Re[2]: Midi looping

[miguel.barella@poyry.com.br (MAT)]

     Sorry for showing you this JamMan very nice "feature"...
     
     Keep me informed if you find a workaround.
     
     I am also including the explanation I got from Bob Sellon:
     
     
     Miguel,
     
     There are two possible causes for the noise I can think of; one you 
     can   fix the other you can't.
     
     When Jamman is slaved to an external MIDI clock it determines the end 
     of   the loop by itself based on the tempo of the clock and the 
number 
     of   beats selected on the front panel. If the operator tries to end 
     the loop   manually, the resulting loop will very likely be shorter 
     than it should   be. When the loop is too short, Jamman restarts the 
     loop twice: once when   the end of the loop (time) is detected and 
     once when the correct number   of MIDI clocks has come in. The 
     solution is to let Jamman close the loop   by itself (DON'T tap a 
     second time). This will get the loops size to it's   best fit.
     
     The second cause is based on the jitter on the incoming MIDI clocks 
     and   the resolution of Jamman itself. At best, Jamman can lock in a 
     loop size   to within half a millisecond (512us). The problem is that 
     most MIDI clock   sources have jitter (timing variations) in the same 
     neighborhood. After   the loop time is locked in, the priority in 
     Jamman is to stay in perfect   sync with the incoming MIDI clock. The 
     problem is that the combined half   millisecond resolution of Jamman 
     and the jitter on the incoming clock   result in the actual size of 
     the loop changing very slightly every time   through. As the loop 
size 
     changes, Jamman either shortens the loop or   replays the very 
     beginning of the loop to compensate resulting in   potential clicks 
     and pops. With the PC itself being slaved the jitter   gets worse and 
     so do the clicks and pops.
     
     As I said, there is currently no work-around for this other than, as 
     you   said, not playing anything at the loop edge. The only other 
     thing I can   suggest (which is equally klugey), is to place 
something 
     percussive at   the splice point which will tend to mask the noise. I 
     am looking at the   problem, however, and will let you know if I come 
     up with anything.
     
     If anyone out there has any suggestions on how to deal with this, I'd 
 
      love to hear it.
     
     Bob Sellon
     Lexicon/Stec
     bsellon@lexicon.com
     
     
     Regards,
     
     Miguel
     


___________________________ Separador de Resposta 
______________________________
Assunto: Re: Midi looping
Autor:  "Robert S. Carter" <rsc4@hmivax.humgen.upenn.edu> na INTERNET
Data:    17/07/1900 11:38


Y'know I never really heard the glitch so much until I went home after 
reading your post and listened carefully. Now it's gonna bug the hell 
out of me. Thanks a lot :). Some loops it's not so bad but yeah it can 
be annoying.  Once layered with the multiple loops I get going on the 
Nord synth I think it's barely noticable (or at least it used to be). I 
geuss I might reconsider ways use the JamMan as the master. The problem 
is I need clock before recording a loop and the JamMan won't do that. 
But now I'm thinking I could record a blank loop to generate clock from 
the JamMan and then use punch-in mode to record the now synced synth. 
I'll have to try this tonite. BOB. 
     
MAT wrote:
> 
> Robert S. Carter:
> 
> >So I send that clock to both the JamMan and the synth and the two are 
> >as one. Let's say I start an arpeggiated pattern on the synth (synced 
> >to clock), I start recording a loop on the JamMan and what pops out is 
> >a perfectly timed replicate of the pattern. I can then use that as a
> >backdrop for more syncronized synth patterns and noises. 
> 
> How do you manage the annoying noise (sounds like a small glitch) that 
> appears next to the loop boundary when the JamMan is receiving MIDI
> clock?
> 
> I tried to get rid of it with all kinds of tricks I could think of but 
> did not succeed.
> 
> Miguel
>