[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Date Index][
Thread Index][
Author Index]
Re: analyse repeater's clock; was Re: Repeaters Clock out
At 01:57 PM 8/5/2002, Jon Wagner wrote:
>I have been thinking for a long time now to put this unstable clock issue
>to
>rest by analyzing the midi clock output of the repeater on an oscilloscope
>here at work. Only problem is that it could be a little tedious to
>manually
>measure each clock pulse to collect data.
you need a scope with infinite persistence mode, and probably some special
triggering to just capture the clock bytes at the same spot each time.
most
modern digital scopes have that. I could help you more if you really want
to try to do a jitter analysis like this.
>Has anyone done this and gotten
>real data for the stability (jitter, drift, ect) of the midi-clock?
no, but when repeater first came out it had a weird clock problem that I
investigated because it was causing the EDP to sync with the wrong loop
time. I made a simple max patch to measure the time difference between
clocks and log it. If I remember right I found that the 23rd clock pulse
of
each quarter note was coming out really late, almost right on top of the
24th pulse. That was tripping up the EDP, and probably other devices as
well. I'm pretty sure Electrix fixed this in the 1.1 update though. I
never
checked into it again.
>Also it
>would be interesting to see if certain conditions improve or harm the
>stability. Is there a way to analyze this somehow in software - for
>example
>just record the midi stream with accurate timestamps on each clock signal,
>then use excel for some basic calculations.
I found that MidiOx is ok for this, and certainly easier to get data out
than trying to use max. it timestamps midi events and logs them so you
could see what the clock is doing. We used it for debugging various midi
sync issues during LoopIV development and it was quite helpful. For some
reason MidiOx has a 2ms resolution in the time stamps. I don't know if it
is a MidiOx thing, my pc, or what, but I was a little annoyed by it. It's
good enough though to see if anything is wrong, since the clocks are
typically much further than 2ms apart, and 2ms difference shouldn't be
enough to cause a problem.
kim
______________________________________________________________________
Kim Flint | Looper's Delight
kflint@loopers-delight.com | http://www.loopers-delight.com