🚀 go-pugleaf

In article <ajb.671294655@hela.iti.org> ajb@iti.org (Al Boehnlein) writes:
>I have an idea for a DSP project that might interest the guitar players out
>there.  I was wondering if it would be possible to write a program that
>would produce tablature of a song.
>
> ajb
Good luck! Unlike many musical instruments, the guitar often has 22 frets
(half-step quantized pitches) and the distance between string pitches is
only a 5th. This means that there is considerable ambiguity as to where
in the tabulature you would place a note. e.g.: the following are ALL ways
to finger the same C note:

     3   5   7    10  12    15    17  19    22
|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|
|0|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|
|-|-|-|-|0|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|
|-|-|-|-|-|-|-|-|-|0|-|-|-|-|-|-|-|-|-|-|-|-|-|
|-|-|-|-|-|-|-|-|-|-|-|-|-|-|0|-|-|-|-|-|-|-|-|
|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|O|-|-|

A good guitarist can hear the subtle changes in tone (spectral components)
and can, with the help of the process of elimination (there can be no more 
than 6 simultaneous notes, and the order is usually obvious), can often
determine a WORKING fingering. A computer program would have to be clever
enough to know which fingerings are possible, which are not possible, which
are probable, anatomically speaking. The task would be much easier if you
had access to the guitarist, and could put a hex pickup on his guitar. This
would eliminate the inherent ambiguity.

This is a project that has long interested me, but my DSP experience is 0.
My '81 college Sr. project was a monophonic real-time (analog) "harmonizer", 
and part of the problem was determining apparent pitch. Producing tabulature
would have been merely an information display problem. But then there's the 
polyphonic nature of the guitar...

-- 

Ron Anderson			"He is no fool who loses what he cannot keep to
rona@videovax.tv.Tek.com	gain what he cannot lose." Jim Elliot, martyr

On a related (well, sort of) note, regarding Electronic Guitar Tuning,
that was precicely what our first Electronic Design project was this quarter.

And let me tell you, that it is mucho difficult to quantify a vibrating 
guitar string without resorting to DSP technology (we were using strict
analog).

It was basically a mic, and amp, a filter to get rid of hi-freq garbage,
and then the most futile attempt at a half-a-dozen methods to get a good
signal to send to a freq-voltage converter (to determine if a string is
in tune or not).

Depending on 1) how the string was plucked, 2) where it was plucked,
3) how hard it was plucked, 4) how in-tune it already was,  and 5) the
phase of the moon, we got different results. Harmonics were popping in
and out of the signal as the string amplitude diminished, and I can under-
stand 1), 2) and 3) for this cause.

Most surprizing was 4): Apparently there are natural harmonic resonating
frequencies of the body of the guitar which heighten or dampen various
harmonics. It also changed depending on how I dampened the vibration of
the guitar with my body.

All in all, a very frustrating project!  Now to write up the report and explain
why it didn't work 100%....


			Steven Christensen
			Univ. of Cincinnati

-- 
Steven V. Christensen
U.C. College of Eng.
schriste@uceng.uc.edu
For the adventurous: svc@elf0.uucp

schriste@uceng.UC.EDU (Steven V. Christensen) writes:

>On a related (well, sort of) note, regarding Electronic Guitar Tuning,
>that was precicely what our first Electronic Design project was this quarter.

>And let me tell you, that it is mucho difficult to quantify a vibrating 
>guitar string without resorting to DSP technology (we were using strict
>analog).

> [ much frustration deleted ]

According to the Hal Christiansen book (I've forgotten the title, but
it's a huge book on computer music) you get much less harmonic junk
if you put the pickup in the middle of the string.  Pickups are down
at the end particularly to get harmonics.  And, of course, you want
a separate pickup for each string.

While we're on the subject of electronic musical instruments, are
there any "lab i/o" MIDI boxes available?  That is, you get a MIDI
port, a DSP, and a bunch of A/D & D/A lines.  You download programs
via SYSEX and wire up your own stuff to the box.

Lance Norskog
of