=== shnatsel is now known as SergeLion [13:23] scott-work: holstein has been asking about the kernel, he wants it to be changed to -generic. What are your ideas?:) [13:24] BTW, any news that a new kernel will be placed in 12.04.1? [13:56] smartboyhw, I didn't get that from what I have seen holstein say. [13:57] OK [13:57] Maybe I'm a bit wrong in describing [13:57] there is acually very little difference from generic to lowlatency [13:58] len-dt: The diff is RT, right? [13:58] holstein, has said that a lot of people would do just fine with generic. [13:58] It is not a RT kernel, that is different again. [13:58] Oh man [13:59] real time or preempt requires a patch to the kernel source [13:59] lowlatency uses the same source but builds with different options [14:00] (same source as generic) [14:00] That makes a huge difference in maintenance work [14:02] There are some dangers with using a RT kernel. a bad application can lock the system up solid. [14:04] RT works well with multi-core systems where c-groups can be used to set some cores RT and at least one to allow some time for other processes to run. [14:05] that provides some safety to the system with the ability to shutdown a rogue process. [14:10] smartboyhw: ? [14:10] Well, the kernel thing. [14:10] smartboyhw: i want us to be testing mainline generic kernels for rt priv support [14:10] OK [14:14] holstein, how are your channels numbered in ffado for your FW IF? Are they channels 1 to n or 0 to n [14:14] len-dt: they were 1 through 8 [14:15] so far with the new stack, the are left1 right1 left2 righ2... so on [14:15] i dont think ive ever had anything labeled 0 yet len-dt [14:15] So if someone has a FW device that is 0 to n that would be the driver for that device. [14:16] BTW, we really need testers, the 12.04.1 images? only len-dt and I reported to ISO QA Tracker:( [14:16] There is someone on LAD who seems to have a FW device that comes up in Jack as 0 to 8 or whatever. [14:16] len-dt: interestings... it used to bother me the inconsistency.. now i think of it as "the ghost in the machine" ;) [14:17] smartboyhw: imagine what its been like before you were here [14:17] If it was alsa a fake device with renumbered channels would fix it. [14:18] holstein: You mean that I am rude? [14:18] never really bothered me len-dt ...as long as i can tell whats what [14:18] He's having problems with things that auto connect [14:18] smartboyhw: ? ...no, im just saying, before you came along, it would have just been len-dt having tested the iso [14:18] holstein: Oh....:( [14:19] smartboyhw: im a glass half full kind of guy [14:19] smartboyhw, and that means only 32 bit [14:19] you might think of it as "oh know, we need testers" and im thinking "wow, we have another tester" [14:19] holstein: OK, that's weird [14:20] lol... i said, "oh know" ;) [14:31] holstein, the generic kernel could be made so that a boot time commandline switch would do the trick. [14:32] len-dt: put me down for testing that [14:32] anything to make life easier [14:32] assuming that would be easier [14:32] holstein: :) [14:32] i feel like i know just enough about it to be dangerous [14:32] holstein: It IS dangerous [14:32] holstein, I'm not much better [14:33] len-dt: nah... you are our top contributor, and i already have a christmas card picked out for you! [14:34] holstein, so how do get our branches uploaded? [14:34] len-dt: I will write one to you and also holstein and actually almost everybody:) [14:34] len-dt: great question! [14:34] for astraljava , stochastic , or scott-work [14:34] or knome maybe [14:35] I think people are busy with .1 right now. [14:35] yup.. and thats good too [14:35] True [14:35] .1 is what interests me most personally [14:35] i consider non lts releases "testing" [14:35] .1, .2, ,3 and .4 interests me [14:36] it has been suggested to me (should be on the list) that we make a new package called -docs [14:37] to replace ubuntu-docs which we don't ship. [14:37] len-dt: Good suggestion [14:37] probably not a bad idea... depending on our target audience i suppose [14:37] anyone know where the branch is kept? [14:37] Dunno:) [14:37] if we target current windwos/mac users doing audio, then it really isnt all that important [14:38] those folks dont look in the OS for help like that [14:38] i think thats a big change for new-comers [14:38] It's a GENERAL guide [14:38] Well the help stuff for nautilus is missing and maybe others too. [14:38] there is no "windows-studio" or "macOSX-audio" or whatever [14:39] its a little foriegn i think for folks to come to from windows for sure, where *everything* is different [14:39] not that we have to address or respond to that [14:39] just thinking outloud [14:39] Ya, the idea of connecting applications together is a linux thing. [14:40] which is something we can celebrate, and is really one of 2 things that makes it impossible for me to leave linux for audio or whatever [14:40] I am just thinking that without the docs these apps look broken. [14:40] len-dt: well.. are they? [14:40] They have a help button that gives an error [14:40] nautilus is really doing some odd things i know.. main ubuntu is going with an older version [14:41] len-dt: i think we should consider another FM [14:41] maybe just the XFCE one.. and soon [14:41] The user can choose thunar if they like [14:41] i think keeping nautilus in the LTS was a good call, and an effort to keep a legacy vibe with gnome2's apps [14:42] but, if nautilus is changing from that, and thunar is more like gnome2's nautilus, then maybe we should consider that [14:42] nautilus provides two things (I can think of) that we would not have otherwise. [14:42] or, give up on the lagacy look and feel as well... which im OK with too [14:42] file search, and remote connect [14:43] I use both. [14:43] len-dt: theres always something im missing when running xubuntu.. i use and prefer gigolo now for remote connect [14:43] len-dt: not that we need to include it... [14:43] len-dt: i also use something like kupfer for "search"... again just me [14:43] We have gigolo, but not all the backends [14:44] len-dt: i find it odd that nautilus is the only one that really has all this [14:44] all this in by default, and out of the box [14:44] nautilus just comes with everything, thunar needs extras to do almost anything [14:46] holstein, this probably means nautilus has a bigger memory footprint [14:47] but then thunar with all the extras to make it work the same probably uses a lot of memory too. [14:47] maybe more [14:52] len-dt: yeah, i personally like having one app, even if its bigger/slower that does it all [14:52] len-dt: especially since i feel like we need those things, since we used to have them, and users arguably need them and expect them [14:53] I have had people ask how to search, and I use remote for keeping my webpage up and backup stuff. [14:53] thunar is great though, and i suppose to be fair, someone should add that functionality and test the performance [14:54] prolly not something we need worry about soon hopefully [14:54] I have had some flaky things with thunar when it auto opens an inserted USB drive. [14:55] This is not a problem in 12.10 where we don't auto open things [15:15] smartboyhw: my latest talk with apw indicated that we will not receive an updated lowlatency kernel in time for 12.04.1 release [15:15] scott-work: Thanks [15:16] smartboyhw: not a deal breaker though, you think? [15:16] OK, so the image won't be changed anyway, and I can focus on #ubuntu-testing now. YEAH [15:16] as len-dt and holstein report, the generic kernel has been getting better and better as RT features have been rolling into the main branch [15:17] however, i would still point out that the lowlatency kernel has demonstrated it can improve latency by at least a factor of 2 (i.e. cut latency by half) in some cases considerably larger factors [15:17] scott-work: cool!.. we'll see if its ever up to the task we need it to be [15:17] ! [15:17] i think holstein was using his firewire device with 1.97msc latency, i have used dell computer onboard sound with 1.97 msec latency...mind you, this is stable latencies [15:18] smartboyhw: re: testing in ISO QA - as long as we meet the minimum number of tests we can get the image released, so i wouldn't get too upset or dramatic if we don't have more testers [15:18] yeah... i can get quite a bit done under 1ms with the old RT kernel and firewire [15:18] doesn't mean we don't want more, rather it isn't dire [15:18] on 10.04 [15:19] ! [15:22] len-dt: uploading branches - i would subscribe -sponsors and announce in #ubuntu-motu about needing sponsor (i disremember the link currently) [15:23] re: thunar, i wish it would do tabs :( [15:23] i think the developer stated in a forum that thunar would never do tabs, that is confusing to me [15:24] scott-work, micahg had already started looking at them. I don't want people to do the same work over again... [15:24] ! [15:24] sorry :) [15:24] nautilus is now set up much like chromium [15:25] (no menu bar) [15:28] holstein: In my estimate, the -generic kernel is nowhere near usable for live audio [15:28] holstein: But, if you can talk UTK into replacing it with -lowlatency, the problem would be solved [15:29] I'm in a sweet position at the moment. For the next month, I'm coding 24/7 [15:29] I'll try to wrap up the testing scripts as soon as possible [15:29] ailo: That's serious [15:30] I go wonky when I code that much... or even less. [15:30] I have so many things that need to be finished in that area, and I'm looking forward to having finished it all within this month [15:31] I got headaches when coding... [15:31] I mean big projects... [15:33] I think on both Mac and Windows you can get pretty low latencies. The downside is, most programs won't let you know when there's a "xrun" [15:33] And I don't think you can get the same latencies as you can with -lowlatency, on either Mac or Windows [15:33] Could be wrong, but that's my general experience [15:34] -generic however, is not nearly as reliable as either Mac or Windows [15:37] I share holstein's hope for a simpler future for pro audio on Linux. jack could be fully automated, and all of the sound systems could be seemlessly working together, without performance issues. All you'd need was one audio control interface, with a window for making connections [15:39] Full blown session support. Just save any configuration, or make it into a template. Use whatever applications you want, for anything, with anything [15:41] in a similar manner, i had some interesting feedback on UX and content creation from my post: https://plus.google.com/u/0/100313956509426913392/posts/953vgozPPDm [15:42] astraljava: Please tell me if I can add a upgrade testcase for US 12.10, I'm now a testcase admin:) [15:42] i think this shows a desire by non-geek musicians for a simpler way to use ubuntu studio for content creation [15:42] Someting I would think is [15:42] sorry.. [15:43] :) [15:43] Something I think is coming soon is more advanced web based applications. In all sorts of areas [15:43] BTW, very strange: Someone asked that if we are only doing LTS->LTS releases. Strange.... [15:45] oh, and i just found this blog: http://audio-and-linux.blogspot.com/ [15:45] interesting stuff [15:46] anybody remember an application for starting and monitoring jack from the panel? the developer has hung out in this channel quite a bit (note: this pre-dates falktx) [15:46] think it was a german developer [15:47] I think maybe only holstein would know [15:47] ;) [15:50] laditray sort of does that [15:50] qjackctl can be made to do that too [15:50] It has an option for it [15:51] it was something like "jack-panel" or similar [15:51] used pyjack or was related to pyjack [15:51] Guys, we have a problem, the people doesn't know the difference between Mythbuntu and Ubuntu Studio. [15:52] which people [15:52] "the people"? [15:52] Well, my friends in Ubuntu QA Testing... [15:52] Lots of people do know [15:52] They seemed to think the two are the same thing... [15:53] meaning? [15:53] They don't even know what's Ubuntu Studio:( [15:54] I think anyone who is looking for what studio provides will not have that problem. [15:54] Hmm....That is the most awkward thing I've ever heard about Ubuntu Studio... [15:55] studio is not for everyone [15:55] OK, but that's sad...:( [15:55] it does fill a large number of rolls, but there may be other flavours that do one thing or some things better. [15:56] I don't think that is sad [15:56] ! [15:56] That is why there are flavours in the first place. [15:56] Hmm [15:57] you can tell them that mythbuntu is for mythtv/PVR functionality (recording and watching tv on a computer) and ubuntu studio is for content creation [15:58] That's what I did. [15:58] Actually, does Ubuntu Studio and Mythbuntu use the same kernel? I mean 0lowlatency? [15:58] *-lowlatency [15:59] found my application - jackpanel: http://www.hans-baier.de/wordpress/jackpanel [15:59] Congrats [15:59] Someone who is interested in creating will not look at myth for long, someone who is setting up a mutimedia sysem will go there. [15:59] ! [16:01] ailo: you have an estimate about the generic kernel? you think it ever will be appropriate for US? [16:01] scott-work, I can make qjackctl look the same as the controller... the tray display is different though [16:04] News: I am now Ubuntu Testcase Admins! [16:07] holstein: Not with its current design and configs. And it might also depend on the sound system. I can't really tell. All I know is that without the preempt config enabled, latency might be ok, but very frail. How to change that, I don't know? [16:07] ailo: did you see len-dt 's statement? a boot time "argument" to the generic kernel? [16:08] If they can make the preemption more dynamic, allowing single applications to be very preempt(and whatever resources they need), just while they request it, that might work on a -generic/standard kernel [16:10] But, is that already how it is on -lowlatency? I don't even knowe [16:10] holstein: You can boot the -generic kernel to support threadirqs with a boot argument [16:11] ailo: i have no idea... [16:12] holstein: The only real difference between generic and lowlatency right now is two configuration options [16:12] Bye for now, off to bed! [16:13] Apart from the threadirqs patch, which can be replaced by a boot time argument [16:13] ailo, RT is that way. If you use no applications that ask for preempt then RT behaves like generic. [16:13] bye smartboyhw [16:13] len-dt: Then, -generic must be that way too. But, perhaps the Hz 1000 makes a tiny difference though [16:13] Bye holstein len-dt scott-work ailo [16:14] smartboyhw: bye [16:15] Can HZ1000 be set for boot time? [16:16] len-dt: Don't know. Don't think so. It would be interesting to find out about both configs though. I would think it's just a matter of adding patches to the code [16:18] Even if that was so, you still need to make the change happen manually. It won't be activated until the user requests it, either using a gui that supports such a feature, or by editing the grub config file by hand [16:18] run it past the kernel team (upstream) [16:18] We could ship our own grub config. [16:19] PR wise it would be a loss :) [16:20] Being able to say low latency kernel is nice. [16:26] len-dt: It would be kind of sweet to be able to do things with -generic that no one thought could be done. But, as most people don't know what -lowlatency is, they don't realize it is already being done [16:47] ya, I know. The kernel team knows too... so there is a general why bother attitude with them [16:48] most folks likely just bail thinking they need a realtime kernel and dont know what that means [16:51] There's -server, and -generic. Since -generic is not meant for servers, it might just as well be configured for preempt, and not voluntary preempt, as now. [16:51] Come to think of that [16:52] I should build two more kernels [16:52] One with preempt, but with slow HZ. The other vice versa [16:52] Which of the two does the most difference, really? [16:53] Might be -generic would suffer no problems at all being configured for preempt. [16:53] For whatever targets it is built for (a wide range) [16:54] ailo: no hit in battery life on the laptops? [16:56] holstein: Perhaps, due to the higher Hz [16:56] The sense is that cpu governing has minimal impact even. [16:56] Yea [16:57] there's gotta be some kernel tests for finding that out at least [18:09] http://www.phoronix-test-suite.com/ [18:10] hmm, it's in the repo in fact [18:10] phoronix-test-suite [18:10] len-dt: ^ [18:18] There's one test that measures battery power usage [18:19] There's also one called idle-power-usage [18:22] len-dt: scott-work, I can make qjackctl look the same as the controller... the tray display is different though === jussio1 is now known as jussi [18:29] heheh, i was going to apologize for getting into a meeting which spilt into lunch and that's why i'm now replying [18:29] when i was interrupted during this apology [18:30] len-dt: ailo: holstein: i would like your input on a thought: [18:30] i want to make it easier for users to experience the different content creation methods [18:30] i have mentioned various changes we could do, e.g. different panels, menus, icons, "docks", etc [18:31] this would streamline the various immediate options (or uses) based on the desktop [18:31] would using a different (and additional) set of xsessions be useful? [18:31] (consider this if we were to include some applet in the panel for controlling jack) [18:31] (or maybe a -controls applet as well) [18:35] or perhaps i should ask...if you had a perfect setup for audio to make it as simplistic and easy for you to use, how would it function? [18:35] forget scripts or existing apps, think abstractly! [18:35] think future perfect state [18:35] shift paradigms [18:40] scott-work, ... extra xsessions.. I am not thrilled with this idea on current HW. All of the workflows we support tend to be resource intensive. The idea of "wasting" resources on more than one xsession hurts. [18:41] Switching the look and feel along with what is running sounds fine. [18:43] That would mean saving everything in a session. (takes time) before starting the next. [18:45] Of course, with 64Gram and 16 cores... anything may be possible. However, when we get there the software will have evolved to use it all :P [18:45] scott-work: I'm going to implement jack controls to ubuntustudio-controls, and I'm kind of playing around with how to make it really simple. Obscure details that aren't important for running jack by hiding them in an advanced settings category. [18:46] With falktx jacklib.py, it's pretty easy to control jack [18:47] len-dt: I don't know what exactly an x session is, but what's the difference between loading one application at a time to loading them all at once (assuming I'm not misunderstanding something) [18:48] Current HW today should be regarded minimum dual core with 2 GB of ram [18:49] Some notebooks and netbooks have a bit slower cpus' but the 2GB limit is quite fair [18:50] I think a few people still use P4 based cpu's with older graphic cards, but they can hardly be used with a modern web browser [18:59] len-dt: the idea of a separate xsession would be to log into it, so therefore it wouldn't be extra resources, would it? [19:00] ailo: i really like falktx's cadence and i have asked him to incorporate a patchage functionality into it and it would be THE application used to control jack IMO [19:01] ailo: so i certainly appreciate controlling jack, but what else would -controls do? should these functions be parsed into separate apps? [19:10] scott-work, why do the logout/login thing? The idea of mode changing is to achieve the same thing in the current session. There should be less time from one mode/session type to the next that way. [19:12] i don't have a problem with that, i was just brainstorming about possible implementations [19:13] would we need to switch modes during a live session? should we? i don't know [19:13] can you change panel applets, desktops, panels, etc easily without logging out? [19:13] The main thing missing is to be able to know what apps have been started in this "subsession" so we know to exit them. [19:14] Ya, it can be done. But better would be to change them on the fly. [19:15] So far as changing modes as subsessions and for on the fly applications... there can be more than one kind of mode... called by different names :) but the idea is the saem in either case [19:16] okay :) [19:17] I don't know if xfce-panel is the best panel to use with this. [19:17] i was sitting in the meeting at work thinking about how jackpanel could change our interpretation of the desktop, or at least what we expect from it, and i thought about how the panels could change for different categories of work flows (audio vs graphic vs video, not just different audio work flows) [19:18] len-dt: but xfce-panel uses xml files, is that not good? [19:18] with the xfce panel to shift things the files have to be moved in and out. [19:19] and i realized that we already control the panels, etc via xsessions and realized we could provide different xsessions so the user could log into a particular session for a specific category of work flow [19:19] with different xsessions you need different users too [19:19] len-dt: could we not have a couple of specific xml files (e.g. foo.xml, bar.xml, etc) and copy them over default.xml ? [19:19] if you used different xsessions but the same home dir.... [19:20] len-dt: why would we need different users for different sessions? i'm doing that right now actually [19:20] i have my laptop with the default unity session and then i added xfce to it as well [19:20] It would be easier to mv a link to different .config/xfce4 directories [19:20] i use unity for one thing and the xfce as another [19:21] okay, a link would work then, correct? [19:21] xfce and unity store their desktop in different places [19:21] if a link didn't work, two mv operations would. [19:22] oh, i think i understand what you are saying [19:22] multiple xsessions for the same DE and user would cause some conflicts [19:22] The one problem is alacarte [19:23] changes made there are stored in .config/menus (I think) [19:23] Though again that could be switched out [19:24] we would have to make a decision about what settings would stay with the session and which would move with the mode. [19:24] maybe call it workflow or something else. [19:31] A new workflow would require the user to tell us they wanted to create a new one. Would need an icon that said create new workflow. In which case we would mv the current WF out and replace it with nothing. When the panel started it would grab system default. [19:33] scott-work, next problem would be how to ship it. Our program would have to start up before xfce stuff. grab the wf syetm defaults and install them... the same as xfce does. [19:47] len-dt: are you considering changing the entire or portions of the desktop per actual work flows (even between specific audio work flows)? [19:47] s/specific/discreete [19:48] for example... [19:49] let's say we shipped support for audio-recording, audio-generation, audio-podcasting, and audio-live work flows [19:49] -recording for recording live instruments (including vocals) and -generation for synths and such [19:49] i would not expect the desktop to change between those [19:50] i think we could include jackpanel applet integration and side panel icons for each of the work flows and keep things consistent [19:50] however, for other types of work flows (video, graphical, etc) we might consider changing the desktop or excluding things (like the jack applet) [19:51] again, i would prefer to keep the menu stock in these cases (since we are addressing work flows with side panels or similar) [19:54] scott-work, anything is possible... certainly small things like changing the background could be done. But I was mostly thinking about the panel [19:55] len-dt: i am asking when have you been considering a change necessary [19:55] were you thinking about changing between each and every discreete work flow or just between the general types of work flows? [19:56] That is open. [19:57] The amount of work to do it is kinda constant no matter how often you switch. [19:57] I think there are a number of workflows that are combined, though. [19:58] scott-work, you started by saying be open (my words not yours) to anything. [19:58] i understand :) [19:58] Really this is where my workflow bar was going [19:58] It is a panel that can change. [19:58] i was just asking your expectations about how often changes would be required based on work flows [19:59] scott-work, I think the availability of the software will dictate when and where it gets used. [20:00] as people have it to use, new ideas of how will show up, extensions will get made. [20:01] oh, i agree [20:01] but based on the current application sets i'm not sure too much needs to be done for different setups [20:01] i was just curious as to people's expectations about how many would need to be made [20:02] The thing when starting is to come up with a good framework for storing the workflow so those extensions don't require a rework of the whole thing. [20:03] I think in some ways many of the workflows we support (including graphics ones) have a lot of the same needs [20:05] agreed, in fact i'm working on a blog post that basically says the same thing (i.e. develop a flexible and/or extensible framework rather than hard code a static solution) === jussi is now known as Guest83481 === game2 is now known as Guest53091 [21:02] I found this very interesting video about music synthesis on Bob Owsinski's blog: http://bobbyowsinski.blogspot.com/2012/08/new-musical-instrument-monday-ractable.html http://www.youtube.com/watch?v=0h-RhyopUmc [21:03] click last link for video [21:25] scott-work, that could be a nice IF for audio.... as opposed to sound gen. [21:25] I was thinking a drawing tablet could be too. [21:47] len-dt: IF? [21:48] i wonder if patchage or jack connections can be made an applet too [21:48] this would put most jack infrastructure functionality in the panel rather than another application in the window list [21:51] scott-work, should I have made ui? User interface [21:52] ah, i understand