Harry's Rant was: Guitar synths (Don's rant)

[WeAreAs1 at aol.com]

matti at devo.com wrote:

<< Don T. mentioned another Roland g-synth that uses a 
distorted version of the vibrating string as the attack 
sound and then switches over to a VCO once it has "found" 
the correct pitch.  >>

For some reason, I can't seem to find the original message of Don's from 
which this was paraphrased.  I hope I didn't carelessly delete it - I'm 
really enjoying this thread, and I'm always interested in Don Tillman's 
thoughtful posts.  However, it should be noted that the Roland GR-300 floor 
module doesn't quite work exactly that way.  It does indeed use a distorted 
and lowpass filtered version of the guitar's string output to analyze for 
pitch-to-period conversion, but it doesn't put that distorted string sound 
into the audio path either before or after the VCO has settled on a stable 
pitch.  

When you strike a note on the GR-300, unless you intentionally mix some 
straight or hex-fuzzed guitar into the output, the sound you hear is 
generated by the VCO, period.  The VCO's generate something close to a regular
 sawtooth wave, whereas the hex fuzz puts out a more spiky/squareish 
waveform.  There are actually two VCO's, one of them using the guitar 
string's period to start and reset a linear ramp generator, and the other one 
using a voltage derived from the peak amplitude of that linear ramp as a 
pitch-to-voltage converter to drive the other VCO, also a linear ramp 
generator.  Both VCO's, as well as the hex fuzz signal, feed into the same 
Mixer>VCF>VCA, so you hear them all attack at the same time.  If you enable 
both VCO's and the hex fuzz, you're actually getting three sound sources per 
note (although the hex fuzz signal and VCO #1 are necessarily locked in phase 
with each other, so it can't ever sound like three free-running VCO's).

Furthermore, the aforementioned hex-fuzzed signal is not the same signal as 
the distorted string signal that the GR-300 internally analyzes for pitch 
conversion.  The hex-fuzzed signal is actually generated at the guitar, not 
the in GR-300 module, and it completely bypasses the GR's pitch conversion 
circuits, going straight to the filter, VCA, and output.  The GR-300's 
minimum attack time is internally pre-set to a slight ramp, to compensate for 
the slight latency of the VCO circuitry, so even if you're using both hex 
fuzz and VCO's into the filter, you hear both of them attack at the same 
time.  This might be what Don was referring to as the "distorted version of 
the vibrating string as the attack sound...then switches over to the VCO".  
I'm sorry if I'm sounding nit-picky here, but I hope my description shows the 
distinction and clarifies his comment a bit.

BTW, the only guitar controllers that Roland made that had the built-in hex 
fuzz circuitry were the G808 and G303 (the two that we've seen Pat Metheny 
playing).  All of the other models (G202, G505, and G707) did not have the 
additional hex fuzz stuff, just the (undistorted) six-way pickup signal and 
straight mono guitar output.

Nevertheless, Don is quite correct that the GR-300 system has essentially 
zero delay (taking into consideration its slight attack ramp in its VCA/VCF 
envelope).  If you ever get the chance, try one sometime.  You'll be amazed 
at how much it's like playing a "normal" guitar, especially if you've been 
beating yourself up playing any of the modern-day MIDI guitar synths.  It's 
no wonder that Pat Metheny, one of the truly great musical geniuses of our 
time (and a very savvy and demanding techno-nerd), still uses it on every 
recording and in every concert.  If you've ever seen him play his orgasmic 
masterpiece "Are You Going With Me?" on that axe, you know just how 
emotionally expressive and incredibly musical a guitar synthesizer can be.  
(In the right hands!)

Michael Bacich

P.S. - I would be interested in discussing the GR-300 with any other list 
members who have one, or are interested in them.  Besides my previous 
ruminations on increasing its gestural picking response, I have this crazy 
idea of trying to use the GR-300's internal pitch-to-voltage circuitry to 
drive the VCO's in a Korg Poly 61 or PolySix (since both of those six-voice 
synths have Korg's unusually stable linear VCO's, which need no temperature 
compensation, and which might mate well with the GR's already accurate 
Hz-Volt pitch CV output).  This, hopefully, would increase the GR's limited 
sonic palette, and also give it access to programmability for live use.  
Anyone who has thoughts on this is encouraged to reply, either privately or 
to the list.    - MJB