[linux-audio-dev] Report from the ICMC

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