Re: [Jack-Devel] Installers for Jack 1.9.8 (SVN 4356) for Windows 64 and 32 bits

PrevNext  Index
DateSat, 30 Apr 2011 13:06:05 -0700
From Devin Anderson <[hidden] at charityfinders dot com>
ToPeter L Jones <[hidden] at drealm dot info>
Cc[hidden] at jackaudio dot org
In-Reply-ToPeter L Jones Re: [Jack-Devel] Installers for Jack 1.9.8 (SVN 4356) for Windows 64 and 32 bits
Follow-UpDevin Anderson Re: [Jack-Devel] Installers for Jack 1.9.8 (SVN 4356) for Windows 64 and 32 bits
Follow-UpPanos Ghekas Re: [Jack-Devel] Installers for Jack 1.9.8 (SVN 4356) for Windows 64 and 32 bits
On Sat, Apr 30, 2011 at 12:34 PM, Peter L Jones <[hidden]> wrote:

> Aha, that'll be because I typed it wrong... ;)
> http://dino.drealm.info/peter/media/jacklogs.7z

Thank you. :)

Interesting.  Both tests show that the timer wakes up within 1
millisecond of the ideal send time (sometimes early, sometimes late),
which means the timer code is introducing, at most, 2 milliseconds
peak jitter.  I can probably get that down to 1 millisecond of peak
jitter by introducing a second wait of 1 millisecond when the timer
wakes up early, but I don't think I can do better than that without
hackish busy-waiting due to Windows limitations on wakeup resolution.

That leaves incoming MIDI timestamp offsets and the MIDI device driver
as potential issues that I can see.

I'm going to take a look at the MIDI input handling code, and see if I
can find any problems.  If not, then I'll add debug statements that
will be output in verbose mode.  I would really appreciate another
test at that time if you're still willing to help. :)

-- 
Devin Anderson
devin (at) charityfinders (dot) com

CharityFinders - http://www.charityfinders.com/
synthclone - http://synthclone.googlecode.com/
PrevNext  Index

1304193990.1672_0.ltw:2,a <BANLkTinMOwZvy=hYuWBmjWq4Wwc3OqS-TA at mail dot gmail dot com>