Theoretical Vibrato Question

[ZiggyZipgun]

Does anyone know of or could give suggestions for a vibrato circuit that doesn't raise the pitch, but instead just modulates between the note played and however much below it the Depth control allows?  I never use vibrato or chorus, but I've been kicking around ideas for an envelope vibrato, and I think it was in Principles of Orchestration by Rimsky-Korsakov that I read that sharp notes are more harsh on the ears than flat ones, so [in that context, probably brass or wind instrument] vibrato should never go above the pitch being played.  String vibrato is much more subtle, only varying a slight amount, and not really the kind of thing you'd be using a normal vibrato pedal for, nor would you want it on all the time.  Just kicking around ideas.  Any thoughts?

[ZiggyZipgun]

Did I lose everyone with the words "Rimsky-Korsakov"? 

Anyone?

[R.G.]

No, not lost. Some of my best friends are Rimsky-Korsakovs. 

I'll have to think about that one. I have a couple of ideas, but it may well be complicated to do. The problem is that you lower the pitch by doing a continuously-increasing delay. At some point, if you ever quit increasing the delay, then the signal reverts back to normal pitch, but delayed more in time. You then can't increase the lowered pitch back up to zero delay, because that signal's already gone. If you decrease the delay the signal then is increased in frequency.above what it would be otherwise. There's no source of slowed-down signal to increase back to normal. DSP can do what you want, because it does not rely on the rate of change of delay to do vibrato; or at least, there are also other ways to decrease pitch that are not purely delay based.

To add to Rimsky-Korsakov's observations on higher pitch sounding harsher, there is a bit of history. We use A=440Hz as a pitch standard these days. It was not always so. European instruments and orchestras used to tune to their own pitch standards. At one time, A was down at about 420 in most places, but the progression was for most places to drift higher in frequency over time. Some of them got well over 440. Players from other areas had a hard time tuning to the local pitch in some cases. That's why the 440Hz standard was adopted, to synch everyone up.

However, it is common for soloists on non-fixed-pitch instruments, like violin and its ilk to play a tiny bit sharp. This apparently has the added attraction of making the instrument stand out from the backing accompaniment. So harsh, or just attention getting, being a bit sharp can be musically useful.

[ZiggyZipgun]

Definitely think about other ways to do this - I am trying to stay away from DSP, but I'm now second-guessing the earlier assumption; maybe the pitch heard is midway between the sharp and flat fluctuations.  It makes sense, though the rate can play a role in that as a slower modulations might make it harder for the ear to "blend" the two pitches.  I'm aiming for a dynamic vibrato styled after classical vocalists; at lower, regular volumes the effect is very subtle or non-existent, but in louder passages, it's always there.  Naturally most guitarists would just use finger vibrato, but you can only go above the pitch and back and consistency is tricky if you're already bending up to the note.  Could be interesting...though maybe a DSP prototype would determine if it's worth working out properly.  Let me know what you come up with!

[Uma Floresta]

j/k :p

[David]

j/k :p

ROTFLMBO!      

[ZiggyZipgun]

Maybe we can find a small electric motor with enough torque to move the Bigsby arm automatically...something like a motorized B-bender! 

Not quite what I had in mind.

[Mark Hammer]

Too many acronyms in those last two posts for me

Just to expand on RG's note (and particularly since my first impulse was "Aw, that's easy.....hmm, maybe not so much".), if one were to use a BBD chip to keep changing the sample rate, and consequently the pitch of the signal, slowing the pitch down would mean you'd have to somehow eventually catch up.  Or to put it a different way, with time-based vibrato, each little piece of time that you use up to drop the pitch, has to be complemented by a period fo time where you clock the signal faster so that you "make up for lost time".  That is why, of course, a BBD/time-based vibrato will normally HAVE to fluctuate between sharp and flat.  The sharp periods are where you catch up to where you ought to be after dragging your ass for a few tenths of a second.

So, as RG wisely notes, "How could you conceivably keep 'losing time' without ever having to make it up?".  That's a real conundrum, and probably the very reason why any analog solution to the challenge would not likely fit in a 1590BB or DD.

[David]

Too many acronyms in those last two posts for me

Just to expand on RG's note (and particularly since my first impulse was "Aw, that's easy.....hmm, maybe not so much".), if one were to use a BBD chip to keep changing the sample rate, and consequently the pitch of the signal, slowing the pitch down would mean you'd have to somehow eventually catch up.  Or to put it a different way, with time-based vibrato, each little piece of time that you use up to drop the pitch, has to be complemented by a period fo time where you clock the signal faster so that you "make up for lost time".  That is why, of course, a BBD/time-based vibrato will normally HAVE to fluctuate between sharp and flat.  The sharp periods are where you catch up to where you ought to be after dragging your ass for a few tenths of a second.

So, as RG wisely notes, "How could you conceivably keep 'losing time' without ever having to make it up?".  That's a real conundrum, and probably the very reason why any analog solution to the challenge would not likely fit in a 1590BB or DD.

ROTFLMBO = Roll on the Floor, Laughing My Bottom Off.

As to the "maybe not so much", let's take a step back here.  This subject has actually come up before, albeit in vastly different guise.  Consider.  We're talking about vibrato, right?  As in pitch change?  I can think of two ways to come up with an "apparent" pitch change:
1)  Create a Doppler effect with something like a Leslie speaker.
2)  Do actual pitch change in a fashion analogous to how Magnatone did it on their amps.  This could be tough.  I think at least one of the components is made out of unobtainium.

[RickL]

I'm wondering if a vibrato that always bent down from the original pitch wouldn't just sound like an up/down vibrato around a slightly lower original pitch.

Imagine a bending down to 438 Hz from 440, then back up to 440, using a sine wave. Wouldn't that sound exactly like bending up to 440 from 439, down to 438 and so on...?

My guess is it would sound like you were just adding vibrato to a slightly flat original note.

[ZiggyZipgun]

Vocal vibrato usually doesn't go above the pitch at all, but it also doesn't go very low, so 440 to 438Hz is still very subtle.  Also, to quote a TC Electronics manual:
TC research has shown that when the speed of a modulation effect is increased, its depth should be decreased accordingly to achieve the same amount of perceived modulation. When the Golden Ratio parameter is set to On, this value is automatically calculated.  Golden Ratio is a feature inherited from the classic TC 2290 processor.

[Mark Hammer]

I think what we're really looking at here, assuming such a thing can be made, is a "Hank Marvin emulator"!  That is, a nice slight shimmy that mimics loosening and retightening the strings, but without going above pitch.  Of course, if you're playing alone, it doesn't really matter.  The trick is when yo're playing with someone else, and you have to create the illusion of being in tune with them.

[R.G.]

Do actual pitch change in a fashion analogous to how Magnatone did it on their amps.  This could be tough.  I think at least one of the components is made out of unobtainium.

Actually, although there's a lot of smoke and mist wafted over Magnatone amps, the reality is not only knowable, but known. The Magnatone vibrato was a real vibrato, and it was done the same way we'd do a vibrato made from a phaser. You make a phase shifter and then just don't mix the dry signal into it. Analog phase shift stages cause a real time delay; the awkwardness is that the delay is not constant with frequency, getting bigger as frequency rises. But it does work to make a noticeable vibrato.

Magnatone amps used two vacuum tube analogs to the univibe style phase stages. They had to use variable phase resistors which would withstand the high voltage environment of the tube preamp, and they used voltage variable resistors. All they had were Metal Oxide Varistors (MOVs). Their MOVs were not the sharp-threshold MOVs used for AC power transient protection that we have now, but rather had about a 10:1 change in resistance over about a 70V range. These parts are the "unobtainium" part of the story.

About 12-13 years ago, I have built Magnatone style phase stages with triodes and LED/LDR pairs. They work OK. Very complicated and expensive to do compared to a solid state phaser followed by a tube preamp, but they work.

[ZiggyZipgun]

I'm definitely aborting any attempt at making a "down only" vibrato, but I am going to put together an envelope vibrato.  Not something that you're going to use on every song, but in the Floydish vein of music that I normally function, it could be useful.  The theories I'd read about sharp pitches in vibrato definitely doesn't seem to apply to all instruments, and certainly not to guitar, since I use finger vibrato all the time.