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

[Tom May]

"Theo" <t.hogers at home.nl> writes:

> Now that's a nice guitarsynt.
> This neuro-net thing is surly going places.
> Still the delay is around 12 to 13 ms, what is way to long. :-(

The page at http://www.musicindustries.com/axon/compare.htm gives
AX-100 timings of 11.7 to 12.9 ms for low notes on the low E string.
And the FAQ at http://www.musicindustries.com/axon/axonfaq1.html says

  According to Andras Szalay, the "inventor" of the Axon . . .
  tracking times for guitar are in the 4-13 msec range.

And later it says
 
  There will be no difference in tracking speed or response whether
  you use the Axon's internal sounds or an external tone generator.
  . . . The Axon's internal soundboard is merely on a parallel MIDI
  connection with the MIDI out on the rear of the Axon unit. What this
  means is that Axon deals in MIDI, and only MIDI, all the time.

And on http://ourworld.compuserve.com/homepages/bluechipmusic/tech.htm
Andras Szalay says:

  The delay is either 
      1 period of the string frequency + 0.5 msec 
      or about 6 msec, 
  whichever is greater.  This amounts to 13 msec for the lowest note
  of the guitar. 

Here are my actual results measuring the delay between the guitar
sound and the synthesized sound from several tone generators.  (The
timing for the sixth string improves to match the timing for the fifth
string on notes above the fifth fret or so.)

            Internal
String      Daughterboard    VL70-m    Matrix-1000
1           12-13ms          19-20ms   16-20ms
2           12-13ms          19-20ms   16-20ms
3           12-13ms          20-22ms   16-20ms
4           12-13ms          20-22ms   16-20ms
5           12-13ms          19-20ms   16-20ms
6           18ms             22-30ms   20-22ms

> A faster processor may help or a more optimised neuro engiene.

I don't think the neural net calculations are time consuming.  The
time it takes the initial picking impulse to reflect from the bridge
and fret causes some delay (Andras's 6ms?), and so does the MIDI
transmission (2-3 ms?) and processing (?).

> Now who's going to DIY this one???

Well you can start with US patents 5929360 and 5717155.  Interesting
stuff.

fTom.

> Cheers Theo.
> 
> 
> ----- Original Message ----- 
> From: Hallgeir Helland <hhelland at mailandnews.com>
> To: DIY <synth-diy at node12b53.a2000.nl>
> Sent: Sunday, May 21, 2000 11:55 AM
> Subject: Re: Harry's Rant was: Guitar synths (Don's rant)
> 
> 
> > Hi all,
> > 
> > I've been reading this thread and it's quite interesting.
> > 
> > FWIW: the AXON guitar -> MIDI seems quite nice.
> > I just recieved a demo video of it, and I am impressed.
> > It uses neural net technology to recognize the attack
> > of the signal, so the tracking delay is minimal.
> > It even recognizes the picking position - you
> > can for example choose three different sounds
> > just by varying the picking position.
> > 
> > http://ourworld.compuserve.com/homepages/bluechipmusic
> > 
> > I'd really like to have one!
> > Hallgeir
> > -- 
> > HELLAND MUSIKK TEKNOLOGI
> > -=  Hallgeir Helland  =-
> > hhelland at mailandnews.com
> > 
> >