[05:05] <len-nb> micahg, did we get to -setting?
[05:06] <micahg> no :(
[05:06] <len-nb> Ok I won't test then.
[05:16] <micahg> len-nb: I'm sorry, I'll try to get to it in the morning, I've been running around most of the day
[05:26] <micahg> len-nb: once I actually fix the history, I'll be able to update -settings quickly again
[21:09] <ttoine> zequence, http://blog.ttoine.net/files/2013/01/Performance-MixBus.txt
[21:12] <zequence> ttoine: I think it's too bad you didn't do tests with the Ubuntu Studio desktop as well
[21:12] <zequence> And, personally, I find gnome-shell more interesting than the classical interface in this context as well
[21:20] <zequence> ttoine: Did you get xruns at 1024 frames/period?
[21:21] <zequence> There are so many variables between the systems too
[21:21] <zequence> It's not possible to say it's the graphic card that makes the difference
[21:23] <zequence> The only thing that is interesting I find is the big contrast in CPU usage on the first machine
[21:36] <ttoine> zequence, no x-rub at all
[21:36] <ttoine> run
[21:37] <ttoine> It was not possible to test Ubuntu Studio on all pc: there was already a 12.04 LTS and I did not have the right to install something else
[21:38] <ttoine> zequence, and yes, that's why I would be very interested to test with more nvidia videocards
[21:38] <ttoine> http://blog.ttoine.net/2013/01/28/harrison-mixbus-et-ubuntu-12-04-chez-azarecord/
[21:39] <ttoine> I did the tests for azarecord
[21:40] <ttoine> they would like to be sure that Ubuntu was a good choice for a Mixbus workstation
[21:40] <ttoine> they asked to stay with Unity in order to look like Os X
[21:40] <ttoine> they didn't like xfce
[21:41] <ttoine> f I have the possibility, I will continue to test this session on other pc, from different generation, and with different hardwares
[21:43] <ttoine> I think that scott would be interested. any news from him ?
[21:52] <zequence> ttoine: So, what you did was make sure Harrison Mixbus would function at all for Unity, basically?
[21:53] <zequence> ttoine: I'm not sure what the purpose of the test was, and what you found from the results
[21:56] <ttoine> zequence, XFCE was not the choice of Azarecord. So the purpose was to test wich other desktop was possible for production. Speaking a bit with the support of Harrison Mixbus team by mail, I learned that with Linux, it was very different depending of hardware and drivers (restricted, free, and versions)
[21:57] <ttoine> So the aim was to test different cpu, different gpu and if possible different mix of cpu and gpu. And as it is very interesting, I think it would be great to test on other hardware (core i5, amd cpu with nvidia gpu, etc.)
[21:58] <zequence> ttoine: So, what you did was basically make sure Mixbus would be able to start, and would not crash, and would run on Unity.
[21:59] <zequence> The differences between the hardware are so vast, that it's impossible to make any conclusions based on which GPU or CPU
[21:59] <ttoine> the matter is when if find someone with a computer running Ubuntu, It is always possible to get it for tests, a few time. But people don't let me install Ubuntu Studio. So I did the test starting from Ubuntu
[22:00] <ttoine> zequence, at this time, it is not possible 
[22:00] <ttoine> zequence, it was really considered to use gnome classic, actually
[22:01] <zequence> I've actually stopped using proprietary drivers with Ubuntu now
[22:02] <zequence> Both nvidia and ATI can do 3d graphics and HD video with the free ones
[22:02] <zequence> ATI is more or less fully functional, but I do think one should take care with expensive cards, and make sure they don't overheat
[22:03] <ttoine> zequence, what is strange, is that Mixbus uses cairo for rendering, and the harrison effects use OpenGL. and there are a lot of differences between gpu drivers. for example, the gui of harrison effects has a known display bug only with amd restricted driver. and harrison is aware that because a bug between cairo and nvidia restricted driver, Mixbus display is not stable when compositing is enabled
[22:03] <zequence> nouveau drivers seem to be fully functional at this moment, but I'm not sure yet. 
[22:03] <zequence> ttoine: Did you try with the open source driver?
[22:04] <ttoine> zequence, the free driver of AMD doesn"t manage HD7000 at this time. If you look at my test, the amd A6 has an integrated HD7000, so compulsory to use the restricted driver at the moment
[22:04] <zequence> Might be that 12.04 is not the best platform for the free drivers, don't know
[22:04] <ttoine> zequence, even with the x-edgers ppa ....
[22:05] <zequence> I didn't realize the logic for not keeping jockey in system settings for 12.10, until I actually tried the free ones
[22:05] <zequence> I just played Halflife Beta on the Linux Steam client, using nouveau drivers
[22:06] <ttoine> on my thinkpad, I did the test with the nouveau driver. It was simply impossible to use Mixbus with compositing enabled. The matter when using unity 2D or Gnome classic without effects is that when you change of windows, example form editing to mixer, there is a peak on the cpu, and so, x(runs
[22:06] <zequence> Might be good of us to pass these things on to the developers
[22:07] <zequence> We should probably add that to our testing routine, if we ever get one
[22:07] <ttoine> the other matter with 12.10 is that compositing is not stable enough
[22:07] <ttoine> I spoke of that with didrocks, he told me that it was not the aim for 12.10. it will be aim for 13.04
[22:10] <zequence> Well, it's starting to look very promising for open source graphic drivers
[22:11] <zequence> I would have thought what we have right now was impossible
[22:11] <zequence> ..then, I don't know how those drivers work
[22:13] <zequence> I'm going to work on sketching out some kind of barebone for testing
[22:20] <ttoine> zequence, both amd and nvidia free drivers are made with the same framework, if I understand well. the difference is AMD provide infos, and nvidia let people do reverse ingeneering
[22:21] <ttoine> it is clear that 13.04 will be a good test for a more open source Ubuntu, with less restricted drivers
[22:23] <ttoine> zequence, http://www.phoronix.com/scan.php?page=article&item=ubuntu_1210_amdstock&num=2
[22:23] <ttoine> test in october 2012, free against restricted driver with several amd gpu cards
[22:30] <zequence> ttoine: Wow, still pretty big difference. I was running HF2 using free ATI drivers with full settings, and it went pretty well, using Wine
[22:30] <zequence> HL2*
[22:30] <zequence> But, I wasn't measuring frame rate, and the card was quite new and fast
[22:31] <zequence> Plus, HL2 is not that new anymore, even if some levels do require a bit of power
[22:41] <zequence> I put down some stuff here https://wiki.ubuntu.com/UbuntuStudio/Testing
[22:41] <zequence> I think the best way to get continuous testing done on the full array of things we need to get tested, automated testing is the only way to get it done
[22:42] <zequence> Structured manual testing for the ISO mostly
[23:04] <ttoine> zequence, good
[23:05] <ttoine> zequence, the matter of restricted against free driver performance is mainly for applications using OpenGL. At first, one could think about Blender
[23:05] <ttoine> But even Ardour 3 will use Cairo and will be dependant of graphic performance...
[23:11] <zequence> ttoine: It's not just a matter of how much CPU it uses
[23:12] <zequence> It's also how much priority the drivers demand
[23:13] <len-1304> zequence, do the drivers have options?
[23:13] <len-1304> (the open ones)
[23:15] <zequence> len-1304: I don't anything about the drivers. Just that some hardware+drivers seem to impact stability more than others
[23:16] <zequence> stability being audio performance
[23:16] <zequence> ..in this case
[23:16] <len-1304> Maybe we should be asking questions of the open video driver authors about different use cases.
[23:16] <zequence> I absolutely think it would be a good idea for us to report any problems we find
[23:16] <len-1304> High speed video/gaming may not be the best for audio.
[23:17] <zequence> And to make them aware of what we find important
[23:19] <zequence> I don't think the aim for the free drivers is 3d gaming. I'm pretty sure they just want to make the drivers give the user full access to the cards functionality
[23:19] <len-1304> Makes sense.
[23:20] <zequence> high speed video for gaming could be == 3d modelling
[23:20] <zequence> which in our case is mostly Blender
[23:20] <len-1304> another topic :)  PA-jack
[23:21] <len-1304> The web site for pulse says that it should be possible to pass the number of channels to use.
[23:22] <len-1304> I would think for most uses 2 would be more correct than 7.1 plus extras.
[23:22] <zequence> Yeah, with jack that would probably make most sense, most of the time
[23:22] <len-1304> The default is to add as many channels as hardware ports
[23:23] <zequence> the pa to jack module could perhaps be enhanced, and be made configurable
[23:23] <len-1304> The web site does not give usage
[23:23] <len-1304> module-jackdbus-detect
[23:23] <len-1304> This module automatically adds JACK sinks and sources whenever the JACK server is started. For this to work, you need to use JACK 2, and enable JACK's D-Bus interface.
[23:23] <len-1304>     channels
[23:23] <len-1304>         The number of channels to create for both module-jack-sink and module-jack-source. If omitted, the sink will use the number of physical output ports and the source will use the number of physical input ports registered in the JACK server. 
[23:23] <len-1304>     connect
[23:23] <len-1304>         Takes a boolean value. If enabled (the default) PulseAudio will try to connect its ports to the physical playback ports of the JACK server. 
[23:24] <zequence> Well, diwic (David Henningson) is the sole author of that module, so he would know
[23:24] <len-1304> I have used the connect no option before with success, but the channels option is new
[23:24] <zequence> I'm guessing he might have written that text as well
[23:25] <len-1304> So it may use a different syntax than other modules.
[23:26] <len-1304> load-module module-jackdbus-detect connect=no
[23:26] <len-1304> seems to work
[23:26] <len-1304> but channels=2 does not.
[23:26] <zequence> I see
[23:29] <ttoine> zequence, is there a kind of Ubutu Studio install with just the system, the tweaks and no applications ?
[23:31] <zequence> ttoine: I would say, install ubuntu-minimal (network install, but don't install any desktop), then add the ubuntustudio-desktop meta. But, not sure if that actually is the close to or the same as what you get when installing from ISO
[23:32] <ttoine> it depends id -desktop meta package asks for -audio, -video, etc...
[23:38] <ttoine> zequence, in your automated testing, it could be interesting to have a score
[23:38] <ttoine> at least audio score and graphic score (for 3D)
[23:38] <zequence> ttoine: Installing the applications I don't think will affect the system performance in any way
[23:38] <zequence> So, one might just as well just install the Ubuntu Studio ISO, if one wants the desktop and the settings
[23:38] <zequence> ..user won't be in audio group, if installing ubuntustudio-desktop, but that much you know
[23:38] <zequence> As for settings.. I'm not really on top of it. I think len-1304 added something
[23:38] <zequence> Might get included if you install ubuntustudio-desktop, or the default settings meta, or however that works
[23:38] <zequence> We should add the option of installing the workflows on our ISO. I'm supposed to investigate adding such a plugin to ubiquitity
[23:38] <zequence> ttoine: the -audio and other workflows aren't dependencies to the desktop meta
[23:38] <zequence> as they shouldn't be
[23:38] <zequence> I'm going to get some sleep. Past midnight, early day tomorrow :)
[23:40] <ttoine> zequence, ok. the only bad stuff with installing ubuntustudio-desktop is that it changes the boot and stop screen