| len-1304 | zequence, just reading about isolated cores (The "isolcpus" kernel parameter) for setting aside a cpu core for a task or thread. Then a thread can be assigned to a core set aside for it. | 04:19 |
|---|---|---|
| len-1304 | I am also reading that IRQs can be bound to a core as well. | 04:20 |
| len-1304 | I don't know how easily certain kinds of tasks could be isolated. Run most of the OS on core 0 and the real time stuff on core 1. | 04:23 |
| len-1304 | I know jack has 3 threads, of which one needs rt and higher priority. I wonder how well other audio app isolate threads. | 04:25 |
| len-1304 | It seems almost all of the processor chips these days have at least two cores... even the lowly atom. | 04:27 |
| len-1304 | Lump|AFK, Asterisk seems to have a jack driver. | 07:25 |
| zequence | len-1304: The interaction between jack and PA is not flawless for me. Initial start of jack worked, but retrying it didn't work very well | 09:06 |
| zequence | PA froze | 09:07 |
| zequence | needed to kill it | 09:07 |
| zequence | Actually, this was only when having PA use jack sink | 09:08 |
| zequence | This worked, but it's still buggy. Start rhythmbox. Start jack. Set PA to use jack sink and source. Stop jack (PA sets to use your onboard card). Start jack (PA is still using onboard, but Rhythmbox uses jack) | 09:11 |
| zequence | Different applications in PA are able to use different interfaces | 09:12 |
| zequence | I think I'd like to have the PA module automatically change to jack when starting it, and change back to whatever was default when stopping it | 09:13 |
| zequence | len-1304: I'm going to make a patch for the module. Feature freeze is coming up March 7th. We should think about what else we want to have uploaded until then | 09:28 |
| zequence | I want it to set jack by default | 09:29 |
| zequence | With the fixes to both jack and PA, I think this release will be tons better just because of that | 09:29 |
| zequence | Ah, the uibiquity plugins | 09:37 |
| zequence | or plugin | 09:37 |
| zequence | I want that to happen too | 09:37 |
| zequence | But, I won't have time until Fri, Sat | 09:37 |
| smartboyhw | Hey scott-work | 12:04 |
| smartboyhw | zequence, ^ | 12:04 |
| scott-work | hi smartboyhw | 12:04 |
| smartboyhw | scott-work, zequence and I are planning for a seperate release team for Ubuntu Studio... | 12:04 |
| zequence | scott-work: Yes, it seemed like the best way to do it, since smartboyhw is managing that pretty well. What do you think? | 12:05 |
| scott-work | can you tell me more about it? | 12:05 |
| zequence | scott-work: Just that we make a team called -release which would instead of -core get the privilege to do the relase stuff | 12:05 |
| scott-work | zequence: that sounds like a good plan | 12:06 |
| zequence | and add smartboyhw to that team. He's the most active on that side of development, and is probably the best person to oversee that | 12:07 |
| zequence | scott-work: smartboyhw started working on a wiki page for release procedure https://wiki.ubuntu.com/UbuntuStudio/ReleaseProcedure | 12:07 |
| smartboyhw | zequence, I was copying the link too:P | 12:08 |
| smartboyhw | scott-work, so can zequence start creating the team? | 12:13 |
| zequence | The team is created, and I asked for the privileges to be moved over to that team | 12:36 |
| zequence | smartboyhw: Again, great work on the wiki. Would you mind if I did some editing just to make it easier to read? I might do that at a later date, if you don't beat me to it. | 12:38 |
| zequence | Just me being pedantic | 12:38 |
| zequence | Not a priority. There is plenty of stuff to improve on the wikis we have | 12:38 |
| smartboyhw | zequence, no worries:) | 12:51 |
| zequence | len-1304: Another difference in graphic cards could be not only which driver, but what type, AGP, PCI or PCIe. | 13:17 |
| scott-work | smartboyhw: zequence: sorry, busy at work. creating the team sounds like the right thing to do | 13:22 |
| scott-work | zequence: what do you need me to do to get the privliges moved to that team? | 13:22 |
| smartboyhw | scott-work, zequence has already made the team and requested stgraber in #ubuntu-release to change things | 13:22 |
| smartboyhw | scott-work, nothing needed for you to do for us:) | 13:23 |
| scott-work | ah, good, good. and i presume that smartboyhw is already added to the team, no? | 13:23 |
| smartboyhw | scott-work, yes | 13:23 |
| smartboyhw | I am added | 13:23 |
| scott-work | oooooh, good work on the wiki page, smartboyhw . that's looks really nice :) | 13:24 |
| smartboyhw | scott-work, :) | 13:24 |
| len-1304 | zequence, assuming you are talking about PA 3.*, I have found that if there is only one audio interface turned on in PA that jack becomes default when it takes the device from PA. | 14:12 |
| zequence | len-1304: Actually, I find that applications started after jackdbus is running will automatically use jack. Even when PA is not set to use jack | 14:13 |
| len-1304 | For me even those started before | 14:13 |
| len-1304 | I have audio going through pa to the IF, start jack, the audio pauses then goes through jack. Stop jack it pauses then goes back to the IF | 14:14 |
| zequence | The PA master volume doesn't have any effect on those applications that use jackdbus when PA is set to use another interface | 14:15 |
| len-1304 | I hadn't tried that. That is a bug. If you report I will try it later and confirm. | 14:16 |
| zequence | len-1304: It's not really a bug | 14:17 |
| zequence | Since PA is set to use another IF, the master volume only affects streams going to that IF | 14:18 |
| zequence | In pavucontrol you can actually set each device to use it's own interface | 14:18 |
| zequence | This you can't do with the regular PA settings Gui which you get with Unity and Gnome3 | 14:18 |
| smartboyhw | yeah | 14:20 |
| len-1304 | The master volume should affect whatever stream the audio gets sent to. I think the PA people would call that a bug | 14:20 |
| zequence | len-1304: I heard from David that they are planning some changes on that whole thing | 14:21 |
| len-1304 | But I am not sure if that is PA or some other part of the system. I will play later. | 14:21 |
| zequence | len-1304: Consider that the master volume makes settings for a specific card. | 14:21 |
| zequence | Not for all of them | 14:21 |
| len-1304 | There are still some changes in git that are not in the current release. | 14:21 |
| zequence | like what? | 14:21 |
| len-1304 | being able to set the bridge channel number. | 14:22 |
| zequence | It's there, but I haven't got it to work | 14:22 |
| len-1304 | It's in the docs but not release. | 14:22 |
| len-1304 | It is in the source/sink but not in the jackdbus detect. | 14:23 |
| zequence | You're right | 14:24 |
| len-1304 | I asked them about it, that was what I was told. | 14:24 |
| zequence | It's in the git tree that Ubuntu is using as a source for the package, so if they do an update before feature freeze, it'll get included | 14:24 |
| len-1304 | you can use pactl to find out what the current modules can do. | 14:25 |
| zequence | or rather, it's a bzr branch | 14:26 |
| len-1304 | Anyway when it does get into release, I would suggest setting channels to two as default. | 14:27 |
| zequence | Yep | 14:28 |
| len-1304 | It would be what people would expect, what most people would need, and use a lot less resources. | 14:28 |
| zequence | actually, we might need to request a patch | 14:29 |
| zequence | Don't think they will update to anything else than official releases, unless they're fixing a bug | 14:29 |
| zequence | I'll make a patch and ask David to add it | 14:30 |
| len-1304 | k | 14:30 |
| zequence | I'll also make one to set the channes to two for the module | 14:30 |
| smartboyhw | zequence, len-1304 try this v | 14:41 |
| smartboyhw | http://rafalcieslak.wordpress.com/2013/02/10/vmodsynth-1-0-released/ | 14:41 |
| smartboyhw | By cielak (which zequence should have saw him in UDS-R getting the winner of the automated testing writing competition) | 14:42 |
| zequence | smartboyhw: I see you are making a package out of it | 14:44 |
| smartboyhw | zequence, yes helping him a bit (and to practise my packaging skills apart from the Kubuntu ones) | 14:44 |
| zequence | smartboyhw: I would suggest you make sure no one is already doing that for Debian, and if not, sending it there | 14:45 |
| smartboyhw | zequence, no Kubuntu packages are ALWAYS -0ubuntu1.... I got people watching over me:P | 14:45 |
| zequence | smartboyhw: Yes, but this is not a Kubuntu package | 14:45 |
| zequence | it belongs in Debian Multimedia team | 14:46 |
| smartboyhw | zequence, I know.... I might persuade him to put it into debian one day...But normally his software doesn't get into the main archives (Debian or Ubuntu) | 14:46 |
| zequence | smartboyhw: Why don't you put it there? | 14:46 |
| smartboyhw | zequence, because he doesn't want to and I don't have a debian system:P | 14:47 |
| zequence | Saves you the trouble of adding it to Ubuntu, and also makes sure it is spread through out the Debian based environment | 14:47 |
| zequence | Why not? | 14:47 |
| smartboyhw | zequence, I didn't say I want to add it to Ubuntu (nor did he) | 14:47 |
| smartboyhw | We just want to keep it in a PPA that's all | 14:47 |
| smartboyhw | I wasn't sure if he will maintain this project for long also | 14:47 |
| smartboyhw | Since he is using it as his university homework | 14:48 |
| smartboyhw | s/homework/project/ | 14:48 |
| kubotu | smartboyhw meant: "Since he is using it as his university project" | 14:48 |
| smartboyhw | Thank you kubotu | 14:48 |
| smartboyhw | LOL | 14:49 |
| zequence | If it's GPL, it isn't really up to him anyway. I haven't tried it, but if it is useful, it should be packaged for Debian, no doubt | 14:49 |
| zequence | smartboyhw: Does it support jack? | 14:51 |
| smartboyhw | zequence, I will ask the dev | 14:51 |
| smartboyhw | zequence, NO it doesn't support JACk | 14:52 |
| zequence | Yeah, just read about that | 14:52 |
| zequence | this makes it a lot less usefull in deed | 14:52 |
| smartboyhw | grrrr right? | 14:53 |
| smartboyhw | zequence, alright let me try to play with debian packaging | 14:56 |
| zequence | astraljava: Re: music to have on when working by the computer - I've been listening to a lot of Aphex Twin. He's done a lot of tracks, but too few to have on permanently. | 15:11 |
| zequence | If there were only more of them, I wouldn't need anything else | 15:11 |
| zequence | smartboyhw: If you'd like to do packagin for Ubuntu Studio, I'd recommend packaging linux-sampler, and other packages that aren't part of the Debian distro | 15:33 |
| zequence | smartboyhw: falktx already does that, and you could look at his packages as examples, even if they aren't from scracth | 15:34 |
| smartboyhw | zequence, well ok let me finish calligra for Kubuntu first:P | 15:34 |
| zequence | scratch* | 15:34 |
| smartboyhw | zequence, ok. Also I need to do some iBus packaging for our Hong Kong people:P | 15:35 |
| smartboyhw | zequence, linux-sampler isn't in Ubuntu? | 15:35 |
| zequence | smartboyhw: Preferably, we try getting those into Debian, if possible. But, if not, then at least it would be good to keep a official PPA for stuff we can't include (may be we can't even keep it in a PPA - we'd need to investigate). | 15:35 |
| smartboyhw | zequence, OK | 15:36 |
| zequence | linux-sampler is GPL, with an addition, which makes it non-free | 15:36 |
| zequence | The addition is for protecting against using it in commercial hardware samplers and such, but it causes problems | 15:36 |
| zequence | It's the only sampler that can load gig files on Linux. One of the only real samplers around | 15:37 |
| zequence | not bug free though, from what I've experienced. At least not when used with qsampler, an interface for it | 15:37 |
| zequence | Debian has all of the packages linux-sampler depends on, except for the actual sampler server | 15:37 |
| smartboyhw | ok | 15:39 |
| contrapunctus | lmfao, kubotu XD | 15:43 |
| === zequence_ is now known as zequence | ||
| len-1304 | zequence, there is a new version of PA again today. | 21:49 |
| zequence | len-1304: Ok. Let me check. | 21:59 |
| len-1304 | It showed up in my daily update. | 22:00 |
| len-1304 | (for 13.04) | 22:00 |
| zequence | len-1304: I was just looking at the changelog (apt-get changelog pulseaudio). Didn't seem like there was much added | 22:01 |
| zequence | A quick bugfix to the package itself | 22:02 |
| len-1304 | Not the master mixer then. (I still need to play with that.) | 22:03 |
| zequence | Made by the same guy who got the lowlatency kernel uploaded initially | 22:03 |
| zequence | Who also used to hang out here until fairly recently - themuso. I met him briefly at UDS | 22:04 |
| len-1304 | I remember him yes. | 22:04 |
| zequence | len-1304: I don't think how the master level works is a bug. Sort of like the master level on a mixer doesn't affect a bus level | 22:05 |
| zequence | Only if the bus goes through the master | 22:05 |
| zequence | In this case, the bus can be made to bypass the master, and connecting directly to another IF | 22:05 |
| len-1304 | Ya, but I need to play with it so I have a better understanding of what you mean. | 22:05 |
| zequence | len-1304: Don't know if you need two devices. Perhaps only one will do, if it's your delta device | 22:06 |
| zequence | You do have the usb device, so that should do | 22:06 |
| len-1304 | I have two PCI cards. | 22:07 |
| len-1304 | an ice1712 and an ensoniq. | 22:07 |
| len-1304 | but I have the ensonic turned off in pulse. | 22:07 |
| zequence | I have three. builtin, hdmi from the graphic card, and the delta. This builtin is strange in that I'm not able to use with jack for some reason | 22:08 |
| zequence | Works fine with PA and alsa. Very weird | 22:08 |
| len-1304 | I have found that having pulse connected to a second card while it is also connected to jack causes coupling problems. | 22:08 |
| len-1304 | My internal IF is turned off in bios. | 22:09 |
| len-1304 | (the ac97 is too noisy for me.) | 22:09 |
| zequence | It is something that you can't always choose, as is the case with my hdmi. And, the stuff should of course work with any card configuration | 22:10 |
| zequence | Haven't done any testing with disabling my other devices | 22:10 |
| len-1304 | It works ok but having pulse see another card while it is bridged to jack affects the latency. | 22:11 |
| len-1304 | The minimum latency becomes that of the highest latency device in my experience. | 22:11 |
| len-1304 | (YMMV) | 22:12 |
| zequence | In my case, I have two devices that don'w even start with jack | 22:12 |
| zequence | The third one is my delta device. On certain machines, I've been able to run it at 16 f/p, but only with jack. Not with additional jack apps | 22:12 |
| zequence | Works now too | 22:13 |
| zequence | on this machine | 22:13 |
| zequence | If I run jackd from the command line, I don't get xruns. But as soon as I open an application, it gets messy | 22:14 |
| zequence | I'm even able to run jackd with PA at 16 f/p, but with loads of xruns | 22:16 |
| zequence | Maybe MBs are just more different from each other than one would think? | 22:16 |
| len-1304 | I can run at 16 f/p as well, but just about anything can make an xrun. I can run at 32 for hours with no xruns with audio running through it. | 22:19 |
| len-1304 | MBs are different. But more than that I think each one needs to be gone over before using it for low latency audio. | 22:19 |
| len-1304 | bios settings make a big difference | 22:20 |
| zequence | len-1304: But, if you have the other device enabled, you can't run jack at 16 f/p? | 22:21 |
| zequence | I'm thinking this is because of the module, as it might make jack crash at low latencies | 22:21 |
| zequence | As some applications do, when pushing it | 22:22 |
| zequence | oh, I'm talking about the PA module, of course | 22:22 |
| len-1304 | I haven't tried that here on this machine. but on my netbook the internal audio only goes down to 128f/p and when I have it enabled my USB IF is the same but if I turn it off in pulse then I can get the USB device down to 64. | 22:23 |
| zequence | len-1304: But, is that only when you use the PA module? | 22:24 |
| len-1304 | Yes. | 22:24 |
| len-1304 | Only when bridging pa-jack | 22:25 |
| zequence | I'm having a weird bug on PA. I can't select an input device. The mixer crashes. It seems confused. None is actually selected | 22:25 |
| zequence | len-1304: That's the part I can't really figure out. Why a device that jack is not trying to run would affect jack lowest possible latency. Other than the PA module having problems getting in sync and making jack crash | 22:26 |
| len-1304 | Ok, I enabled the ensonic and can still run at 32f/p | 22:27 |
| len-1304 | I should see if I can run the ensonic in jack at 32... | 22:27 |
| zequence | 32 should work for most devices, but 16 is a different story | 22:28 |
| len-1304 | a few more xruns with the other device connected though. | 22:29 |
| len-1304 | With jack running the ensonic pulse is running 10% less cpu because it is only sending two channels to jack instead of 10. | 22:31 |
| len-1304 | unable to start the ensonic at 16. | 22:32 |
| zequence | That would be highly expected, yes | 22:32 |
| len-1304 | Ok, jack starts the ICE at 32 with pulse connected to both jack and the ensonic, but there are xruns all the time. | 22:34 |
| len-1304 | with the ensoniq turned off in pa I could run for several minutes at least before the first xrun | 22:35 |
| zequence | len-1304: Try without the PA module. | 22:35 |
| zequence | you could just run jackd from the terminal: jackd -d alsa -d hw:n -p 16 | 22:36 |
| len-1304 | unloaded module - no xruns. | 22:38 |
| len-1304 | cpu is really low too :) | 22:38 |
| zequence | I think what happens with PA is that the module is a part of PA, and if it has more stuff to do, then it will work less well with jack | 22:38 |
| len-1304 | but that is to be expected | 22:38 |
| zequence | Another problem may be hardware interrupts, if you're using both PA and jack in paralell, but without the module | 22:39 |
| zequence | I don't see however how PA would affect jack latency directly. If jack uses a device, PA is not | 22:39 |
| zequence | And PA is strictly a jack application when connecting to jack. And some apps are really lousy at low latencies | 22:40 |
| zequence | That's my theory, anyway | 22:40 |
| len-1304 | I think pa defaults to a lower latency than jack and so with the bridge it has to keep up with jack's latency. | 22:40 |
| zequence | len-1304: So, PA is able to run at 16 f/p on your onboard? | 22:40 |
| zequence | While jack is not? | 22:40 |
| len-1304 | no | 22:41 |
| len-1304 | I mean the bridge module has to run at the same latency. | 22:41 |
| len-1304 | pa doesn't | 22:41 |
| zequence | The module may be connected to jack, while PA is using another device as output. The other device works flawlessly, while only jack gets audio droupouts | 22:42 |
| zequence | Yeah | 22:42 |
| zequence | The module is like a jack application | 22:42 |
| zequence | But, it's also depending on what PA is doing, I guess | 22:42 |
| zequence | So, if PA is busy, that might mean less resources for the module? | 22:42 |
| len-1304 | PA takes about twice the cpu as jack at any latency with the bridge connected. | 22:43 |
| len-1304 | PA seems to use almost nothing compared to the module. | 22:43 |
| len-1304 | with the module unloaded pa is using 1/5 what jack is | 22:44 |
| zequence | Can't say I know much about jack coding, but perhaps the module could be made a lot more efficient. David has often said it's more of a beginning towards a real tool | 22:44 |
| len-1304 | 16 frames may run without xruns, but soon as I start guitarix... 1000s of xruns before the gui even shows up. doesn't slow down after either. | 22:50 |
| zequence | yeah, I haven't heard of anyone being able to run applications xrun free at 16 f/p | 23:02 |
| zequence | Often jack used to crash for me, if I tried | 23:03 |
| zequence | next release of the kernel skipped one part of the SRU process, as they were in a hurry, as it seems | 23:05 |
| zequence | the part being verification testing. Something I don't do, since generic already does it, and it's the same kernel anyway | 23:05 |
| len-1304 | for 12.04.2? | 23:05 |
| zequence | both Precise and Quantal | 23:06 |
| zequence | having problems making the patch. PA build dependencies are unmet. libjack is the wrong version | 23:26 |
| zequence | ah, I need to install it manually | 23:27 |
| zequence | it removes some stuff | 23:28 |
| zequence | libjack0, that is | 23:28 |
| zequence | Ah, I'm too tired. It's past midnight. Just writing the damn patch description is more work the generating tha patch | 23:34 |
| zequence | heh, the last kernel update fixes quite a bunch of bugs | 23:38 |
| zequence | This one was particularly interesting, bug #1095315 | 23:38 |
| ubottu | bug 1095315 in linux (Ubuntu) "[USB-Audio - FastTrack Pro, recording] No sound at all" [Undecided,Fix released] https://launchpad.net/bugs/1095315 | 23:38 |
| len-1304 | I think that was why I didn't buy the device. | 23:39 |
| len-1304 | I could see the the ports but get no sound through | 23:39 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!