[LAU] MIDI jitter - Was: outfitting a computer for songwriting in linux?

Ralf Mardorf ralf.mardorf at alice-dsl.net
Mon Jul 27 07:56:40 UTC 2015


"Academic studies have indicated that jitter becomes noticeable between
1 and 5ms, with a weighting around 2-3ms (Van Noorden 1975, Michon
1964, and Lunney 1974). Note that these tests only measured the
perception of jitter while listening, not while performing. Intuitively
I expect performing musicians to be even less tolerant to it."

"As sample buffer sizes are reduced, the jitter reduces"

http://expressiveness.org/2012/12/04/midi-jitter

When we measured the MIDI loops a few years ago people didn't believe
that I'm able to notice the jitter and they posted links about "good"
drummers who are more out of time than MIDI jitter is. Even while it's
true that a good musician is out of time too, it's done with human
touch, the human "jitter" provides the good groove. Machine's jitter
breaks the groove, it's arbitrary jitter, The musicians jitter is
aligned to the emotion the music should communicate.

However, the link above also claims that _it is noticeable_ and we
should keep in mind that a MIDI loop measurement doesn't care about
MIDI jitter of the receiving synth and jitter caused when the audio
signal is synced to the MIDI events. When recording very short click
sounds with a short attack, it's still hard to find the position were
the audio signal starts, when taking a look at the audio files, so
measurements a hard to do.

Another nice test is to record the MIDI events parallel to the audio
signal. With a high resolution for the MIDI events (quasi no
quantisation) playing the audio and MIDI track parallel should be
identical, but it isn't.

Taking a look at audio recordings of MIDI tracks, I noticed that jitter
peaks are periodically. Even while the graphics doesn't share the IRQ
with the audio gear, the periodic jitter peaks differ when using
different graphics.

Btw. I tried to find some of the tests we made a few years ago, but
searching the archives is not easy to do. At the moment I don't have
time to search my mails for that tests.


More information about the Linux-audio-user mailing list