Hi there Daniel and others! I was able to change this value and rebuild the driver. This has allowed me to set buffer sizes down to 32 frames. Now I am investigating what could be the source of so many xruns and audio problems at lower values that should work based on results from osx and windows. I am seeing intermittent streams of xruns with jack and very high utililization of the xorg process when this happens. I have also disabled my layla20 interface which has cleared up some glitches. So now I am in the process of trying to audit resources and eliminate any contention for that device at the irq and pci level. I have noticed that the USB ports I connect to have a default pci timer latency of 32, while the layla like pro interfaces has 192. I can get audio playback at lower buffer sizes than 1024, but the appearance of xruns and buffer errors still seems random and system related at this point. Hopefully if I can eliminate these other things, I can get a sense of what this device is capable of in linux. One bug: I am getting full scale static on the rig control input whenever it is monitored. Another possible bug is the device not being available on boot up, though I noticed enabling plug and play in my bios did make it available on boot once. Not a big deal since it is external and easy to reconnect. I'm not sure of what all the values near .periods_bytes_min mean in the source code, so I am not sure how my test value of 32 may interact with things like urbs and others. @Alsa folks: One question I have is whether is is necessary to rebuild all alsa sources every time I change the audio driver section or whether I can get away with just building the alsa-driver portion. It seems like these separate sources for driver, utils, firmware and libs isn't much of a favor for the user/listener. If they all relate and shouldn't be used separately, then why not pack them all together? To Daniel, without me understanding how the rest of the audio section works, is a patch warranted? Thanks! -jonathan On Aug 8, 2007, at 7:26 PM, Daniel Mack wrote: > Hi, > > On Aug 9, 2007, at 5:44 AM, Jonathan Leonard wrote: > >>> I'm not sure whether your jackd problems are acutally related to >>> the driver at all. >>> Have you tried your jackd config with any other driver, maybe >>> some onboard soundcard in your computer? Probably someone with >>> more jackd experience can give some more substantiated >>> information :) >> >> Yes I have tried the same config with another soundcard. The >> Echoaudio Layla20 which has worked reliably with a range of settings. >> >> +++/alsa-kernel/usb/caiaq.c >> Ln 63: >> .periods_bytes_min = 4096 >> >> Is this limit cool for low latency? Is it possible to get lower >> buffer sizes than 1024 frames? > > Why don't you go ahead and change this value, recompile the driver > and test it? If that works for you, send a patch to the alsa-devel > list and let the other developers know about that. I copy'n pasted > that value from another driver in a very early stage of development > - it's possible that it also works with a lower value. > > Best regards, > Daniel > ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ Alsa-user mailing list Alsa-user@xxxxxxxxxxxxxxxxxxxxx https://lists.sourceforge.net/lists/listinfo/alsa-user
