmonics at high settings of resonance.

Unfortunately, many cool sound features are only possible on the REV2 when using self resonating filters, so to me it's rather important to have a good working calibration... half my own presets rely on this... any sound that use the Audio Mod parameter in conjunction with a highly resonant or self oscillating filter, to create bell like FM tones will require the filters be in tune, or they would sound like a sick orchestra playing.

Noticed when I was just testing factory presets. So it wasn't even the self oscillating that was the issue. Just out of tune until I ran the calibration

Re: Voice to voice filter variation on high resonance
« Reply #6 on: July 19, 2018, 12:46:01 PM »
Noticed when I was just testing factory presets. So it wasn't even the self oscillating that was the issue. Just out of tune until I ran the calibration

Which calibration?  I guess it’s possible the pitch wheel was miscalibrated, because otherwise I’m not sure that makes sense.  None of the oscillators are analog, after all.

Re: Voice to voice filter variation on high resonance
« Reply #7 on: July 19, 2018, 01:14:42 PM »
Noticed when I was just testing factory presets. So it wasn't even the self oscillating that was the issue. Just out of tune until I ran the calibration

Which calibration?  I guess it’s possible the pitch wheel was miscalibrated, because otherwise I’m not sure that makes sense.  None of the oscillators are analog, after all.
I know. It made zero sense to me as well. I just know it was 1/2 step off from my Jupiter and Rhodes and once I ran the filter calibration it was rock solid in tune. Can't explain it, but it happened and that worked. Did not run any pitch wheel calibration and that seems fine. Regardless, working as intended now and maybe my experience will be helpful to someone else if they experience the issue.