/srv/irclogs.ubuntu.com/2011/01/29/#ubuntustudio-devel.txt

ScottLfalktx_, did you need something earler?02:10
falktx_ScottL: i just asked about what CMS you guys prefer02:11
falktx_joomla, drupal or wordpress02:11
ScottLdrupal for the website02:11
ScottLtroy_s, your comments the other night about dvd authoring being dead got me thinking about what an audio distribution could offer a musician02:14
troy_sScottL: LOL. Your mind works in strange ways.02:14
ScottLtroy_s, brainstorming a few ideas and will send an email to the mailing list later to see if anyone is interesting or wants to contribute02:14
ScottLtroy_s, i enjoy thinking outside the mainstream box, most people scare me with the way they think02:15
ScottLtroy_s, of course most of my "thinking" is really just ADHD02:15
ScottLtroy_s, but your comments/questions months and months ago about "what is ubuntu studio and who is it for" has stuck with me and i think about it often02:16
troy_sScottL: Don't blame the lunatic.02:17
troy_sScottL: We are utterly fearful of outwardly rejecting audiences (inner geek / nerd perhaps?) that we utterly fail to deliver to anyone in any great capacity. 02:17
ScottLtroy_s, no, no, it's a good question and it's one that probably should be asked periodically02:17
troy_sScottL: Just track the progress and you will see what I mean. In fact, track back far. I'll wager that the tendrils run deep.02:18
ScottLin some ways it might be like peeling away the layers of an onion, once you eliminate what it isn't you are left with what it is02:18
troy_sScottL: The core of all is audience as Jim Krause says - "Audience governs all." and in that you can sum up about all of Don Norman's work and plenty of other famous folks.02:21
ScottLtroy_s, i haven't really found a good way to articulate my thought yet, but i would like to explore how ubuntu studio can facilitate today's musician with exposing their music/video to others02:22
troy_sScottL: Audience gives context and helps to define needs. Saying 'everyone' is a complete lack of respect for design. It cannot be done. It makes horrible assumptions about cultural values, about age demographics, about prior knowledge, about societal norms, etc.02:22
troy_sScottL: Erm... that wouldn't be Ubuntu Studio would it? Unless you had some sort of single click to publish thing with Jamendo or Magnatune. And even then, that doesn't work.02:23
troy_sScottL: That whole rubbish of unfettered such is just anarchy... it doesn't work worth a rat's ass. Look at Android's market for example.02:23
troy_sScottL: So realistically, there is no magic bullet for a musician there either. Can you create work with it? I have no clue. I do know that if the audio works are as gimped as... uh it shall remain nameless, then the answer is no and it is hopeless to aim even as high as intermediate.02:24
ScottLtroy_s,  i was thinking about tools that would help convert to formats for vimeo or youtube and then a mechanism to automate posting their work to the same along with soundcloud, bandcamp, et al02:25
ScottLtroy_s, and maybe even including a tag line to send via identica02:25
ScottL"listen to my new song at..."02:26
ScottLbut this is just a single thought line from a single person02:26
ScottLmaybe i spark someone else's imagination and they think of something completely perpendicular but utterly brilliant02:26
troy_sScottL: Sounds interesting. No idea if that appeals to a given audience type. 02:26
ailoScottL: I think documentation for standard workflows is a good start.02:29
ailoScottL: If we think unix-like maybe such things like ladish will let us join all our little programs together. A lot seems to have to do with software design.02:31
ailoStill haven't tried that myself, though :)02:32
ailoAdding ways to publish on the web, aught to be providing info and links02:35
ScottLTheMuso, i have added a patch to the gnome-classic default xsession bug 70271203:10
ubottuLaunchpad bug 702712 in ubuntustudio-default-settings (Ubuntu) "set gnome-classic xsession as default in natty" [Undecided,In progress] https://launchpad.net/bugs/70271203:10
ScottLhi rlameiro 15:39
rlameiroHey ScottL :D how are you?15:39
ScottLrlameiro, doing well, fixed many things lately, lots of people working on good stuff for the future :)15:42
ScottLtoday is also by daughters birthday party, i'm not really looking forward to having a crapload of pre-teen girls in the house today and "sleeping" over tonight15:42
rlameiroNice... when is alpha 2 going out?15:42
ScottLi say "sleeping" because they really don't seem do any of it15:42
rlameiroooops...15:42
ScottLwe did this last year as well15:43
ScottL!schedule15:43
ubottuUbuntu releases a new version every 6 months. Each version is supported for 18 months to 5 years. More info at https://wiki.ubuntu.com/Releases & http://wiki.ubuntu.com/TimeBasedReleases15:43
ScottLhmm, that's not it15:43
ScottL!alpha215:43
ScottLfebruary 3rd is alpha2 according to15:43
ScottLhttps://wiki.ubuntu.com/NattyReleaseSchedule15:43
rlameirohummm almost there15:44
ScottLhopefully we get the gnome-classic default xsession patched and included before then :/15:44
rlameiroScottL: what do you thing that are the most important testing procedures for US?15:44
rlameiroI was thinking in making at least 3 audio tests....15:44
ScottLi still have to see about fixing the tasksel as well, but i'm not worried about getting it necessarily for A215:44
rlameiro1 - Onboard /souncard test15:44
rlameiro2 - USB interface test15:45
rlameiro3 - Firewire test15:45
ScottLall good tests15:45
rlameiroI think making to much test can be bad15:45
ScottLailo has been giving thought to testing as well, and documenting some of his and holstein's findings15:46
rlameirohumm ok15:46
ScottLthey really starting digging into what it takes to make a firewire interface work15:46
rlameiroI need to talk to them15:46
ailoYeah, it would be good to make some documentation for testers. 15:46
rlameiroailo: yeap...15:46
ScottLrlameiro, i think you and ailo might work together to cover more ground and faster, maybe divide the work15:46
rlameirothat was suposedely my job.... but i am slacking a lot....15:46
rlameiroshame on me15:46
ailoI've been looking into how to test lowlatency15:47
ScottLlol, it's all voluntary and other things get in the way15:47
rlameiroScottL: tru... like lilypond15:47
ScottLailo, i'm thinking of taking some time this weekend and testing the new -generic kernel (2.6.38) and then the -lowlatency and see how it compares then15:48
rlameiroailo: what firewire interface do you have?15:48
ailorlameiro: I guess it should be pretty straight forward, just collect a list of things that need to be confirmed15:48
ailoI don't have firewire, but I worked with holstein and a guy named tanders to do some tests15:48
rlameirohumm15:48
ailoAlready did some initial tests on -generic15:49
ailoBut we would need to have some more variables than just testing jack and maybe jack + a program15:49
rlameiroI tought  myabe doing an Ardour session with lots of plugins and at least 4 output channels to test firewire...15:49
ailorlameiro: You don't have any issue with firewire at all?15:50
rlameiroailo: where are ou from /TimeZone?15:50
ailo+1, but maybe 0 just now15:50
rlameiroailo: didnt tested yet in natty15:50
ScottLalong with testing the three audio interfaces that rlameiro mentioned, we should have them verify that they are running with -rt privileges15:50
rlameirook I am on Zero..15:50
rlameiroI need to go now to a meeting with a composer15:50
ailoThen I'm on +1 :)15:51
rlameiroi wil come latter to see about it15:51
ailorlameiro: I will be here. Just ping15:51
rlameiroScottL: also maybe generic kernel test and abogani kernel test15:51
ailorlameiro: Let's make a table of things. 15:52
rlameiroScottL: is the Natty US ISO usable?15:52
rlameirocan i install it on disk?15:52
ScottLrlameiro, i believe it is, but when you begin to login you will need to select your xsession at the bottom15:53
rlameirook15:53
ScottLit will default to "Ubuntu Desktop" or something similar, you will need to select "Ubuntu Classic Gnome"15:54
rlameiroso i will install it tonight ant test it with my firewire and see what is happening15:54
rlameirowell15:54
rlameiroI need to go now15:54
ScottLthis is the patch that i mentioned earlier, this will cause it to default to the gnome-classic xsession without having to pick anything15:54
ScottLbye rlameiro 15:54
rlameirocya15:54
ailoScottL: I did some testing on the new -generic and as earlier I can confirm realtime from the latest -generic16:26
ailoSo, no reason to add that to the test now.16:27
ailoBut it could be added to a full Documentation of testing a new distro, or a new release of a distro16:27
ailoI did just a simple testing table according to what rlameiro proposed now. http://paste.ubuntu.com/559948/16:28
ailoMaybe we can post the tests to the mail list, later to do some comparisons. That could be fun16:43
ScottLailo, awesome ideas16:53
ailoScottL: It think we should have a script for the test. Then post the results from the script to the mail list. Should be the easiest solution.16:54
ailoJust one test, to rule them all16:55
ScottLLOL, but an awesome idea too18:06
ailorlameiro: I had some ideas before about testing19:10
ailoWhat about making a script, post it on the mail list and let users add their results the list19:10
rlameiroailo: fire them up19:10
rlameirowhat kind of script?19:10
ailoI don't know yet, but at least dealing with jackd19:11
ailoAt first I went with your suggestion and did this http://paste.ubuntu.com/559948/19:12
ailoBut, in order to be precise I think we need a script, that many can use and compare19:12
ailoWe don't need so many testers, perhaps, but it could be fun to post it on the mail list anyway19:13
rlameiroyeap19:14
rlameirowell, it would be nice to make a sript that automated the test...19:14
rlameiroThat would be awesome :D19:14
ailoIf we do a script, we could start it  ./script -d alsa -d hw:1 19:14
ailoLike we would start jack19:14
rlameiroyea19:15
ailoThen collect info from the system, start some audio program and collect the results into a textfile19:15
ailoI don't know what programs or what kind of action to take. though19:17
ailoHave to eat. Be back later..19:17
falktxhey guys19:21
falktxhey rlameiro19:21
rlameirofalktx: :D19:21
falktxI have good news19:21
falktxI made my realtime-31 package based on abogani's one, and it's building19:22
falktxit it builds ok for lucid, then the same should happen for maverick too19:22
falktxwe may have some realtime kernels soon...19:22
rlameironice19:23
falktxcompiling the kernel takes so much time...19:24
falktxyay! the kernel compiled!19:54
ailofalktx: No graphic drivers with those, right?19:56
falktxailo: not yet19:56
falktxailo: this is still the 31 kernel, in a few days I'll upload 3319:56
falktxailo: abogani's 33-rt kernel is outdated though...19:57
ailofalktx: 31 for multiple distros, and 33 too?19:57
falktxailo: yes!19:57
falktxailo: 31 is building on maverick now19:57
falktx(lucid was fine)19:57
falktxailo: natty will start build soon19:58
ailofalktx: I'm sure some firewire users may have use for them. Did you check out the rtirq script yet?19:58
falktxailo: oh, not yet19:58
falktxailo: what do I need to do?19:58
ailofalktx: Don't ask me :). But I know you will need it. 19:59
falktxk19:59
ailofalktx: http://alsa.opensrc.org/Rtirq20:00
falktxailo: I basically just need to update it, right?20:00
ailoThe rtirq needs to be updated at least for the new firewire stack, so that it gives priority to firewire20:00
ailofirewire being the new name, don't remember the old name20:01
falktxailo: latest release of rtirq is from 200920:04
falktxlucid is already up-to-date20:04
falktxif anyone wants to try:20:06
falktxhttps://launchpad.net/~kxstudio-team/+archive/kernel/+sourcepub/1478697/+listing-archive-extra20:06
falktxinstall the *.all.deb and *.i386|amd64.deb (depend on your arch)20:06
ailofalktx: Here's how to install the rtirq http://my.opera.com/coreymwamba/blog/2009/01/03/automatically-realtime-priorities-using-rtirq20:07
ailoI'm not sure if that is the correct way to do it on Ubuntu though20:08
falktxailo: there's a "simpler" way for ubuntu:20:08
falktxupdate-rc.d rtirq defaults 50 1020:09
falktxreboot time, testing realtime kernel20:11
falktxrealtime 31 kernel working fine20:14
falktxthings seems a little slower though20:15
falktxthe desktop at least...20:15
ailofalktx: Did you try jack yet?20:17
falktxailo: will do that now20:17
ailoAre you using the rtirq?20:17
falktxailo: yes, installed by default in kxstudio20:17
falktxmy realtime test always involves wine20:18
falktxin generic, everytime I start a wine app, I got 1-3 xruns20:18
falktxhaha, now I got 0!20:18
ailoI usually get xruns no matter kernel when starting some apps20:18
falktxoh, 1 when adding new vst20:19
falktxhaha, I should the news on the forums20:19
falktx*should post20:19
ailoHow about latency compared to -lowlatency?20:20
ailoor -preempt in your case20:20
falktxailo: i did not test lowlatency yet20:20
rlameiroailo20:20
falktxailo: i have to do tests later20:20
rlameirohttp://testcases.qa.ubuntu.com/UbuntuStudio20:20
falktxwow, Maverick rt31 also compiled!20:20
rlameirowill try to add some procedures to this testcases page20:20
rlameirogoing to make some food and eat20:21
ailorlameiro: I guess I could help fill in some spaces there. I could start with the latency testing script. 20:36
falktxhm, guys, cya20:40
falktxrlameiro: eu n devo ir la, tenho mais k fazer....20:41
falktxxau20:41
rlameiroailo 20:55
rlameirowell, i dont know20:55
rlameiromaybe we should do the script first20:55
rlameiromaybe define a clear testing procedure20:55
ailorlameiro: What should the goal be?20:56
rlameiroafter that inplement it on a script20:56
ailoI don't think it matters if we start from both ends at the same time20:56
ailoI'm not a great scripter, but I thought I could at least start with making jack start and stop in different frames/period and then catch xruns and log them20:59
ailoAs a barebone20:59
ailoFor audio testing20:59
ailoI'm thinking about what form the results should have, what the script should accomplish21:01
ailoMany things I think needs to be done manually anyway, so maybe the script can be left simple.21:03
ailorlameiro: What do you think?21:03
rlameiroI am with you :D21:04
rlameiroailo: in python?21:04
ailorlameiro: Bash is hard enough for me :)21:05
rlameirook, i dont know mch of bash21:05
rlameirobut we can try to pull it off21:06
rlameirobut before that we need to know what to test for21:06
rlameirofor instance21:06
rlameirofirewire normally runs on 3 periods21:07
ailorlameiro: Why is that, anyway?21:07
rlameiroI think it is related with the firewire bus21:08
rlameiroalso usb card should be run like that at low latencies21:08
ailo3 periods per buffer should work as default for all devices, right?21:09
rlameiroat leas for FW ones21:09
rlameiro*leasr21:09
rlameirogrrr21:09
rlameiro"least21:09
ailoI don't think there is any problem with pci, so if it works best for firewire, I think we might as well keep it at 3, to make things easier21:10
rlameirowell, but it is quite diferent setup for pci21:10
rlameiroyou can achive lower latencies21:10
rlameirofor instance21:10
rlameiro2 buffers at 256 has lower latencie than 2 buffers at 25621:11
ailoOf course, but maybe we don't need to make a very strict latency testing21:11
rlameirowell yeah21:12
rlameirowe could try to do 2 diferent types of test21:12
ailoIf we do, we should plug in outputs to inputs and measure the real latency, but that becomes more work.21:12
rlameiroone for extreme values and another Standard21:12
ailoThere is a script for latency, latentor21:12
rlameiroit could be used21:13
ailoI'm looking for it now21:14
ailoHere's some info on jack latency testing http://wiki.linuxaudio.org/wiki/jack_latency_tests21:14
ailoThe latentor script was on that page http://rg42.org/gitweb/?p=latentor.git;a=snapshot;sf=tgz21:18
ailoBut it will only work if you connect inputs to outputs21:18
ailorlameiro: I also asked about testing kernels a couple months back. Here's the discussion http://lalists.stanford.edu/lau/2010/12/0400.html21:22
ailoTo do very specific tests, that may take some more work, but just to get a general idea, we can still keep it very simple21:23
rlameiroailo... we could try to make a pd patch very, very, very cpu intensive with fft etc and run it without gui :D21:25
rlameirothat would surely stress the ystem21:25
ailorlameiro: Good idea21:25
ailoI haven't tested yet the difference, but I feel pd is slightly harder to get to work without xruns than Ardour. Using pd at least gives us total control of CPU intensivity.21:27
ailoBut that test would show the relationship between stress and xruns, instead of latency and xruns21:27
rlameirowell, you cant do only one test for that21:29
ailoMaybe do a series, where jack starts first, and then after a while pd is started. Xruns will follow from starting pd, but those can be ignored21:30
ailoAs well as shutting down pd21:30
rlameiroyes, we need to give it some time to open and strat the patch21:31
rlameiroafter some time you only need to stop pd and then jack21:32
rlameiroafter that a new test21:32
ailoYes, that would work21:32
ailopd should track cpu to scale the intensity so it adapts to the system it is tested on21:34
rlameiroyes we can look at the load and if it goes avobe 98 shutdown a module and if it goes lower than 97 connect other21:36
ailoThat takes care of the audio testing, but there's all the other things to worry about, which seem like it should be presented as a list of manual tasks to go through, one by one.21:38
ailoLike boolean problems, does it work, or not21:39
ailoBut we can't have that for every piece of software. I think the tests should be directed at checking hardware21:41
ailoGraphic cards, cd drives, things like that.21:41
rlameiroyeap21:45
rlameiroi think ubuntu already has some kind of test on that21:45
rlameiroont the system menu or something21:45
ailoYes, so we only need to do things that are Studio specific21:45
rlameiroyeap21:45
rlameiroaltough we can direct them for that test for other stuff21:46
rlameirobut this is also fo ease of testing inside of US dev21:46
rlameironot for every user21:46
rlameirofor instance when there is a new kernel build we need to test it and see if it works or not21:46
rlameiroonly after that it would be available at ubuntustudio ppa21:47
ailoWould you want to do the pd patch and I do the script that starts it?21:49
rlameiroi can try21:51
ailoI can help with pd if needed21:51
rlameiroI could ask to the list the best way to stress DSP :D21:52
ailoI'm sure pd list people can too21:52
rlameirothey like that stuff :D21:52
rlameirocrash computers and stuff21:52
ailohehe21:52
rlameirome too by the way21:52
ailogood night rlameiro. I will work on the script tomorrow and make a quick version.22:05
rlameirook. I dont know if i have a patch for tomorrow22:05
rlameiroi can try22:05
ailono matter if the patch is empty. At least we have the backbone working :)22:06
ScottLof course, all the testing is relative and would be hard to pinpoint exact comparisons between various user/hardware data22:52
ScottLbut hopefully we can divine vast trends though, e.g. -lowlatency generally is able to perform at approximately X msecs less that -generic without xruns22:53
ScottLand comparing my data (with pci audio interface) with holstein's data (with firewire) might not have any direct correlation either unfortunately22:54
persiaailo, If you create a latency testing tool, please have it sample at various system loads.  Frequently a setup that provides the lowest possible latency at low loads can scale poorly, so someone who runs more effects, or someone with lower-power hardware needs a different arrangement.23:28

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!