/srv/irclogs.ubuntu.com/2011/02/11/#ubuntustudio-devel.txt

ScottLrlameiro, there is still quite a bit of work going on with the kernel i believe00:13
ScottLrlameiro, i've been having troubles myself and i really hope the solve them soon :P00:13
rlameiroScottL: well, maybe we should not post that testing untill stff is more stable00:14
rlameiroand continue to make test internally and look at them..00:14
ailorlameiro: Don't think your firewire problems are related, though00:14
ailoBut, the variation in test results is interesting. Just a week ago, I was getting the same results from generic as from -lowlatency00:15
rlameiroailo: yea, i need to do an update test again00:16
rlameiroi also tested it with 8x8 channels00:16
* persia is amused at the use of "internally" for a transparent public project00:16
rlameiro6 in channels where quiet but still it is a stress00:17
rlameiropersia: internally as not a definitive result :D00:17
rlameiropersia: not as obscure :D00:17
rlameirothis channel is public with logs :D00:17
persiaI know.  I'm still amused.  Even though everything we do here and on the mailing list is public, it's still "internal" in a sense.00:18
* ScottL wonders how many people come here and read through the logs or the emails00:28
rlameiroI do sometimes :D00:29
rlameiro1 for me :D00:29
persiaLots of folk likely subscribe to the mailling list, or read the archives.00:30
persiaMost of the folk who read the IRC logs are probably at least occasionally active here, or on the mailing list.00:30
astraljavaSooo.... wonder if Canonical could use a guy with some C++ and Qt skills? Looks like I'm gonna run out of work soon...09:43
aboganiastraljava: https://tbe.taleo.net/NA3/ats/careers/searchResults.jsp;jsessionid=E22F4015F5C1F74559120C79A16C04F0.NA3_primary_jvm?org=CANONICAL&cws=109:55
astraljavaabogani: Thanks! Way cool of you. :D09:56
aboganiastraljava: :-)09:57
ailoabogani: You heard persia, right? Is the kernel getting into the repos now?10:36
ailoThe -generic is nowhere near good for audio right now. 10:38
aboganiailo: Right. I'll double-check it this weekend. In any case I'm not sure that it is the right thing to do.10:59
ailoabogani: Why are you not sure?11:15
ailoI don't know what is going on with the -generic kernel, but until we can get a garantee that it will be able to do reliable low latency work, I would hold on to -lowlatency.11:22
ailo-lowlatency is the right technical term, but I wouldn't mind calling it -multimedia and let it be a multimedia alternative to the -generic, as long as the -generic is not able to provide the same performance.11:24
persiaThat would be unfortunate.  There are many "multimedia" use cases where it's probably not appropriate.11:25
ailopersia: Like what?11:25
persialike watching youtube11:25
persiaremember, "multimedia" is a *big* word.11:25
ailoIt is. You mean flash is not as fast with -lowlatency?11:26
persiaShould make no difference at all.  Might cause a (very small) decrease in battery time if I'm watching youtube on a portable device.11:26
ailoIt uses up more power, right?11:27
persiaA little bit.11:27
persiaDepends on what you're doing, really.11:27
persiaIf you're mostly idle, it will use more power.  If you're using your machine hard, it makes no difference at all, and may even save some power.11:28
persiawatching youtube is mostly idle for most recent devices.11:28
ailoMaybe we should add some info about that for the user11:28
persiaHardly anyone cares.11:28
ailoFor me, watching flash uses up most of my CPU. But that's because the graphic card is not working with it, right?11:29
ailoWithout drivers I can't even watch some videos full screen11:29
persiaThe only people who care are those who need low latency for some reason (audio or video production, audio/video streaming on insufficient HW, telephony switching, etc.)11:29
persiaThat's unrelated.  You aren't likely to notice difference there between -lowlatency and -generic11:30
aboganiindeed11:31
ailoFor audio low latency is needed as soon as you do something live. Play a virtual instrument, monitor while recording, or as I do, live processing11:31
persiaNo it isn't.11:31
ailoFor audio/video, I don't know11:31
persiaIt's needed for doing what you say under certain combinations of system load and driver quality.11:32
ailoI can't very well play a virtual instrument, if there is an audible delay11:32
ailoFor me it's totally based around human perception limitations11:32
ailoAnd -lowlatency nails it.11:33
ailoI mean, there's even a little room beyond that limitation11:33
persiaSure.  But the difference in delay between the average -generic kernel and the average -lowlatency kernel depends heavily on the load on your system (both raw CPU load, and context-switch count), the specific IO hardware you have, and the raw performance of your equipment/11:33
ailoAfter my recent testing, the difference is huge11:33
persia-lowlatency does better at higher system loads, or with buggier drivers.11:34
ailo-generic is not anywhere near acceptable right now11:34
* persia is talking in general "right now" is ephemeral11:34
ailoCan someone garantee that -generic will be able to perform as well as -lowlatency11:36
ailoFor me -lowlatency is a safe bet. I was surprised to see that the -generic performed as well for a little while. I have no idea why, and why it is not doing that now.11:37
persiaThere are no guarantees.  This is free software and a volunteer project.11:37
ailoThen, for the sake reliability, I at least would want to keep the -lowlatency as an audio alternative in the repos.11:38
persiakeep?  I'd like to put it in the repos, but will only do that if the maintainer wishes, as otherwise it will bitrot.11:39
ailoWell, put then :)11:39
ailoI don't know much about the work that goes into maintaining the kernel and how things work over at Canonical. I would think that -lowlatency would be useful, not only to Ubuntu/UbuntuStudio, but also to all other distros (and specifically audio distros) based on Ubuntu. So, from my ignorant point of view, I think the -lowlatency should go in.12:05
persiaMaintaining a kernel is a lot of work.  It's one of the most active, fast-paced, and confusing open source projects out there.12:06
persiaI'm not sure what Canonical has to do with anything in this context.12:06
scott-workabogani: we spoke earlier about someone learning to compile the -lowlatency kernel and i am still interested in learning to do so13:02
aboganiscott-work: Ok. Start cloning my git repos from kernel.ubuntu.com/git.13:04
scott-workabogani:  apparently because of the differences in our time zones this may be a slow process :P13:05
scott-workabogani: i will do that tonight at home when i expect you will be asleep :)13:05
scott-workabogani: but bear witness, this is something that i will peservere and see to completion13:05
aboganiscott-work: Could you recall your TZ to me?13:06
scott-worki am central US, in the state of texas13:07
scott-workabogani: ^^^13:07
scott-workabogani: i recall that you are in Italy but i don't know where geographically whithin her13:07
aboganiscott-work: I'm in central too. I was born in Prato 17 km near Florence (Tuscany).13:08
scott-workabogani: i meant that i am in CST time zone, my apologies, in the US using the term "central" is interchangable with CST13:11
aboganiscott-work: So it is incredibly early at you.13:14
scott-workabogani: i prefer to be at work at 6:30 am so that i may work undisturbed by colleagues, as supervisor i receive significant interruptions13:15
aboganiscott-work: Understand.13:18
scott-workabogani: but please don't take that to mean that i consider talking to you (or anyone on this channel) as an "interruption" :)13:24
scott-workif i did so then i wouldn't log into IRC ;)13:24
aboganiscott-work: ;-)13:30
scott-workdoes anyone also know of a good guide for creating themes? i would really like to understand some basics before we begin to make significant changes to the ubuntustudi-theme package13:32
aboganiscott-work: In any case I want you know that updating (that is the maintenance) of -lowlatency kernel is very trivial. 3 commands for ubuntu-natty-meta-lowlatency and 5 for ubuntu-natty-lowlatency.13:34
persia(plus lots of testing, bug tracking, etc.)13:34
scott-workabogani: wow, that really is insignificant!13:35
scott-workbut my personality prohibits me from trying to do something without having some generalized understanding of what is happening13:35
aboganipersia: touche :-)13:36
scott-workwhen i do something i like to understand what it should be doing and how it affects other things13:36
scott-workwhich i suppose aligns with what persia said as well :P13:36
persiaabogani, You may have scripted things so resync is quick, but I know it's a fair bit of work to keep the kernels working.13:36
aboganipersia: Kernels don't change significantly after release so we have a good chance that -lowlatency still in good still working.13:38
aboganiargghhhh13:39
* abogani needs coffe13:39
persiaabogani, heh, except for all the SRUs.  Seems there's one every couple weeks.13:49
scott-workabogani:  i belive it is approximately 3:30 pm on Friday for you.  is that correct?14:25
scott-workif so then you are +7 hours in relation to me14:25
scott-workthat's not as bad as i thought had originally thought14:26
aboganiscott-work: Exactly14:33
ailoscott-work: I will gladly take the lead on the Theme design, as long as we try keep it as trivial as possible. I have already learned some about GTK themes and Metacity.15:54
scott-workailo:  were you heavily involved in something else?  dont' overload yourself16:04
* scott-work apologizes for not remembering but work has gotten rather involved and time constrained today16:05
ailoscott-work: Not really. I'm focused on the -controls right now, but theming is always fun as relaxation ;)16:05
ailoscott-work: Here's the scheme for the ubuntustudio-controls I sent to paultag. http://paste.ubuntu.com/565935/16:32
ailoI updated the extra section16:33
ailoI'm looking into the memlock thing, to find out if there is any reason to edit that.16:33
ailoI'm also looking into the possibility of choosing which type of kernel to boot by default. After a quick look, it seems this needs to be done by adding a custom script to Grub16:35
scott-workailo: i'm running home now, i'll look at it there in about an hour16:40
ailoScottL: After reading more about memlock, some suggest that it is dangerous to use unlimited memlock. The amount of ram needed seems to be related to the software in use, but I guess using 75-80% would be a good thumb rule?16:55
ailoAs default16:55
ScottLailo, i'm not sure about the memlock amount for two reasons; 1) "upstream" is still discussing it and 2) it might depend on how much memory you have in your machine17:27
ScottLe.g. some might recommend unlimited, but for a computer with less memory, say 512meg or even perhaps 1gig on certain systems, you may not want unlimited17:28
ScottLtherefore i feel like some users might still need to adjust their memlocks settings17:28
ailoScottL: Im pretty sure its about how much you need to run programs, however percent-wise, it will differ when having little or much memory17:28
ScottLnow perhaps this is a fringe use case, and not even within two standard deviations from the typical17:28
ScottLmaybe we should send an email to the jack-devel mailist list and ask for their "expert" opinions17:33
ailoScottL: I'm for adding a control. Just adding some info too, that explains what is useful to use, and what could be dangerous.17:34
ailoDefault values: rtprio 95, memlock 75%17:35
ScottLailo, that sounds good :)17:35
ScottLalso ailo please keep in mind to implement a new theme will require more that just developing or designing it, we will need to make sure it is build and installed during the image building process17:42
ScottLthat later part for the current theme is where i have concerns about removing or modifying those parts for the new theme17:43
ailoScottL: I wouldn't mind learning about that too. If we do something drastic for this release, it would have to be simple to do.17:46
ailoScottL: I know a little about packaging, but don't have a lot of practical experience.17:49
ScottLsometimes i sound like such a naysayer :(17:51
ailoScottL: Paul Davis says there's no danger in using memlock unlimited.17:55
ailoScottL: I can not find one clear case where memlock would have been a definitive problem. Also, rtprio should not be changed either, I think. Right now, I'm thinking there's no need to add controls for those. 18:09
ScottLhi rlameiro  :)19:23
rlameirohi ScottL :D19:23
ScottLailo, that sounds good, if it does every become a problem we can always update the package19:23
ScottLrlameiro, did you ever get an install to work correctly?19:24
* ScottL is taking a break from recording music for the rpm challenge :)19:24
ScottLrlameiro, and what was your install methodology?  "upgrade" from vanilla ubuntu or use the dvd for full, fresh install19:25
rlameiroScottL: fresh install19:25
rlameiromanual partitioning19:25
ScottLdid it work for you?  are you able to record?19:26
ScottLi remember you saying you were having trouble with xruns19:27
rlameiroyeap, i have a lot of them19:28
rlameiroi think i really need an RT kernel to have somewhat reliable session using Firewire19:28
rlameirowith my chipset19:28
ScottLrlameiro, is this a different computer than you used before?19:31
ScottLrlameiro, or did you previously use an -rt kernel when you didn't get xruns?19:31
ScottLwhat is different this time?19:31
ScottLbesides the xruns, of course :)19:31
rlameiroI used a RT kernel before19:31
rlameirokarmic19:31
ScottLah!19:31
rlameiroso i think it is really related to this19:31
ScottLi think holstein was rocking the -lowlatency kernel into insanely low latencies19:31
rlameirolow latency doesnt solve firewire issues unless you have a TI chipset...19:32
ailoBut you said you had problems with -rt too, right?19:32
ScottLyou might try it before going to -rt19:32
rlameiroailo: not as much as lowlatency19:32
rlameiroon RT i had like 2 Xruns an hour19:32
ailoIve heard bad things about the ricoh chip19:32
ScottLi think i will spend some time talking to rnbc next week to understand more about his rtirq package19:33
ailoBut, if the problem is solved with -rt and rtirq, then there is probably an issue with irq sharing. rtirq will only work with -rt19:34
rlameiroyep, my firewire chipset has a lot of shared IRQ19:34
ScottLailo, sorry, i just now checked my mail and read your email to jack-dev list, good job and paul answered!  kinda like rubbing elbows with royalty :P19:36
* ScottL is going back upstairs to record some vocals and maybe another guitar track19:36
ailoI spent some time talking with ffado people about irq sharing. They said the problem should be solved by the people developing drivers for linux. Graphic drivers and such. I'm not sure about the details, but it seems hard to get away from the problem with irq sharing on vanilla / generic kernels right now.19:36
ailoThe problem is mainly on laptops too. 19:37
rlameirowell, i am on laptop :(19:39
ScottLrlameiro, i liked your answer to the alesis question on the mailing list..."why are you using OSS?"   LOL19:57
ScottLi'm pretty happy with the rest of my tool chain for guitars, bass, drums...but not my vocals, i'm still struggling with that :(19:58
rlameiroScottL: right isnt it?19:58
ScottLrlameiro, i believe so!19:59
ScottLbut there are those who would argue that OSS should be preferred over ALSA20:00
ScottLbut then again, for those with hardware that only works with OSS, can we argue that it isn't needed ?20:00
rlameirohttp://createdigitalmusic.com/2011/02/the-79-virtual-analog-console-now-on-both-mac-and-linux-harrison-mixbus/#comment-121524720:03
rlameiroScottL: hey :D PipeManMusic show me this20:11
rlameirohttp://www.harrisonconsoles.com/mixbus/Linux_screen_1.png20:11
rlameiroits ubuntu studio20:11
=== detrate` is now known as detrate
ScottLi saw that rlameiro from my rss feed, it looks fooking awesome, doesn't it22:11
rlameiroScottL: yes :D22:12
ScottL[lsd] in the #opensourcemusicians channel was helping me with qtractor and midi...he freaking rocks too :)23:10

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