Subject: [linux-audio-dev] Report from the ICMC
From: Ivica Ico Bukvic (ico_AT_fuse.net)
Date: Mon Nov 08 2004 - 03:23:26 EET
Hi all,
I just wanted to share with you a quick report from the ICMC. The conference
was much fun although many events were overlapping and therefore it was
impossible to attend everything.
A number of new open-source audio-oriented software was demoed many of which
are also available for the Linux platform.
My Linux demo went relatively well (apart from the 2 consecutive X-server
crashes, courtesy of the crappy binary-only ATI's driver, and an odd issue
where disconnect of hdsp pcmcia did not yield acknowledgment from the 2.6.7
kernel and therefore subsequent reconnect did not recognize the card's
presence). There were ~20+ people in the audience at any given time
throughout the presentation (some of them were there only for a portion of
the presentation, so people kept coming and going).
Stuff I demoed (in no particular order):
*Jack
*Freqtweak
*Jamin
*Pd/Gem
*Latency tests
*Ardour
*Sweep
*Rezound
*Audacity
*gAlan
*Rosegarden
*Qsynth/fluidsynth
*ZynAddSubFX
*VKeybd
*Cinelerra
*Blender
*Xine/DVD playback
*Spiral Synth Modular
*KDE productivity
Stuff I wanted to demo but ran out of time:
*Hdsp latency/performance
*RTcmix synthesis language
*Supercollider
*Csound
Etc.
Stuff I observed preparing and doing this demo (please understand that I
have no idea whether these problems are result of my own setup or are
justifiable bugs -- nonetheless I am including them here in hope they may
shed some light towards their resolution):
*ZynAddSubFx crashes when a lot of polyphony is created even though the cpu
utilization is not topped-off
*Jamin crashes when pushing limiters to the extereme (not consistently)
*fluidsynth has some weird looping problems (I am trying to track this one
down as apparently this is specific to my setup)
*Ardour's real-time preview via sliding the timing bar does not work (I do
have one release prior to the latest, though)
*Spiral Synth Modular crackles a bit even with large buffer sizes when using
OSS (haven't had the time to do exhaustive tests nor did I try using Jack
object as of yet)
*Audacity has some instabilities when used in conjunction with Jack
*There is a real need for Jack to be able to restart (should a need for its
restarting ever arise) without losing all of the connections
Stuff that really rocked:
*I was able to run Jack on my via82xx laptop soundcard using rt mode with
64x4 buffers (5.33ms at 44.1KHz) *almost* rock solid (there is still some
issues due to crappy ATI driver and obviously 2.6.7 kernel that is still
sub-par to the 2.4.x kernel series performance, but that mainly amounted to
perhaps a 1 xrun/minute).
*I was able (although after the demo) to run hdsp with 64x2 (2.9ms at
44.1KHz) *almost* rock solid (random occasional xruns but no xruns when
adding new connections and/or apps to the Jack session).
*I tested the PD with the hdsp's 64x2 buffer size and used the latency test
patch by connecting audio out with audio in on the multiface and got ~6ms
:-).
The general reaction from the audience was quite positive as many of the
users who even were familiar with the Linux Audio scene were impressed by
the diversity and quality of the software offering as well as the
flexibility of the Jack's framework and the overall user-friendliness of the
UI/desktop environment.
I would like to therefore use this opportunity to once more thank all of you
for being such active and generous contributors to this great community.
Without you none of this would have been possible!
----------------------
The panel on the "Standards from the Computer Music Community" was very much
interesting as it covered many of the important facets of today's computer
music scene, but also more importantly revealed some of the greatest
strengths of Linux, including (but not limited to):
*openness of the standards and therefore ability to generate umbrella
meta-standards (i.e. LASH)
*longevity
*ability for a new standard to supersede the reigning old standard solely
based on its merit (i.e. Alsa vs. OSS), and not due to its commercial PR
and/or widespread use
*Minimization of the misrepresentation of the standard's features and/or
abilities
Having had this wonderful opportunity to be a part of such panel (many
thanks to Matt Wright for inviting me!) has truly reinforced and clearly
defined the advantages as well as strong reasons for being a part of this
community.
Many thanks go to all of you who have generously offered your insight in
these issues.
Time permitting, I will post a more in-depth list of the ideas I've covered
during the panel. Matt Wright should also have slides ready on his site
sometime soon.
Best wishes,
Ivica Ico Bukvic, composer & multimedia sculptor
http://meowing.ccm.uc.edu/~ico/
--- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.788 / Virus Database: 533 - Release Date: 11/1/2004
This archive was generated by hypermail 2b28 : Mon Nov 08 2004 - 03:38:11 EET