Search

  • 0

Getting the data to emulate guitar tremolos

D., 15 august 2020
Getting the data to emulate guitar tremolos

First: yes, this is about vibrato, not tremolo, technically. But Leo Fender used the terms backwards, and that's kind of stuck, so we're going with it. We could also call it a whammy bar, but that term kinda became popular around the time the popularity of Jazzmasters was at rock bottom, so it would feel weirdly anachronistic. So, "tremolo" it is, for the purposes of this post.

When you modulate the pitch on a synth or a sampled instrument, whether using the pitch bend wheel or an LFO, it usually doesn't matter what MIDI note you're playing - any note would get its pitch modulated by the same number of cents as any other note. That's generally reasonable for violin-family instruments, too, and for monophonic wind instruments. It doesn't work too well for guitars - if you're playing a melody, it's fine, but if you play a chord, it just sounds too clear and in-tune and doesn't get all weird like they do in real life. It just doesn't sound anything like the classic Chris Isaak wobbliness you'd expect.

Now, we're going to totally ignore bending strings with the fretting hand here, and just deal with tremolo bridges and tailpieces. There are lots of variations - the Bigsby, the original Strat design, the Floyd Rose, the Steinberger Trans-Trem, the Gibson Maestro Vibrola etc. Some of them bend the pitch only down, some can do both up and down. A few only go up. Their maximum range varies a lot, too. We happened to be sampling a Jazzmaster type for a big sample library, so that's what we used for this video - at least we know what this one does!

So, that's what it does - the pitch of some of the strings changes by just about three times the amount of the pitch of the other strings, and that's why using it while playing more than one note makes things sound all detuned and weird. The thin strings get modulated less than the thick ones, and the wound strings also get modulated only a little more than the plain ones, though they are much thicker. It seems like the main cause of the variation might be the string core diameter... perhaps. At least that's a plausible hypothesis for now.

We'll keep gathering more data, and maybe someday we'll be able to make a generalized model that takes into account the tremolo type, the string type, string gauge, and length of string beyond the headstock and bridge (if those have any influence), but for now we have this...

typeJazzmasterBigsbyFormantaStrat

Mastery

bass Kahler
high E3834263297
B7166637194
G10375134749852
D776861489544
A8059686098

79

low E

100100100100100

100

Other than the Mastery tailpiece which seems like it pretty much behaves like the pitch wheel on a synth, and which we'd very much like to see replicated, the same general pattern shows up, though there's a good deal of variation in the G and D strings. Even the rare bass Kahler trem behaves not too terribly differently from guitar trems, though its bend depth on the D string is quite shallow.

The Formanta is a Soviet guitar with a Bigsby-ish trem which also behaved more or less the same as Western trems, but its high E string broke and got temporarily replaced with a 0.017 gauge string (an old G). Its pitch didn't move as much as the other high Es, so it seems that while larger gauge strings change pitch more, pitch will also change less when the string has less tension on it.

The max depth with the bar pushed all the way down is only 80 cents on the high E with the Strat-style trem (which was a Danelectro Hearsay from the 90s), and a little over 100 cents on the Jazzmaster-style. We don't have max depth data for the rest, though the bass Kahler was able to go down 100 cents on the G string without bottoming out.

We'll keep gathering more data and adding to this - anybody got a Gibson Vibrola, a Kahler roller bridge, a Steinberger Trans-Term, a Bass VI with a trem or some weird rarity which you feel like testing out for us, please contact samples@karoryfer.com.

A couple more notes: upward bends (where possible) seem to work the same way as downward bends, only with a smaller max value. Any note on the same string gets bent by the same amount, consistently, with some tiny amount of variation.

With a floating bridge or tailpiece, bending one string by a half-step makes the rest of the strings go flat by about 5 cents.

There's also the matter of how out of tune things will be after they come back; mercifully, we decided we don't need that much realism in virtual instruments, and left that one alone.

If you're an instrument developer and would like to put this kind of thing to use, we've also explained vibrato LFOs in a tutorial on the SFZformat site - that also covers how to make vibrato asymmetrical, which can be useful with Strat-types.

Cart

Your shopping cart is empty

Buy your first product