Midi Latency, was Re: Bob Sellon JamMan upgrade was midi-locked delays and stuff

["Jon Wagner" <jondrums@hotmail.com>]

> Interesting comments from Bob regarding the whole MIDI latency issue in
the
> JamBoy.  To be fair, I have found that the MIDI clock (both sending and
> receiving) is fine in the JamMan, and it plays well with most other MIDI
> stuff, even EDPs and 'peaters  MY issue has been, and continues to be, 
>the
> latency in tapping in and out of loops (it is slight, but my stuff is
highly
> rhythmic, and those miliseconds add up....to frustrations).

I honestly can't completely agree with you here, max.    Yes there is
latency introduced when using midi to control timing critical events.
However, even if your pedal has a 5ms latency and the Jamman also has a 5ms
latency, that's only 10ms latency, and that is similar to the latency that
is needed for a switch closure!

The real issue in using midi with looping, is whether that latency is
_repeatable_.  In other words if I can rely on the latency to always be 
20ms
from the moment that I press a button, I never have a problem.  Say I press
the button on beat1 to start looping.  The loop will start recording 20ms
late.  Then if I press the button to end recording on beat1, the loop will
stop recording 20ms late too.  My loop ends up perfectly in rhythm!    
There
is a limit to acceptable latency, because it will tend to clip the attack
from any note I might play on beat1 - truthfully, I usually end a loop with
the same sound as I start with, so for me this isn't an issue.

On the other hand, if I can't rely on the latency and it is specified at
20ms +/- 15ms, then we have a problem!  If its 5ms on the loop start, and 
35
on the loop end, then we have a loop that is 30ms too long - BAD news.  I
admit _some_ midi devices can certainly have this problem, but in my
experience I could only imagine this happening in a very poorly designed
device or one that is compromised by a cost cutting design (usually at the
software level).

It is my understanding that a well trained human ear can hear differences 
in
timing as little as 10ms.  Therefore in my book, if the latency is
repeatable to within +/- 5ms we're fine for a looper.  This is not at all 
an
unreasonable goal to expect of a midi chain.

So, not being a jamman owner or user, I have to ask you Max: How have you
compared midi control to foot-switch control?   If you have seriously 
A/B'ed
the two techniques and noticed concrete examples of midi control having
rhythmic problems - I yield to your opinion 100%.  It certainly is possible
for the jamman to have problems with midi timing control.  I'm not trying 
to
be a midi evangelist, I would just hope that you give it a chance with a
quality midi controller.  I will offer to let you try out my behringer midi
footcontroller next time we have a chance to meet up.  Let us all know how
you go...

Jon
ps.  I have seriously A/B'ed midi control to foot-switch control on the
repeater and echoplex and midi control has fared quite well.  I consider my
ear to be a well trained ear in the rhythmic sense and I cannot honestly
detect a difference in timing between the two methods.