[00:00] ailo, would you be able to get all the stuff you learned onto a wiki page? [00:03] ScottL: Well, about ffado, it seems we don't need to know the udev stuff anymore, other than that the rules point to group "audio". In a way, that is a universal problem now on Natty, that one needs to belong to that group. [00:04] In order to get access to the device [00:05] For Ubuntustudio, it's all about belonging to the audio group, then. [00:10] As long as we are on the -lowlatency kernel things are simple. On -rt we might need to adjust rtirtq script to work with firewire. [00:15] ailo, i was just thinking that it would be nice to know what it took for firewire to work, especially six months or a year from now and people don't remember or aren't around anymore [00:16] Sadly, it is Maverick that seems to need the most tweaking. The only problem with firewire could be that the device is not included in the udev rules. [00:16] I mean on Natty, the only problem would be that [00:17] Which, I cannot tell by just looking at the rules. I would think all ffado supported devices are listed there [00:18] ScottL: For Natty it would become a short wiki page. Install Alessio's kernel, become member of audio group, reboot.. [00:20] For realtime we need the kernel. For access to firewire, belonging to audio group, I don't know. [00:31] ailo: cool [00:31] that makes sense [00:31] ScottL: ailo and I were queried today while i went a step and a time [00:32] i tried sudo qjackctl first with the generic kernel and my firewire device with no luck [00:32] i booted the lowlatency kernel and it worked as sudo [00:33] all i did permission-wise was add myself to the audio group [00:33] holstein, i was asking alio if someone can document what it took to get firewire working so if we have to come back later then we know what should be going on [00:33] BUT, we're going to try and find someone to verify [00:34] and we both had an idea for a verifier [00:34] ScottL: i say, after one other person has the same results [00:34] one or more of course* [00:34] with a different firewire device than what i have [00:34] we'll document it [00:34] ScottL: maybe the kernel stuff gets sorted [00:35] and natty can have abogani 's -lowlatency kernel [00:35] holstein, it looks like other issues are standing in the way of getting the -lowlatency kernel into the repos at this time :( [00:36] and in the US-controls, it could be 'click here to add your user to the audio group for firewire [00:36] so i am again doubting if we'll see it for natty [00:36] ScottL: we need that kernel [00:36] period [00:36] holstein, i don't doubt it will be kept out forever, my impression is that there were some other, unrelated issues at hand that was causing several things to be delayed [00:37] if its testing, we got a nice pool of users on it [00:37] holstein: ScottL: I feel satisfied after learning about the udev rules, since it explains how the user gets access to the firewire cards. [00:38] I'm actually editing a wiki page now, but only concerning Ubuntu 11.04 [00:38] ScottL: as long as we can put something somewhere about how to add that PPA [00:38] somewhere in the istallation [00:38] installation* [00:39] ScottL: would i join -motu ? [00:40] and ask? [00:40] or have you already got the axe on it ? [00:44] holstein, i don't think it's testing for anything we are doing, i got the impression it was higher order stuff that effected ubuntu at large [00:45] holstein, BUT you can ask JFo about it, he's the one who mentioned it [00:46] looking at backscroll, JFo said, "it seems that we have so many critical items that are getting delayed (and this one has also been pushed back at least one milestone)" [00:47] right [00:47] i'll ask [00:47] because we need that kernel [00:47] US [00:47] if we are going to compete [00:48] we have the only audio/multimedia distro that doesnt have a RT kernel [00:48] i used to not mind much when we had one in the repo [00:48] and the generic one is getting so much better all the time [00:48] BUT its not there [00:49] we just cant support a large percentage of professionals [00:49] out of the box, OR with a simple install [00:49] AND the -lowlatency kernel + natty seems very nice so far [00:49] for me [00:49] great combo for FW [00:51] that installed by default, and i dont think anyone could complain legitamately [00:53] i feel very strongly about supporting firewire because i believe very many people are using laptops with ubuntu studio [00:53] holstein: I agree. The lowlatency kernel is the perfect replacement for the generic kernel [00:54] if not replacement [00:54] its perfect in this time [00:54] It's a must to get audio working at all [00:54] while the generic one is evolving [00:54] i mean, if this was a couple years ago [00:55] i probably wouldnt even understand how to add a PPA [00:55] or what that means [00:55] 'is it a security risk?' [00:55] i think its a bad call [00:56] especially since its a real 'gotcha' after you've installed it, plugged the firewire device in, and nothing happens [01:11] not to knock what you are saying, but the -generic kernel does me good for my pci delta44 ;) [01:12] right [01:13] but, my device doesnt work at all [01:13] not even poorly [01:13] or at a high-latency [01:13] wont start [01:13] ScottL: But you don't get realtime privilege [01:13] You can see it in the jack log [01:13] and im not trying to say 'wah, my device dont work' [01:14] i know what to do, but lots of folk dont [01:14] they'll just remove ubuntustudio with a bad taste in their mouth [01:15] ailo, when i run qjackctl it has "RT" flashing though [01:15] ScottL: check in the messages though [01:15] i will [01:15] i had a message 'cannot get RT privs' [01:16] but i also get a message to choose to enable rt privelages when jackd is installed [01:16] i forget the exact message [01:16] i'll check later tonight [01:16] in the -lowlatency kernel [01:16] i finally upgraded my music box to the dual core machine :) [01:16] using the alsa driver, not firewire [01:16] well, upgrading is the right word [01:17] i had that message about not getting RT privs [01:17] before adding my user to the audio group [01:17] JACK was running though [01:17] and seemed OK [01:17] holstein, did you start with vanilla ubuntu and add packages? [01:17] i started with ubuntu studio disc [01:17] well, "installed" not started [01:17] ScottL: i had only added jackd at that poing [01:17] point* [01:18] its that test laptop i set up a few weeks ago [01:18] 64bit [01:18] for my install i believe i will already be part of the audio group and i really believe i'm running jack with realtime privs [01:18] but i'll check later tonight as i said [01:18] ScottL: does ubuntustudio add the user to the audio group ? [01:18] by default? [01:18] * ScottL is doing the final update to packages after install [01:18] ScottL: It would be impossible, becuase of cgroups [01:18] holstein, yes, i believe it does [01:19] ailo, this is lucid, not natty [01:19] Ok, sorry :) [01:19] i keep the LTS on my studio box [01:19] hehe, my bad too, didn't mention this before [01:19] ScottL: w00t, that great news on the box :) [01:20] question: if i had installed the 32 bit version on the dual core machine, would it recognize more that 3 gigs of memory? [01:21] i'm guessing that it wouldn't [01:21] another question: if there any benefit (other than memory) to running a 64 bit install? any cons to running 64 bit? [01:23] ScottL: you need a PAE kernel [01:23] if the motherboard supports PAE [01:24] check out.. [01:24] http://www.bandshed.net/AVLinux.html [01:24] nice explaination of the PAE kernel used in that distro [01:24] which, i kinda want to try now [01:24] i was reading about it earlier today [01:24] http://liquorix.net/ [01:25] ScottL: i cant run one thing i want to run [01:25] lightscribe labeler [01:25] in 64bit [01:25] other than that, i havent had any problems [01:26] cant say ive seen any drastic benifits either [01:26] ScottL: One benefit is that build times are a bit faster. Some thing just seem faster. [01:26] my understanding is that wine-asio doesn't work in 64 bit, but that was also a year ago [01:26] probably falktx has an idea about that as well [01:26] falk is up on the 64bit support [01:27] ailo, that is good because i'll be doing some backports soon :) [01:27] holstein: You mean, PAE for 32 bit systems, right? [01:27] ailo: yup [01:27] ive used one [01:27] To get more memory access [01:27] not really for that purpose [01:27] i just got one [01:28] with an ubuntu install [01:28] not sure why [01:28] holstein: I think the only benefit of PAE is that it supports more than 4 GB of ram [01:28] For 32 bit systems [01:30] I think it has to do with adress names. On a 32 bit processor, you have a limit on how many combinations of adresses you can have. [01:30] ive never needed to try it [01:30] but ive heard it works [01:30] So, on the PAE it uses some other way of getting adresses, probably by splitting up the adress into two 32bit parts, or something [01:44] Ok, I added my first ever wiki page on Ubuntu's community documentation. https://help.ubuntu.com/community/Firewire Audio on Ubuntu 11.04 Natty [01:44] hmm, the link is a bit short [01:44] nice, we should link to that... [01:47] ailo, holstein: my understanding is that the -rt kernel was needed to avoid irq conflicts with firewire devices, can this be addressed with the -lowlatency kernel as well? [01:48] ScottL: -lowlatency is the one i was using [01:48] in my tests [01:48] in natty [01:50] ScottL: I don't know if perhaps the new stack has changed that somehow. I suppose, judging on how well it seems to run for holstein, and Asmo, that won't be needed anymore? [01:51] oh, that would be awesome too :) [01:51] holstein: ScottL: I too was under the impression that only -rt gave decent performance in the past. [01:52] Maybe I should not have had spaces in my wiki page name? [01:52] ailo: so far that has been my experience [01:52] i was very impressed when i got down to 1.6ms latency today [01:53] with the -lowlatency kernel [01:53] I'm trying to find if I can move the page. Anyone done that? [01:53] ailo: the wiki page? [01:53] cant move it [01:53] you'll have to make another [01:53] and copy [01:53] Ok. Will do. [01:53] prolly [01:53] dont know who can move them [01:54] none of us i bet [01:56] So we can't delete our own pages either? [01:57] not sure [01:57] I'm not able to now, at least [01:58] I did a renaming, that I didn't like :( [01:59] How about emptying a page, and redirecting? [01:59] maybe [01:59] do we have a firewire one? [01:59] for ubuntustudio? [01:59] we could add to the ubuntustudio preperation one [02:02] Ok, so this is the name of the page https://help.ubuntu.com/community/FirewireNatty [02:03] ailo: maybe it should be audio named? [02:03] not sure [02:04] is this for all firewire devices? [02:04] I'm putting it in audio category now. [02:04] I'm explaining the firewire device support at the bottom of the page a bit. [02:05] cool [02:05] i like it [02:06] holstein: So, this is just reference for an unreleased distro, on our progress. If we want to save it for Ubuntustudio, we could put a link somewhere maybe? [02:08] cool [07:32] * abogani waves all [13:24] morning abogani , ailo , astraljava , AutoStatic , paultag , persia :) [13:24] Good afternoon ScottL! [13:27] Hello ScottL [13:33] hi all! [13:33] ScottL: Do we have some news? [13:36] abogani, about the -lowlatency kernel? not very good news for the moment [13:36] ScottL: :-( [13:37] abogani, JFo said, "it seems that we have so many critical items that are getting delayed (and this one has also been pushed back at least one milestone)" [13:37] abogani, this isn't to say that it will not happen, just not quite yet [13:38] ScottL: So they are quite busy. [13:38] abogani, and i think the UKT was a bit shocked to learn that the -lowlatency kernel requires compile time flag changes, hopefully that facilitates the process :) [13:38] only [13:39] abogani, requires compile time flag changes only i will state it as such from now on [13:39] abogani, but hopefully this knowledge facilitates the process considerably :) [13:40] abogani, and ailo, asmo, and holstein have all had extremely favourable comments about the -lowlatency kernel and firewire :-) [13:41] abogani, so we will definitely keep pushing them to get it into the repositories, even if it is delayed until natty+1 (hopefully it doesn't take that long however) [13:41] ScottL: Good. [13:42] ScottL: What does natty+1 mean? [13:43] After final release? [13:46] ailo: No it means 11.10 [13:47] ok, so UbuntuStudio Natty will most probably need to include the ppa, then. [13:54] abogani, my understanding was that the -rt kernel was required for firewire users avoid irq conflicts [13:54] abogani, does the -lowlatency kernel provide the same benefit? [13:55] Isn't that done using the rtirq script together with the -rt kernel? [14:02] ScottL: The -rt/-realtime kernel don't avoid irqs conflicts: It let us how prioritize irqs execution. As ailo pinpointed rtirq do exactly that job. [14:03] it let use decide how prioritize [14:05] abogani, i'm frightfully ignorant about firewire and irq conflicts, however i trust your knowledge which is why i ask you [14:05] abogani, okay, so if a user is experiencing an irq conflict and they are using the -lowlatency kernel, can they still use rtirq to fix it? [14:09] ScottL: From what I understand the rtirq script only works with the realtime patch. [14:13] Seems to me that ailo is very well informed. :-) [14:19] That's right, rtirq only works with kernels that have the realtime patch [14:20] I don't know so much about the danger of irq conflicts, but I always thought the rtirq script brings down low latency even more. I don't think it helps when there is an irq conflict. [14:21] It doesn't bring down latency [14:21] Sorry, "gives maximum priority" [14:22] It allows you to prioritize softirq's [14:25] And that helps when you have IRQ conflicts like this: 16: 8201 318657 IO-APIC-fasteoi uhci_hcd:usb3, ohci1394, mmc0, eth1, jmb38x_ms:slot0, nvidia [14:25] (that's my notebook) [14:26] I can't use my FireWire soundcard with a generic kernel, only with a realtime kernel [14:27] AutoStatic: How does the rt kernel help avoid conflicts? I'm not sure how to read that, but are there many devices using the same irq? [14:27] Yes [14:27] a USB port, a cardreader, my wifi card and my graphics card [14:28] Ok, so the rtirq picks out one device, and gives it maximum priority within that irq? [14:28] Yes [14:28] I suppose the -lowlatency cannot help there, then. [14:28] I give ohci1394 max priority with the help of rtirq [14:29] And then my FireWire soundcard works [14:29] No, unless you unbind the driver for that USB port, disable WiFi and use Nouveau [14:30] And disable the cardreader [14:30] But I've never tried that [14:34] Autostatic: Have you tried Natty on that system. Would the juju stack make any difference? [14:35] hey guys [14:36] Hello falktx [14:36] ailo: Haven't tried Natty [14:36] AutoStatic: do you have a firewire device? [14:36] falktx: Yes [14:37] cool [14:38] AutoStatic: I always wanted to know how the new stack works without the udev rules thing... [14:38] Can't tell you, I've never tried the new stack [14:39] falktx: We figured it out yesterday, [14:39] ailo: what exact version brings this "fix"? 2.0.1 ? [14:40] Apparently the udev rules are now default and reside in //lib/udev/rules.d/60-ffado.rules [14:40] A part of udev [14:40] And they point to audio group, so one must be a member in order to get access. [14:41] err [14:41] ailo: i though that was the old way... ? [14:41] You still need udev [14:42] falktx: Have you checked that dir /lib/udev/rules.d/* ? [14:42] There a lot of stuff in there [14:42] as the new stack creates nodes too [14:42] and you have to get permissions right on those [14:43] * falktx is still confused [14:44] I don't know anything about udev, but in that folder you can find rules for alsa as well, and pretty much everything else it seems [14:44] yep [14:44] hey falktx, is wine-asio available for 64 bit yet? if i remember correctly, it didn't used to be [14:44] but I already use the udev for firewire in kxstudio... [14:45] ScottL: ubuntu doesn't have wine 64bit, so it's useless [14:45] ScottL: but yes, wineasio can be compiled to work on 64bit, JackWinter confirms it [14:46] falktx: I guess the main difference is just that you don't have to add them. The rules are a part of the standard installation. [14:46] I'm not an udev expert either [14:47] ailo: ah, k [14:47] All I know is that it allows you to set permissions on device nodes [14:47] so the new stack "fix" is a fake [14:47] ehe [14:50] falktx: so, for the ubuntustudiocontrols you would only need to worry about adding the user to audio group, in order to get firewire working [14:50] And installing the -lowlatency kernel, as it seems [14:50] ailo: hm, it doesnt work with generic ? [14:51] falktx: Not heard of it working yet on the generic. [14:51] it doesnt mean it doesnt work [14:51] Don't know if it's because of jackd being in realtime mode by default, and cgroups won't let that happen [14:52] ah, the cgroups stupidity... [14:52] almost forgot about that one... [14:52] In any case, the generic kernel is no good [14:52] k [14:52] does natty has 2.6.33-rt kernel ? [14:53] Nope [14:54] ah, still 31? [14:54] falktx: Do you mean in the official repos? [14:55] abogani: yes [14:55] i havent checked it yet [14:55] falktx: It isn't still exist anymore. I dropped it completely. [14:55] oh [14:56] too bad [14:56] falktx, i don't believe there are *any* -rt kernels in the repos [14:56] just in ppa [14:56] ScottL: lucid has 2.6.31-rt from karmic [14:57] falktx: For same reason (they don't give me upload rights) I don't support that version. [14:57] falktx, i was pretty sure it was dropped, but i could be wrong of course :P [14:57] abogani: it would be nice to have it rt in the official repos [14:57] ... [14:59] ubuntu packages shows linux-rt for hardy, karmic, and lucid...hmmmm [14:59] falktx: https://wiki.ubuntu.com/RealTime [15:00] falktx: Sorry but in last releases I spent only the 15% of my time (which I had dedicated to Ubuntu) for kernel and the rest for Ubuntu's bureaucracy. it isn't acceptable for me. [15:00] abogani: i totally understand [15:00] abogani: but I though you were MOTU ... are you not ? [15:01] falktx: No. [15:02] damn it [15:02] anyone in US dev team is MOTU ? [15:04] falktx, i don't know about juniper.jaxx, but per.sia and the.muso are either core devs or motu, but of course they aren't directly active currently [15:04] falktx: TheMuso, quadrispro, perhaps persia [15:05] abogani, oh yeah, quadrispro....he's core dev i believe [15:05] this isn't to say that per.sia and the.muso don't help when asked because they do :) [15:05] ah, yes, quadispro is from debian [15:06] falktx, he's also a debian developer, so we have a really good vector (along with debian multimedia team) to get packages into debian and therefore into ubuntu [15:06] yep [15:06] * falktx hopes to get ladish into debian soon [15:06] falktx, that would absolutely rock [15:07] falktx, i believe that if ladish gets into debian and then into ubuntu studio and we can do some PR for it [15:07] ScottL: ladish currently depends on laditools, so we need to get that there as well [15:07] falktx, then ladish will be positioned to own the position over jack-session [15:07] that would rock [15:08] ladish would be first and provide more functionality [15:08] i really think some PR and tutorial (video and wiki) and people would ignore jack-session [15:10] What's wrong with jack-session? [15:10] (I don't use session managers myself BTW) [15:11] Why do you want people to ignore it? [15:11] AutoStatic: have you tried it? [15:11] No [15:12] AutoStatic: i did try to develop an app for it, at least the API sucks [15:12] If I'd try it it would become clearer I guess :) [15:12] there's no "open" file or "quit" [15:12] just save, save template and save+quit [15:12] we can save a template but can't open it [15:13] All I know is that Torben Hohn is really enthusiastic about it, guess that blurs my judgement [15:14] Torben Hohn is kinda the progressive guy [15:15] Anyone thought of making startup scripts for qjackctl, that checks whether the system is tuned or not? Opening up maybe studio-controls, if not.. [15:15] Or, A startup script. [15:15] There is the realtimequickconfig script [15:16] AutoStatic, i say that because i believe ladish offers more currently than jack-session is planning to offer from what i have read [15:16] hi quadrispro [15:16] hi ScottL ! [15:17] Yeah ScottL you're probably right [15:17] it appears that you already have to have jack started and ardour open to re-establish connections (which kinda seems redundant since ardour already does this) [15:17] but it's probably something i'm doing wrong or not understanding [15:17] * falktx wishes to make video tutorials soon, but he's afraid of his bad english [15:17] but what falktx said is true, it's very difficult to understand (at least intuitively and how does that fare for a new user) when you are saving or opening [15:18] falktx, i wouldn't worry about your english, there will be many people watching it who either a) aren't native english speakers or b) don't care if you mispronounce a word [15:19] i hope so [15:19] the thing i like about ladish is that it seems to be a single point of startup for all the audio apps, including all connections [15:20] Yeah, i already got the idea too that jack session management is more a dev orientated alternative while Ladish aims at users [15:21] ardour3 now requires jack-session [15:23] Ah, well, I don't use Ardour [15:23] Never will probably [15:23] AutoStatic, really? what do you use and why do you prefer it over ardour? [15:24] * ScottL hopes to learn something cool ;) [15:24] Qtractor [15:24] is that because of MIDI sequencing? [15:24] Yes, that plays a big role, I do a lot with MIDI [15:25] But i also use it to record the band [15:25] that is very understandable, by the way, qtractor should replace seq24 as the default sequencer in natty [15:25] I just like it, it supports all plugin framweorks [15:25] Seq24 is a pattern based sequencer [15:25] Qtractor is not [15:25] falktx, maybe one of the OSMP guys can help you with the ladish tutorial, maybe do voice over is you are still worried about it [15:26] They are completely different applications [15:26] I use them side by side [15:26] Seq24 for my pattern based sequences and Qtractor for overdubs [15:26] AutoStatic: have you tried the Non-* apps? [15:26] Yes. Don't like them [15:26] heh [15:27] They all lack specific things I need [15:27] I'm still looking for a good jack-midi sequencer... [15:27] oh, and AutoStatic, please don't misunderstand my beliefs or intentions: i believe that paul, et al, are doing incredible work and that i don't want to exclude jack-session from ubuntu, ubuntustudio, or anywhere [15:27] No, no I understand [15:27] but i view things from a user perspective and think that ladish offers more for them [15:27] AutoStatic: ScottL: btw, jack-session is in the new PPAs [15:27] just in case [15:28] That's what I wanted to say, you're seeing things from a user perspective and from that perspective Ladish is superior [15:29] AutoStatic: nedko @ #ladi would love to hear that... [15:31] hahaha, i bet he would :) who doesn't like some validation? [15:31] * falktx cheats [15:33] He'll get it ;) [15:33] He has an account on Flattr and I support some of his projects [15:40] falktx: what would you like to see in a JACK-MIDI sequencer? [15:43] AutoStatic: just the normal stuff, using patterns and then put them together on a sequence [15:43] like seq24 kinda does [15:43] But you're not so fond of seq24? [15:44] AutoStatic: I recently developed a plugin for energyXT, which allows to use jack midi input&output [15:44] AutoStatic: seq24 is alsa only afaik [15:44] i wonder who bluebug is? [15:44] hi JFo [15:45] falktx: that's right, I misread your initial post [15:45] I read it as a jack-aware midi sequencer [15:45] hi ScottL [15:45] But you want a MIDI sequencer that has JACK-MIDI support [15:46] AutoStatic: my energyxt work -> http://www.kvraudio.com/forum/viewtopic.php?t=307300 [15:46] AutoStatic: yes, so far only ardour3, non-sequencer and epichord does it [15:46] AutoStatic: now also energyxt too cause I wrote it :) [15:46] And Jacker [15:47] https://bitbucket.org/paniq/jacker/wiki/Home [15:47] Cool stuff [15:47] (The energyXT plugin I mean) [15:48] AutoStatic: is the first way to get midi output support in energyXT [15:48] (in linux) [15:48] I've never used energyXT [15:48] last time I tried jacker crashed on start [15:49] JFo, you have any idea how the Graner's are? [15:51] well they are both wacko if that is what you mean ;-) just kidding [15:51] AutoStatic: fun fact... energyXT does not send midi_off events (0x80), it just send a new note with 0 volume [15:51] ScottL, they are good. [15:51] AutoStatic: i laughed at it when I notice it [15:51] The new house is coming along nicely [15:51] yeah, thatÅ› pretty clumsy [15:52] * AutoStatic that's [15:52] JFo, sweet! i can only imagine how they felt, just devastated and uprooted [15:52] AutoStatic: they store all midi data in an Int, which was hard to work with [15:52] yeah, it was a tough time [15:52] but things are slowly getting back to normal [16:10] AutoStatic: do you know any other jack-midi sequencer? [16:15] morning ScottL [16:15] falktx: Nope [16:16] AutoStatic: my reason for the jack-midi interest is the jack-freewheel support [16:17] AutoStatic: if an app support jack properly (audio, midi and transport), you can render an entire project in freewheel mode, without a single xrun no matter how many synths/effects you enable [16:18] Sounds cool, but there are not a lot of apps that support all that [16:18] I guess [16:19] AutoStatic: yep, that is a sad thing [16:20] What do you mean with rendering actually? [16:20] I talked to Rui about jack.midi in qtractor, but he said not possible at this time [16:20] No, Qtractor relies heavily on Alsa midi [16:20] AutoStatic: hehe, a tool I developed specially for this [16:20] AutoStatic: imagine this [16:21] hydrogen open in song mode, synced with qtractor (audio) [16:21] some effects laying around [16:21] AutoStatic: you set a start time (usually 0), and a stop/end time (3mins by default) [16:21] AutoStatic: my app will start jack_capture, set freewheel mode, and start transport [16:22] AutoStatic: when the transport reaches end-time, transport and recording stops [16:23] AutoStatic: it renders a whole project with no crackles, and mot of time faster than realtime [16:23] *most [16:25] Sounds a lot like what monolithic apps are doing on other platforms [16:26] AutoStatic: yep, but we can get the jack modular setup that way [16:27] using realtime stuff in a laptop doesnt quite work for me [16:27] I don't see the relation between rendering and using realtime stuff [16:28] and qtractor cannot render midi synths to audio directly, so I think this is the best idea [16:28] AutoStatic: ^ [16:28] AutoStatic: I cannot use many synths at the same time without getting xruns [16:33] Well, this is all a bit above my head ;) [16:33] (from my own perspective) === falktx_ is now known as falktx [17:30] hi troy_s, how have you been? [17:30] ScottL: LOL [17:31] ScottL: Apparently you missed the cross posting event. [17:31] ScottL: Been uh... interesting here the past two days. [17:32] ScottL: You? [17:49] maybe im unclear as to what the time issue can be with getting the -lowlatency kernel in the repo [17:49] is the plan to make the generic one work ? [17:50] if thats the case, thats fine [17:53] troy_s, i've been busy, which cross posting? [17:53] LOL [17:54] ScottL: About 35,000 views. Lots of vitriol. Lots of anger. Lots of tangential opinion. Plenty of drama. http://news.ycombinator.com/item?id=2091194 http://lwn.net/Articles/422414/ [17:59] troy_s, sorry was busy getting an answer in another channel [17:59] ScottL: Don't be sorry. ;) [17:59] troy_s, oh yes, i had expected this to make quite a stir [17:59] ScottL: Yeah. It did. [17:59] troy_s, i know you had been critical of this before and was surprised it took this long to make the post, but i'm glad you did [18:00] troy_s, sadly i haven't posted in a long time :( i'm hoping to start up again in a month or so. and keep it going again [18:00] holstein: About the cgroups on generic kernel, all I know is that some jack devs are working on it. [18:02] holstein, initially the blueprint was to get the documentation completed for natty so others (us included) can get other kernels into the repos [18:03] holstein, but JFo was saying that other critical items are stacking up which is delaying other things (our concerns as well, i.e. -lowlatency kernel) [18:04] holstein, so once the "log jam" with the UKT settles then we can expect some movement on the remaining items, including for us :) [18:04] holstein, so, it might be in time for natty, but i'm expecting not, but this means that i expect it for natty+1 hopefully [18:04] yeah, there are tons of things that are getting shifted [18:04] much to the dismay of the team itself [18:04] ScottL: im just wondering how we would get permission to have that kernel installed by default [18:05] holstein, "by default" do you into the repos or onto the ISO/disc? [18:05] well, ideally in the repo [18:06] but if thats not going to happen [18:06] ScottL: Could it be possible to add a repo on top of the Ubuntu one? [18:06] can we get premission to have a non-official repo item inculded? [18:06] included* [18:06] im just following this RT in maverick thread in the mailing list [18:07] i've sent several messages with a nice link [18:07] and still, there are questions [18:07] valid ones [18:07] about how to get an RT kernel [18:08] holstein: well, isn't that mainly because of Maverick? [18:08] holstein, well, i don't think it's a matter of getting "permission" as it is logistics, i.e. _how_ would one include the repo with the installation [18:08] holstein, which would probably be modifying a package to add that line to the sources.list [18:10] holstein, but this doesn't really get the item (or kernel) installed on the image, just provides a path to get it [18:10] holstein, but ultimately, with the image build system that exists, i don't think it's possible to include something not in the repos [18:15] i think other distros (avlinux, kxstudio) build their images from an extant system, meaning they either have a working install that they make an image from (ala remastersys) or they build it in a chroot [18:16] i think falktx used to use remastersys but moved to building their own image via script [18:16] sorry, phone.. [18:16] essentially i believe that falktx has automated this process:https://help.ubuntu.com/community/LiveCDCustomization [18:17] no problem holstein [18:17] but i do not know if he build his install in a chroot as they wiki page describes [18:17] ScottL: i would be fine with abogani 's ppa added by default [18:17] and documentation [18:18] how do we make that happen for natty? [18:18] what can i do* to make that happen? [18:18] BUT the automatic build deamons (buildd) start with "seeds" names of packages desired and it "germinates" which means it finds all the depencies and somehow makes an ISO image that works [18:18] that last part is beyond me though [18:19] well, somebody knows [18:19] holstein, i think we need to have a meeting soon and determine what changes we would like to make to any of the ubuntustudio-* packages [18:19] and it cant be rocket science [18:19] thats something i could find out [18:19] falk would know that i bet [18:19] ScottL: agreed [18:19] ScottL: What's the deal with existing Ubuntustudio packages? I don't know the procedure. Can they all be updated before final release? [18:19] i just didnt want to miss the opportunity to get -lowlatency in natty [18:20] but the point of the buildd system is that it is automated and all based on what is in the repos, to do otherwise would probably mean rewriting the buildd system [18:20] at least an apt-get away [18:20] ailo, yes [18:20] ailo, but there is a limit though [18:20] ScottL: So we could in a sence add some scripts and default settings to make it really easy to setup everything? [18:21] ailo, according to the schedule, that time frame would be Feb 24th: https://wiki.ubuntu.com/NattyReleaseSchedule [18:21] ailo, possibly, but we need to determine the changes, make sure they're sane, modify the package, update the source, and Bob's your uncle [18:21] if nothing goes wrong somewhere ;) [18:22] At least get Ubuntustudio-controls updated and include kernel install to it, I think. [18:23] keep in mind, not all the proposed or intended changes for -controls needs to happen now, we can stage it between various releases if necessary [18:23] but, i could get on adding abogani 's PPA to the sources right? [18:23] Sure [18:23] that doesnt really change anything functionality wise [18:24] out of the box [18:24] i think adding the a ppa to the sources.list would be better suited for the ubuntustudio-settings package thought, not in -controls [18:24] IF we are going to add it automatically and not by user request (by ticking a box) [18:25] holstein, agreed, the user would need to apt-get it, but at least it would show up in synaptic or apt-get cache [18:25] If we add a script to jackd startup configuration that checks if kernel is installed, if user belongs to audio group, and stuff like that, this could pop up the studio controls, that would let you set everything up. [18:25] if "apt-get cache" is the right apt-get command, don't know if this is correct [18:25] ScottL: im comfotable with that [18:25] comfortable [18:26] sudo apt-cache search linux-lowlatency [18:26] and something would come up [18:27] ailo, i'm not sure that is the correct approach [18:27] is this going to happen everytime jack is started? [18:27] what if the user _doesn't_ want -lowlatency? [18:27] It would only pop up, if something was missing [18:28] On Natty? On Ubuntustudio? We can't use generic, in my opinion. [18:28] ailo, right, but what if the user is comfortable with -generic, will it pop up everytime because -lowlatency is "missing" ? [18:28] oh yes, i keep forgetting about the cgroups [18:28] well, the plan is that the -generic will be what we use though [18:28] * ScottL is going to install vanilla natty today and play with it [18:28] in the hopefully not to distant future [18:28] Consider the -lowlatency a generic kernel, because it pretty much is [18:30] i tried installing yesterday's ubuntustudio ISO and the kernel keeps panicing during installation (but at least it's not terminating with sigterm anymore) [18:30] :/ [18:30] i did vanilla [18:30] and added jack to test [18:30] added the meta packages yesterday [18:30] ailo, but the -lowlatency kernel doesn't have rt_group_sched, correct? and this prevents cgroups from being a problem correct? [18:31] ScottL: Exactly [18:31] holstein, one of the meta packages failed to buld last night though, i'll have to choose carefully which i install, but i think it was the audio meta, so i should be okay [18:31] And, the -lowlatency kernel has everything the -generic has, except it's better at lowlatency. [18:31] i just did ubuntustudio-audio and ubuntustudio-audio-plugins [18:31] i think -video failed [18:35] ScottL: There could be many strategies for a startup script for qjackctl. One strategy could be using a onetime, did you set everything up? [18:37] Also, there would perhaps need to be a dependency to the Ubuntustudio-controls, no matter which audio-related ubuntustudio-package one installs [18:38] The problem is mainly related to jack. Realtime as well as firewire. And both seem to need -lowlatency and the audio group. [18:42] I can understand holstein's concerns on not having things "out of the box". For someone not realizing they need -lowlatency may get really dissapointed. [18:47] (In a way, jackd being in the main repo without a functional kernel for realtime is a pretty serious flaw. ) [19:06] ailo, are you suggesting we modify the jack package? [19:07] i think having access to a RT kernel out of the box is fine [19:12] ScottL: I don't know about jackd. That's still in the main repo and should of course at some point be adjusted to work with new Ubuntu kernels. [19:13] But, Ubuntustudio-settings could perhaps include something that makes it easier to get things setup. [19:14] The simplest approach could be doc on the Desktop, of course. [19:15] The best thing, of course would be to get the -lowlatency in. [19:16] If we were to adress the importance a bit, perhaps we can make them upgrade the importance of the -lowlatency kernel? [19:16] And things would go faster. [19:29] ScottL: By the way, those workflow ideas for audio I guess would be perfect i used as templates on something ladish, eventually. [19:30] (time to sleep..) [22:34] ailo, holstein: okay, finally got a vanilla natty installed and you are correct, sir; i can't get real-time privileges [22:34] ailo, holstein: the audio group was already there and i added myself to it, but still no real-time privilages [22:34] ailo, holstein: and i get horrible stable latency with -generic kernel [22:35] next, i move onto abogani's -lowlatency kernel :) [22:56] ScottL: cool [22:56] glad you're finding the same results [22:56] at least [22:57] * holstein dinner... bbl