'Theory' question: Can changer string travel be calculated?

[Jim Pitman]

Having been an automation engineer for many years, the synthesized approach seems to be the most feasible. Unfortunately, synthesizing the subtleties like attack, palm effects, damping effects, etc is difficult if not impossible currently.
How does one configure the character, of a given string barred at a given fret and plucked at a given distance from the changer, with a given force intensity, character being made up of attack, envelope shape, overtones, pitch, LF pitch change, sustain, etc. What characteristics do you want to have control over and which ones do you want to have automatic?
A synthesized PSG becomes another instrument, (one without as much soul currently) in my opinion.
An electro-mechanical actuation system is more desirable but where's the market?
I could see manufacturers adding such a system to one guitar so they can demo it at a convention. "Dial in your copedent an try out my guitar" - that's about it.

[Jim Pitman]

I believe the door is still wide open to make improvements to the mechanically actuated PSG.
Can you achieve pitch accuracy/good tone/sustain/low overall weight/portability with mechanics - yes.
Has it been done? - not entirely.

[RD Bennett]

Some exciting new stuff here - thanks particularly to Messrs. Cowell and Packard for your very interesting input. I'll have to look into the info here once I've had a few brain-clearing rounds of joe.

On synthesizing the steel-- good chance it would be an entirely different instrument, but it could still be very good at many of the same things that make us love the steel. I see no reason why it would have to sound "soulless" or sparse in its expressive possibilities any more than I think an analog synth sounds that way. (But then, I'm a lifelong synth nerd too.)

I think it *is* in fact possible to create a synthesis engine capable of responding to nuances of right-hand performance. The key is to use trenchant sensing in conjunction with some kind of physical modeling-based engine, which of course has been beaten to death in guitar amplification and effects but has gone curiously ignored in MIDIphile circles.

There may be patent issues as Yamaha and Stanford seem to jointly own a lot of the existing IP surrounding PM-based synthesis. That, or the increasingly homogenized mainstream synth-market demands ("no, just give me YET ANOTHER version of the Minimoog and/or M1!!!") have kept more sophisticated, unique-sounding physical modeling synthesis systems from really taking off in hardware.

But if you look at some of the stuff that was being researched ten years ago in the physical-modeling realm, there's a lot of promise there. There are well-known models for string behavior like Karplus-Strong that seem to have eventually generated some really interesting and somewhat successful results in emulating, say, the varying bowing tactics on a violin. I'd argue that bowed stringed instruments are much more complex to emulate than plucked ones.

[ed packard]

RE calculations...see below link, and check out part IV.

http://www.home.agilent.com/agilent/edi ... &id=866712

RE synthesis...there are alternatives to storing and recalling sampled or synthesized wave forms, thus enabling using the real time nuances such as where and how the string is excited etc.

[RD Bennett]

Also, on "what's the market" - possibly not current steel players, but I think both kinds of technology under discussion could be invaluable in attracting new players to the field and leading to a revival of PSG interest / design changes.

I can't help but think that a lot of younger dudes (like, my mid-30s age and well below) are prevented from getting into PSG by the expense and learning curve. There are folks like Jerry Fessenden who will build you a 6-string PSG to mess around with-- if you know how to find these builders, which implies you've already done your homework. But the expense is still significant, and everyone tells you you'll outgrow the available changes.

Dump the mechanic underpinnings of a steel and you end up with an instrument that could potentially be made more cheaply using a lap-steel sort of framework and mass-production hardware. The parts of the mechanical steel may not cost much, but the setup takes a great deal of experienced labor, time, and specialized hardware, all of which could be cut down. Copedents could be made entirely fluid and variable on the fly. And, heck, you might even find some way to make the instrument one you could play standing up / strapped on.

It's possible that a six-string or eight-string guitar so equipped, with the right tuning / copedent, could cover most of the same ground as a standard mechanical E9th-- possibly *without* requiring as much tuning relearning for guitar converts.

Just some pie-in-the-sky thoughts. I would love to see a full-fledged steel revival, but the instrument as it currently stands has an awful lot of quirks that keeps the thing entirely out of the hands of younger players.

[RD Bennett]

RE synthesis...there are alternatives to storing and recalling sampled or synthesized wave forms, thus enabling using the real time nuances such as where and how the string is excited etc.

Exactly what I mean by using physical modeling - the DSP "imagines the string," has enough information to know what the right hand is really doing, and makes its sonic decisions accordingly. You probably couldn't use MIDI for the link between sensor and synth engine, too much data. It would have to be something custom designed for the purpose.

I just bought one of these last week. It's a self-contained PC for $35 that would easily fit inside a guitar and could potentially run such a synthesis engine. Everything would just need to be coded and designed from scratch, you're not going to find a box in the Guitar Center MIDI section that can come even close to sounding like a steel or a real stringed instrument, no matter how much data you throw at it.