Subject: Re:[OT] RE: Latency [linux-audio-dev] RFC: API for audio across network - inter-host audio routing
From: Charles Baker (baker_AT_charlieb.com)
Date: Fri Jun 14 2002 - 15:30:45 EEST
Joern Nettingsmeier wrote:
> it's not hard to hear. at more than 5-8 msec, you feel that beats become
> fuzzy, and below that, you can judge by sound coloration (comb filter).
> granted, it's a slightly different situation, since you have both a
> delayed and an undelayed signal combined, but it goes to show you can
> hear quite a bit. i know a drummer who claims to listen for the sound
> color of snare and ride together to check if he's in sync...
Well,
what this and the earlier gedankenexperiment about the jazz drummer on an
extremely upbeat tune implies ( and I tend to think it is true) is that
humans can judge delays between direct and indirect sound clearly into the
realm of phase relationships between the two signals (again, giving lie to
the homily I was taught 15 years ago about humans not perceiving phase
relationships in sound.)
Consider however *not* the case of direct / indirect sound delays (such as
amplification issues), but rather the case where one's instrument requires
a constant awaremess of it's inherent latency: Tuba, String Bass, ORGAN,
etc.: these are *still* playable, even in wonderful ensembles such as a fine
chamber orchestra: one merely learns a method of playing that takes into
account the delays...and although a sad picture was painted of the poor
drummer learning to adjust, *it has and can be done*....
Another example to keep in mind (one close to my home, and therefore onr that
comes immediately to mind) would be the humble "Salsa" band of the local
Florida hispanic communities:
Here we have a large group of percussionists, all locked into a fairl;y fast
beat, with a bass player who *continuously* plays at least 80-90 ms. (with
some *easily* upto 100 ms!!) AHEAD OF THE BEAT...while the body of
percussionists are unperturbed, and percussion solos REGULARLY float
"up, out of the tempo" and play solo phrases in a beat *up to several M.M.
ahead* of the other percussionists, then slip back down into the unperturbed
beat of the rest of the group. Meanwhile the bass has maintained *steadily*
their beat "in front of" the rest of the ensemble.
Doubt this? Go listen to a good salsa band.
I point this out for two reasons:
1) Humans can and do play instruments with much worse "response" than
the 10ms we want as minimum in out linux apps.
2) Even with a slow to resond instr. as above (SUCH AS AN STRING BASS IN A
SALSA BAND), subtle tricks with rhythmic prescision can be performed, and
can even be part and parcel of the performance tradition of the instrument.
Once played a work for a friend where I played a bass "drum" instrument built
out of an extremely large sheet of metal...ithad a strike to sound latency of
, well, at least 100ms (32nd note @ 60mm).
A few rehearsals, and one could not tell there was that latency in the
ensemble, unless one watched my hands...my sweety at the time was greatly
amused to see me striking the beast VISUALLY well in advance of the sound. I
just thought of it as a strange feeling , but not difficult.
I think that we *should* aim for latencies that make drummers and guitarists
happy, but I feel that latency has been given too much time recently in
computer music real-time issues. More important to me is the lack of control
data transmitted...we're still on the "turn it on with this volume, then turn
it off" paradigm of instrument control....MIDI has a lot to answer for!! What
about the intimacy of 'breath control' in winds, or the intimacy of finger on
the the string of an unfretted instrument?
There is a whole WORLD of instrument control issues that are just beginning
to be addressed, and to focus on latency is only to ignore what I believe to
be much more important issues of control.
Just my two cents:, YMMV, nothing personal, let's think about it, ok?
Love&Kisses,
CharlieBPhD
This archive was generated by hypermail 2b28 : Fri Jun 14 2002 - 15:29:23 EEST