Vibrato question

[gez]

Although not my favourite effect, I've messed around with vibratos in the past using discrete MOSFETs wired up as phase-splitters.  My experience then was that the vibrato effect lessened with decreasing frequency, though this could have been due to the limited range of the circuit.

Flicking through my notes on phase shifters I came across a snippet which suggests that the reverse effect is true with op-amp circuits, i.e. vibrato decreases with increasing frequency.  Is this correct?  If so is there any way I can wire up a op-amp phase shifter so that the vibrato effect is more prominent on the higher strings?

Although I already have a working discreet version, a op-amp would be the easier option in the circuit I'm working on, hence the question.

[Mark Hammer]

In some respects, your question can be addressed by comparing the vibrato sound of a Uni-vibe, which spreads the phase-shift a little more evenly across the spectrum, with the vibrato sound of a conventional phase shifter, which focuses the phase-shift more.

As well, although I haven't thought it out very far, I'm wondering if some sort of bandwidth restriction in the dry path of a phase shifter might not assist your objective.

Personally, I think one of the best places to start is Larry Spence's modezero.com site.  He has samples of every manner of modulated devices.  Try out some of the vibrato samples there, and if you have schematics of the relevent pedals perhaps you can make some supportable inferences.

Two people worth eliciting comment from are Mike Irwin and JC Maillet, both of whom have a penchant for rigourously applying math to this and an abiding interest in the mysteries of allpass stages.  JC had a note this summer regarding the relationship between pitch-shift in vibrato pedals and number of allpass stages.  If I recall correctly, the notion is that more stages = more "wobble" in allpass-based vibrato circuits.  What this suggests is that you might get the desired frequency-dependent wobble via some sort of crude band-splitting and parallel allpass networks (driven by the same LFO).  For example, 2 stages for low-end in parallel with 4 stages for mids and higher.

I should emphasize that I have absolutely NO opinion about the matter, and am every bit as curious as you are.

[gez]

Now how did I know I was going to get an answer from you?  

If I recall correctly, the notion is that more stages = more "wobble" in allpass-based vibrato circuits.  What this suggests is that you might get the desired frequency-dependent wobble via some sort of crude band-splitting and parallel allpass networks (driven by the same LFO).  For example, 2 stages for low-end in parallel with 4 stages for mids and higher.

That’s a really good idea, only I was hoping for a simpler solution, preferably something that only requires a single op-amp stage.  

The discreet circuit I came up with in the past only uses one stage and is very subtle.  Whether it’s the nature of these circuits or by fluke,  frequency modulation seems to be less prominent on the lower strings, which is what I require for the first stage of a circuit I’m working on at the moment.  Although said circuit fits the bill sound wise, it requires a buffer and uses a LDR, and even though I did the whole lot with a 4007 (using one of the FETs to modulate a LED), it takes up more space than a op-amp, uses more components and I’d prefer not to use the LDR.

Before I start major upheavals to accommodate some new circuitry on the breadboard, I just wanted to know whether I’m going to get a similar phenomenon with a one stage op-amp design, i.e. frequency modulation more prominent on upper strings.  Although it's not that clear, my notes suggest the opposite effect.  Anyone?

[gez]

Now how did I know I was going to get an answer from you?  

If I recall correctly, the notion is that more stages = more "wobble" in allpass-based vibrato circuits.  What this suggests is that you might get the desired frequency-dependent wobble via some sort of crude band-splitting and parallel allpass networks (driven by the same LFO).  For example, 2 stages for low-end in parallel with 4 stages for mids and higher.

That’s a really good idea, only I was hoping for a simpler solution, preferably something that only requires a single op-amp stage.  

The discreet circuit I came up with in the past only uses one stage and is very subtle.  Whether it’s the nature of these circuits or by fluke,  frequency modulation seems to be less prominent on the lower strings, which is what I require for the first stage of a circuit I’m working on at the moment.  Although said circuit fits the bill sound wise, it requires a buffer and uses a LDR, and even though I did the whole lot with a 4007 (using one of the FETs to modulate a LED), it takes up more space than a op-amp, uses more components and I’d prefer not to use the LDR.

Before I start major upheavals to accommodate some new circuitry on the breadboard, I just wanted to know whether I’m going to get a similar phenomenon with a one stage op-amp design, i.e. frequency modulation more prominent on upper strings.  Although it's not that clear, my notes suggest the opposite effect.  Anyone?

[gez]

I need some ginko biloba!  I copied this snippet from a post just the other day (puretube?) and promptly forgot about it.

"The number of stages (phase shift vibrato) affect how far the lower notes are bent from their original pitch.
(the higher the original frequency, the further the phase-shift in each stage - very simple stated...).
BBD-pitchbending is uniformous for all frequencies.... "

Answers my question, sorry for wasting time!