[08:11] hi [08:14] can someone tell me if it is possible to backport soundconverter in 12.10, maybe 12.04 ? Bug #1077508 [08:14] Launchpad bug 1077508 in soundconverter (Ubuntu) "2.0.1 is bugged, update to 2.0.4" [Undecided,Confirmed] https://launchpad.net/bugs/1077508 [08:15] ttoine, yes.... [08:16] ttoine, use requestbackport command in terminal (install ubuntu-dev-tools first) [08:20] smartboyhw, is it not possible to use launchpad ? It is fixed in Debian, but not in Ubuntu... [08:28] ttoine: The procedure is to use the command tools (it's integrated with launchpad) [08:28] ttoine: Let's add a workitem for it. We need to start a routine for backporting (that workitem is already present) [08:30] I'll add a new blueprint for backporting [08:30] First, breakfest [08:32] zequence, what's on breakfast today?;P [08:57] smartboyhw: Bread an cottage cheese (forgot my avocado at home) [08:57] zequence, oh;P [09:10] ttoine: I feel the whiteboard on public relations covers just about everything. Is there something there you feel is missing? https://blueprints.launchpad.net/ubuntustudio/+spec/ubuntustudio-public-relations [09:11] I think it might be good to wrap up the technical side of PR within the next couple of weeks, and start posting actively [09:16] Scott has made a good job of skething things out [09:16] zequence, you are right, it covers almost everything [09:17] zequence, sorry, I am a bit overloaded those days. I had my workshop sunday about recording with Ardour [09:17] and was at the JDLL.org all the week-end [09:18] will be better tomorrow, if I can find some time to sleep. I think I will try to make an illustrated doc on "recording music with Ubuntu Studio and Ardour" [09:20] I meet a guy using CentOs and Planet CCRMA to make music. He did a workshop. Nobody understand anything. Too complicated... And with Ubuntu Studio, all worked out of the box. [09:23] ttoine: :) [09:25] ttoine, good [09:31] ttoine: I updated this page with some info now https://wiki.ubuntu.com/UbuntuStudio/PublicRelations [09:32] It's just an edited version of the whiteboard [09:39] zequence, ok. I have to go out of office. See you later. [09:40] cherio === smartboyhw_ is now known as smartboyhw [11:44] zequence (or any other guy in the dev team) please approve https://code.launchpad.net/~smartboyhw/ubuntustudio-default-settings/UbuntuStudio/+merge/135123 (you can't not merge it, 0.40 is up into raring-proposed already... [12:55] hop [12:55] ttoine, hop what?:P [13:11] Beer hop, naturally. [13:13] * astraljava has this t-shirt: http://bit.ly/USOZjV [13:21] astraljava, nice [13:21] smartboyhw, hop, here again [14:15] smartboyhw, micahg (or whoever releases) does that not us. [14:16] We have been told to leave things as UNRELEASED [14:16] len-dt, er actually I released that and dholbach sponsored it so... [14:16] I am testing the 310 nvidia beta driver. It's fast. Amazingly fast !!! [14:17] ttoine, :d [14:18] from the startup to the login screen, and then switching between apps, etc... all is fast. I never seen that with my thinkpad before [14:18] smartboyhw, why? [14:19] Has there even been any changes made to that source? [14:20] zequence, some minor menu mods [14:20] smartboyhw: So you see, uploading had no purpose [14:20] Two of them, both mine. [14:21] It seems there is now an unofficial version of settings in the wild... [14:23] len-dt: What do you mean? [14:25] if smartboyhw has had someone release something from his part of things as if it came from here... [14:25] released where? [14:26] * len-dt is confused as to what happened. [14:28] smartboyhw, anyway your merge is something we were told not to do. [14:29] I had a one on one conversation with smartboyhw about this. I'm hoping he will ask before acting on issues like this in the future [14:29] Asking is always welcome [14:29] In that case I will say no more and leave settings as is. [14:39] zequence, I have writen a small testimonial on your wiki page. good luck for membership [14:40] ttoine: Thanks. Yea, tomorrow is the day, so let's see how it goes :) [14:41] zequence, you on the 12:00 or 22:00 membership board? [14:42] 12.00. I'm +1, so 22.00 would be a bit too late for me [14:44] Time to go home (zzz..) [14:44] zequence, good I am gonna see the progress. Do they accept live testimonials?:P [14:46] smartboyhw, it is writen that testimonials written on the wiki page are better for people who can't attend to the membership [14:47] ttoine, so that's why I am going to do a "live testimonial":P [14:56] http://www.blackmagicdesign.com/products/decklink/ [14:56] linux driver for pro video hardware [14:57] it is aimed at gstreamer acceleration. I a trying to talk with pitivi devel to see if they could handle that [15:00] I am trying [15:14] zequence: you would think.. but then theres the "ralph" factor [15:15] holstein: First of all, I haven't recommended anyone to use an older lowlatency kernel [15:15] Second, what about ralph? [15:15] * smartboyhw wonders what is ralph [15:15] zequence: you can... im just talking about a user such as ralph from the list [15:16] ralph might be the kind of user who would expect US to support an older kernel [15:16] and should we? [15:17] it might just work fine... i agree that if the user typically knows how to switch kernels, they probably know how to deal with the consequences [15:18] I've considered this. Since newer kernels have been a bit poor, it might have been a plus to use an older kernel, such as 2.6.39, which was the first to support threadirqs, but also very fast [15:18] It's probably not possible to do in the main repo though [15:18] And I think this is a special case, which only a few people will want [15:18] the RT kernel from 9.10 works great [15:19] Well, so does -lowlatency 2.6.37 [15:19] still, like you say few people want it.. and i argue even less need it [15:19] they see the term "realtime" thrown around, and want it... and thats fine [15:20] I haven't done enough live processing with the newer kernels to verify if they are able to give me what I need. I'm not very sure they can [15:20] i say, from a marketing perspective, we should be "the first audio distro providing excellent latency without the need for a realtime kernel" [15:20] And, if that is so, the group of people who would "need" a faster kernel, would be fairly large [15:21] holstein: But, it's not all together true, since -lowlatency is only possible thanks to the realtime patch, which now to a great extent lives in the vanilla kernel [15:21] so, whats not true? [15:22] What I'd like to say is -lowlatency is an officially supported kernel, while -rt is experimental [15:22] sure, but there is no "rt" [15:22] you can make one.. or try the ones from ppa.. but we dont have an rt one.. not even experimental [15:23] holstein: That -lowlatency is not a realtime kernel. Cause, it is, at least to the extent that it matters to the user (since as I said, the vanilla kernel nowadays includes much of the realtime patch) [15:23] It's not hard realtime [15:23] sure.. but if you want lowlatency in a default stock ubuntu.. thats what we have [15:23] But, it's realtime [15:23] and thats what we could talk about.. marketing-wise [15:24] Problem now is that neither a -rt kernel, or a -lowlatency (Debian has a -realtime kernel in their repo) is as good as they used to be [15:24] they didnt for a while though [15:25] and i still say, users typically dont need it.. they just want it, which is fine [15:25] Anyone doing live processing need it [15:25] Something as simple as playing a live synth [15:25] sure.. but who are they? most folks have internal cards.. do podcasting maybe [15:26] not a lot of folks do either effects nor synths [15:26] I don't know what makes you think that, but in my experience, people who make music quite often do live processing in one way or another [15:26] it would be nice to have something to offer them though, since 3 years ago we did [15:27] zequence: i find most users think they need it.. but really dont even have the harware to support it [15:27] We are offering it to them, with -lowlatency [15:27] zequence: sure.. but its not as good as rt from 9.10 [15:27] I'm telling you, -lowlatency is if not exactly as good, then about as good as -rt [15:27] not nearly as good for my firewire [15:27] You're talking about kernel versions [15:27] Not, -rt vs -lowlatency [15:27] The -realtime in Debian repo sux as well [15:27] im talking about what came in the repos for 9.10 vs what is in the repo in 12.10 [15:27] Why? Because of the kernel version [15:28] im talking about. i install and dont know what a kernel is/does [15:28] i can have as good performance as i had in 9.10 [15:28] Add a realtime patch to 3.2, and it will suck just as much as -lowlatency [15:28] why is that? i might not care.. i might just move on. but i likely dont need it anyways [15:29] Because, as I said before, something in the vanilla kernel made it less responsive. And both -rt and -lowlatency suffers from that [15:29] So, what would you do, if you want to have low latency on 12.04, if neither -rt or -lowlatency is good enough? [15:29] Perhaps use an older kernel? [15:29] i would like to change the idea.. marketing wise [15:29] we are not going to have an rt kenrel in the repo [15:29] and the kernel is what it is [15:30] so, we could just try and state the facts.. [15:31] If we could get some hard facts down, from testing -lowlatency vs -rt, we could clearly state how good/bad -lowlatency is. N [15:31] try and get folks to just try the software, and not really think/worry so much about getting sub 1 ms latency [15:31] zequence: i think that would be helpful [15:31] zequence: especially for some commom hardware types [15:31] zequence: you think for 12.10? 12.04? both? [15:31] I don't think most people are concerned with the 1 ms bit, just that they don't get xruns when using their OS for making music [15:31] zequence: i think we could get a nice cross section of harware represented [15:32] Someone would need to build -rt kernels, and add them to a PPA for testing purposes [15:33] If the kernel 12.04 would have been better performing (and again, not the fault of -lowlatency, but the kernel version itself), then I would have felt better about promoting it [15:33] or, just state the facts about what is default [15:33] I'm not sure those kernels are adequate for live processing [15:33] Which sux, bigtime [15:33] "im using default 12.04, stock kernel with x hardware, performance is x" [15:33] ^^ for example [15:34] maybe not a comparison so much as a statement [15:34] YEa, well. We did do those tests between -generic, and -lowlatency, if you remember. I even compared to -rt on 9.10. And at the time, the result was clear. -lowlatency kicked ass [15:34] But, the problem again, is not -lowlatency. It's the kernel version [15:35] yeah... well, hopefully its being addressed [15:35] not much we can do about it at this point.. with the lts.. [15:35] I talked with someone in the kernel team about it in UDS, and they said they had found something weird, but no one knew what it was [15:35] to "fix" it.. i mean [15:35] zequence: great.... [15:36] zequence: i wish we had a representative up there.. someone with our interests in mind... [15:36] This is why I'm not a big fan of LTS for Ubuntu Studio. 9.10 was our LTS [15:36] holstein, if that so it will be zequence :P [15:37] Actually, let me try jack quickly on 13.04.. [15:37] zequence: i should be doing that too.. and can [15:39] It's not good enough [15:39] I mean, the kernel on here [15:39] Well, maybe. [15:39] i could load up the 64bit 13.04 on my main rig and test with firewire when i get my device back over here [15:41] Ok, so this is ok. I don't seem to get xruns at 64 f/p, that's about 2.7ms latency (in jack, but it's not the real latency) [15:42] That's my acceptable limit. 128 f/p works, sort of, but it's a little too much [15:42] If 32 f/p worked, that would be a great safe margin [15:42] I'd need to have the system running for a while to catch any random xruns [15:42] That's the main problem I saw with 12.04 and 12.10. [15:42] Yep, I just got one [15:42] thats how i feel about it.. acceptable [15:43] not rip-roaring like 9.10/10.04 was... [15:43] So, this is probably not good enough for me then [15:43] i have a few devices though... firewire and a few usb's [15:43] And the 3.7 kernel is no better than those on 12.04 and 12.10 [15:44] jta: Hey man [15:44] is anybody way upstream with this issue? [15:44] zequence, jta holstein bye [15:44] holstein, what do ya mean? [15:44] i mean, this is not ubuntustudio specific, correct? [15:44] all newer kernels suck, correct? [15:45] holstein: No. This has only to do with the vanilla kernel source [15:45] zequence: wonder if we can just go up there and sort it out? then it'll trickle in to everyone [15:45] zequence: do we know anyone up there? [15:45] holstein: I'm not sure anyone even knows why [15:46] holstein, zequence the problem is that you can't get official -rt patches for 3.7... They only release for like 2.6, 3.0, 3.2, 3.4, 3.6, 3.8 and so on... [15:46] and its challenging since its not a goal up there [15:46] The linux kernel is from what I understand the largest software project in the world [15:47] So, not easy to keep track of all the changes in the code [15:47] But, I'm sure someone knows something [15:48] It's not something we can do, anyway (until one of us becomes a kernel hacker) [15:49] maybe i'll ask casually in #ardour [15:49] those guys are sharp [15:50] or, if we could get someone like jono using the software and asking (complaining) about it [15:50] There are a couple of guys who work on kernel code, and I believe one or two of them might be in there [15:51] jta: So, did you have a launchpad account [15:52] jta: We're using launchpad for most of the things we do now, so you'd really need to get into speed with that, at least so that we know your lp username [15:53] zequence: not yet, I have been busy wrapping up my event season... [15:53] I will float it to the top of my list so it gets done sooner zequence [15:54] zequence: ok done, it's in the top section of my todo list [15:54] jta: Ah, great. It's just very convenient when doing planning [15:54] And you'll be able to see changes that we do also, and be up to speed with what's happening [15:58] zequence: cool, thanks for the prompt [16:42] zequence, mind if i use your first name on your wikipage? [16:42] knome: Of course not. :) [16:42] zequence, great, thanks [16:51] zequence, i've added a testimony for you [16:53] knome: Thanks a bunch :) [16:54] np [17:41] Actually, I'm going to change my opinion about -realtime vs -lowlatency now [17:41] It's just that -lowlatency 2.6.37 was a great combination, but on later kernels, -rt is still very much ok [17:46] holstein: I just did some testing with -rt on Wheezy. It's 3.2 series. [17:46] It's quite ok, actually [17:47] I'd need to run it for longer periods, to make sure [17:47] I did have some problems with it in the past, but not sure if it was my own fault === jta is now known as jta_afk [18:36] zequence: i was going to look to be sure but i think thats where AVlinux stopped [18:36] 3.2.x [18:37] There's no newer kernel for Debian, in fact [18:37] they build custom ones [18:37] he has a kernel guy.. and he does the rest [18:38] the performance so far is quite nice [18:38] I think their kernel is 3.0.36 [18:38] its not installed on my main machine, so i dont have a good analog to it [18:39] I'm building 2.6.38 right now, to do some tests. I was just at #lad talking about kernels, and we started doing some cyclic tests [18:39] COOL [18:39] thats what we need.. data [18:39] and a kernel dev [18:39] Well, we still can't use -rt, unless we use an older kernel for some of the releases [18:40] Since the realtime patch is not released for every kernel version [18:40] But, if -lowlatency does not improve, I might find the motivation to see about making that happen [18:40] But not sooner than 14.04 [18:40] well, a ppa would be fine, if folks want/need it [18:41] i was hoping by 14.04 we would all be using a generic kernel for everything [18:42] while -lowlatency is at best useful, -generic is not even nearly [18:42] And I don't think that situation will improve easily [18:42] Cause no one is really asking for a kernel like that [18:42] Except for some hippie musicians [18:42] yeah... its not in the main stream desktop need [18:42] servers i suppose dont need it either [18:43] Servers need the contrary [18:43] i tried the liquorix kernel [18:43] That's probably why we're not getting what we need [18:43] servers are optimized for throughput, which is something realtime sacrifices [18:47] holstein: This may answer some of it http://en.wikipedia.org/wiki/Criticism_of_Linux#Con_Kolivas.27_criticism_of_kernel_politics [18:47] Well, maybe not :) [18:48] But, linux on such platforms as Debian/Ubuntu are definately more used for servers than desktops [18:48] And the business is more tuned towards that [18:53] which is fine, if we could get someone like AVlinux's kernel guy to help us... get him up as a motu for a specific kernel [19:00] holstein: I think I'm more than capable of maintaing a kernel, so that's not the issue [19:00] But, coding is another thing all together [19:03] zequence: i think it would be great if it were you [19:04] I'm pretty happy with -lowlatency right now, since with very little effor, we get a pretty nice kernel that can do most people well [19:04] But, if things don't change in the future, I will probably want to see a possible addition of a -rt kernel [19:05] yup.. its a good place to be [19:12] heh, I need to start using bigger partitions for my installs [19:12] 40GB, with /home on a separate partition (mostly), I almost ran out of space just now [19:14] Ah, of course. It's the Linux Steam game [19:31] holstein: So, I booted into the 2.6.38 kernel, and did the cyclic test [19:32] holstein: There's one thing that kind of determines how well it performs [19:33] On -lowlatency 3.7 I got max:~1200, on -rt 3.2:~130, and again, the -lowlatency 2.6.38: ~473 [19:33] On -generic it would be something like ~4000 [19:33] So, 2.6.38 is so damn good, you don't even need more [19:34] But, for some reason there has been a regression since then [19:34] And the current lowlatency just doesn't match up [19:35] Well, the 2.6.38 -generic might be better too of course [19:35] Better than current -generics, that is [19:43] holstein: :(. Actually, 2.6.38 on this release was not a big success, when I had it on for a bit longer. [19:43] It's still better than the current kernel, but not by as much as I had hoped [19:43] The difference is not big enough to be important [19:48] 2.6.38 was part of natty (11.04) which is no longer supported [19:50] micahg: Yea, I'm just testing performance. I === jta_afk is now known as jta