#milkymist IRC log for Monday, 2011-11-07

wolfspraulwpwrak: I forwarded you the faderfox reply02:23
wolfspraulhe is one of the vast majoriy who always hits reply, not reply-all...02:24
wolfsprauljust a heads up - Jon is still traveling, but in London his m1 (rc2 with l3/l19 fixes) froze on him several times02:56
wolfspraulhe thinks it may be related to using the mouse02:56
wolfspraulso he's in the GUI, moving around with the mouse -> freeze02:56
wolfspraulfirst time I hear this, will keep an eye on it02:56
wolfspraulreboot fixed it, so it was not a big problem, but slightly embarassing with m1 hooked up to a big screen...02:56
wolfspraulrejon: I just mentioned your m1 hangs, maybe cause by mouse movement...03:10
wolfspraulif you have any additional feelings about this (as to the cause), please post...03:11
wolfspraulwpwrak: ok so mathias (faderfox) also suggested the LC2 and especially LD203:11
wolfspraulI can somehow see how the LD2 could make sense03:12
wolfspraulit's very simple, but all endless encoders so it could work well with m103:12
wolfspraulplus it works now (midi), unlike lv3 which first needs midi-usb support03:12
wolfsprauldoes the LD2 interest you at all?03:13
wolfspraulwe could ask him to send a LV3 to Sebastien, and LD2 to you? I can also imagine buying some for myself/Jon/Xiangfu/others03:13
wolfspraulsomehow I think the LD2 will give us quicker results at the beginning, until Sebastien has added usb-midi support (I don't think anybody else can add that quickly, so either Sebastien or no usb-midi for now)03:14
wolfspraulmathias has discounted another 2nd gen model to 99 EUR, though it's the LX2 which doesn't make much sense for m1 imho03:14
wolfspraulmaybe he can give us a better price on the LD2 as well, if we order a few... (in addition to the m1 swap program which I will definitely do because it gets an m1 into the hands of someone in this industry, which is a must for us)03:15
wpwrak(FN 2.0 GUI) well, i already suggested a touch screen/tablet. and yes, i had virtual sliders and knobs in mind :) i still think that would be a very nice complement for more physical controls (or even a replacement for some people)10:38
wpwrak(gui) the patch would then simply define which set of controls it needs and the GUI would set up a suitable virtual layout. of course, one could combine this with a learn mode where you tap the virtual control, actuate the physical control (MIDI or such), and the physical control gets assigned10:40
lekernelyeah, that and also replace the equation editor10:41
lekernelactually maybe even the complete MD process :) as a superset of the current process10:41
lekernelthere are a lot of hardcoded things which can be made more configurable10:42
wpwrakwolfspraul: (faderfox) thanks ! hmm, so doesn't like to make LV3. grmbl.10:45
wpwrak(rejon's freeze) maybe another variant of the MIDI freeze ? i wouldn't be terribly surprised if that wasn't limited to just midi10:46
wpwrakyeah. that was easy. killed it :-)10:47
lekernel?10:47
lekernelyou fixed the midi bug?10:47
stekernlekernel: apart from looking insanely cool, is the reactable really providing an enhanced user experience compared to more conventional methods?10:48
wpwrakwolfspraul: (usb-midi) i think by the time we have the devices, sebastien will probably have implemented the support :) i think he once said the protocol didn't look too ugly10:49
lekernelwell, if you could edit M1 patches with such an interface, it could motivate people to make some10:49
lekernelit doesn't have to be a touch surface btw, a 1st version could work with mouse (and be compatible with the current hardware)10:49
wpwrak(fied) no, it's still at large :-( but i was able to hang my M1 with the mouse, much like i did with midi10:49
lekernelah?10:50
lekernelhow did you do? moving it a lot?10:50
lekernelit never happened to me10:50
lekernelMIDI and mouse use very similar code ...10:51
lekernelso yes, it would make sense that the mouse can cause a crash as well10:51
lekernelmaybe RTEMS doesn't like posting a message into a queue from an ISR?10:51
wpwraki started rendering, created a "busy" scene (not sure if scene complexity actually matters), then jerked the mouse around a bit. i accidently hit the mouse button, so it went back to the GUI, and hung right there10:52
wpwrakhmmm10:52
wpwrakinterrupt synchronization issues would explain a lot10:52
stekernyes, connecting together "boxes" with functionality would probably be more user friendly10:53
wpwrak(cooler gui) first, i'd clean up the language. as a i said, the whole thing is a bit of an insult to the whole IT profession :) we can still maintain compatibility with the old syntax10:54
lekernelhe, say that to the milkdrop guys :p10:54
wpwrakyes yes, i know you're not to blame for that ;-)10:54
wpwrakfor creating effects, i think just having a ton of controls can be very useful for the creative process10:55
wpwrakwolfspraul: sebastien suggested a joysticks because several variables are truly 2D vectors. e.g., points of reference (center of rotation, center of expansion) or shift vectors11:00
wpwraki haven't use the pad as a true 2D control yet, just to control a pair of unrelated parameters11:02
wpwrakbut i plan to do so. there are currently two obstacles: one is the hang, which gets in the way quickly. the second one it the small number of MIDI variables. i need to look into enlarging this set. maybe doubling or quadrupling it.11:03
wpwraknot sure if this is easy or hard. probably not too bad. gui will get a bit crowded.11:04
lekerneluse pages, like on the DMX desk?11:04
wpwrakah, note to self: i should test gui changes in 640x480.11:04
wpwrakyes, that's an option. may be the only choice in the end anyway11:05
wolfspraulwpwrak: ok, let's hope my connection doesn't break all the time11:10
wolfspraulso yes, it seems the LV3 is something he wants to phase out already11:10
wolfspraulafter just barely having added it11:11
wolfspraulalso it's not in his 10-minute intro video of 3rd gen controllers11:11
wolfspraulonly the other 5 are11:11
wolfspraulI will confirm this in my next mail11:11
wolfspraulwhat do you think about the LD2 he pointed to?11:13
wolfspraulhis naming is a mess. it looks like the FX3 is the closest 3rd gen equivalent to the LD2? :-)11:17
wpwrakyes, his naming is very messy. you need a dictionary :)11:18
wpwraklemme check the LD2 ...11:18
wolfspraulI have a number of questions/replies already on his last mail, the only thing I need to know from you is what you think about the LD2, and which one in general you would be happy to receive as an early Christmas gift11:18
lekernelwpwrak, I wouldn't be surprised if some bug in the RTEMS port to LM32 messed up the use of queues with interrupts11:19
wolfspraulis there any software for an android or ipad tablet that would turn it into a midi controller?11:19
lekernelthe LM32 port (initially made by a Finnish company) is "nothing serious" and no one really cares11:20
wolfspraulI think we should just be practical and pick the fastest and cheapest way *now* to get more value out of m1, for ourselves first, then our customers11:21
wolfspraulthat's why I'm a little drastic and would even prefer the LD2 right now, because I can plug it in and make it work fast, without usb-midi support11:21
wpwrakthe LD2 is just rotary encoders (with push function). that may be a bit limiting.11:21
lekernelhttp://www.rtems.com/ml/rtems-users/2010/october/msg00031.html11:22
wpwraki'd still want the LV3, if only to have as diverse set of controls as possible to experiment with11:22
wolfspraulonce we have usb-midi, the LV3 is great for experiments but it sounds like it will be phased out, joysticks causing too much trouble for him11:22
wolfspraulwpwrak: yes, but he says they have great precision and are endless11:22
wpwrak(usb-midi) as i said, it's not really a problem if you have a PC11:22
wolfspraulbut the lv3 doesn't work today, can we really be that optimistic that usb-midi won't turn into a many-months endeavor?11:22
wolfspraulah ok, sorry I forgot that11:22
wolfspraulwell then11:23
wolfspraullv3 your way?11:23
lekernelwolfspraul, send me that LV3 and i'll get it to work (natively)11:23
wpwrak(usb-miid) it works for me. now. very day ;-)11:23
wolfspraulok ok11:23
wolfspraullv311:23
wolfspraulto Sebastien & Werner :-)11:23
wolfsprauland we get him an M1 in exchange, I really love that idea11:23
wpwrakeverybody loves LV3, except for his daddy :)11:23
wpwraka migration path from joysticks could be a touch pad11:24
wolfspraulwpwrak: why do you think endless rotary encoders with push are limiting?11:26
wpwrakthere would only be space for one pad, as opposed to two joysticks, though (well, in theory you could have two really tiny pads, but i think that would be too crowded)11:27
wolfspraulthey have push, replacing the need for keys, at least initially, unless the keys are velocity, pressure or displacement sensitive (will ask about that too)11:28
wpwrakfirst, they're not 2D, so you'd have to coordinate two rotary movement. also, you'd use both hands for a 2D control instead of just one11:28
wolfsprauland same for the faders I would think, only that you cannot so quickly go to either extreme maybe11:28
wolfspraulyes, compared to x/y you need two, agree11:28
wolfspraulthat's indeed bad11:28
wpwraki'm less sure about faster vs. pot vs. encoder11:28
wpwrakfaster and pot have fixed stops, which can be an advantage, a disadvantage, or both at the same time11:29
wolfspraulyes11:29
wpwrake.g., if you have well-defined semantics for the endpoints (such as no movement at 0), then it's nice if you can find that point easily, instead of having to "search" for it11:30
wolfspraulso basically you say we know so little about the exact usage on m1, for now we just want to have all variants to find the optimum use for them?11:30
wpwrakof course, if what you're controlling is endless in nature, such as the angle in a color wheel, a pot/fader would be limiting11:30
wpwrakalso, pots/faders/joysticks retain their physical configuration when software changes. e.g., if you load a new patch, that may be a conflict11:31
wpwrakthough it may be possible to avoid most problems there simply by being careful with what variable you assign to what control11:32
wpwrak(know little) yes, that's basically it11:33
wpwrakand it goes beyond this - we probably also want to give customers a lot of freedom there11:33
wpwrakso it's good if we have a device we can recommend. of course, there's the small problem that he wants to retire it. maybe we need to find out more about why exactly the joystick is such a pain for him11:34
wolfspraulhe wrote it11:34
wpwrakgluing ...11:34
wolfspraulexpensive/lots of work/glue/moq 100011:34
wpwrakbah, who's afraid of MOQ ? ;-)11:34
wolfspraulbut sure he is flexible11:35
wolfspraulI'm not so worried about that eol right now11:35
wolfspraulhe has over 100 in stock/production, is flexible, we on the other hand have nothing but plans right now11:35
wolfspraulso we can focus on what is best to make our side work well asap, and then we will always find a controller11:36
wpwraki'd at least like to try the joystick. see if they feel good and make sense. that will help with figuring out what to do next, e.g., joystick, pad, or abandon 2D and just combine 1D controls11:37
wpwrakyes, i guess if we're in shape to empty his current stock, the MOQ will not be his biggest problem ;-)11:37
wpwrakanother thing that may be worth looking into are pressure/velocity-sensitive buttons11:39
wpwraki could see a use for injecting images11:39
wolfspraulI will ask in my next round/reply11:40
wpwraki played a bit with this: connect the camera, assign video_a to a button, toss snapshots into the effect11:40
wpwraki'm having a bit of trouble with total brightness control.  and the nature of how patches are processed means that i can control the intensity of the snapshot with the duration of the button press11:41
wpwrakbut it may be useful to have a faster way to control the brightness. also, there may be others effects where you want to make something go "boom", where a pressure sensitive button could be nice.11:42
wpwrakneeds experimenting. an alternative could be pot/fader/encoder plus button11:43
wpwrakof course, again, this takes two hands :)11:43
wpwrakfor my demo videos, i used a lot of control inputs on four variables: zoom/blur, rotation speed, wave amplitude (audio sensitivity), and offset from the center11:45
wpwrakthat's about as much as you can do with both hands11:47
wpwraki put all the color controls on the nanoKONTROL2, and they're much slower to access. of course, the color mix is propagated slowly anyway, and the color change rate is something i feel doesn't need adjusting a lot. most of the time, i just keep it at "fast", to get a busy image11:48
wolfspraulwpwrak: and what types of controls worked best for those?11:49
wolfspraulsounds like they could all be encoders, no? zoom/blue, rotation speed, wave amplitude, offset from center11:50
wpwrakmaybe with improved color control, the color section would become more important. but i'll need to work on the math there, to get a useful "compression" of the color space. my current modulation of the original path in 3D space is pretty crude and makes it hard to get predictable results.11:50
wpwraki have them all on faders for now11:51
wpwrakah, i mean the color controls11:51
wpwrakfor the rest, i have amplitude on a pot, offset on a big fader, blur and rotation on the pad11:52
wpwrakthe use of the pad for these isn't driven by them wanting a pad but more my the pad being there :)11:52
wpwraks/my/by/11:52
wpwrakso i need to play a bit more with sending them to other controls. hence the "learn" mode :)11:53
wpwrakwhat's nice about the pad is that it's very fast. this is particularly interesting for rotation. there i can "wiggle" the arms of a star pattern by rolling my finger. or make pulsed rotation by holding one point and tapping another11:55
wpwrakjoysticks may be inferior in that regard. but let's see.11:56
wpwrakafk for a bit11:57
wpwrakback12:13
wpwrakjust found another little MIDI quirk. "running status". that's when a midi message reuses the status byte of the last one. luckily, none of my equipment seems to generate this sort of stuff. anyway, will be easy to support12:14
stekernthat is pretty commonly used12:15
wpwrakah, good to know12:16
wpwrakmidi is full of little surprises :)12:16
stekernyeah, it's one of those bw optimizations, I've learned about that one the hard way :)12:17
wpwraki like how they hid it at the end of the protocol description, as far as possible from other things at the same layer :)12:18
wpwrakstekern: do you know if there's anything on linux that generates such sequences ?12:20
stekerncan't remember if rosegarden did, or if it was some of my equipment that did12:21
wpwrakis there a decent midi sniffer ? i've tried kmidimon, but it always hangs after a few messages :-(12:23
kristianpaulepson emp 176512:25
stekernI think I just used an oscilloscope ;)12:25
wpwrakgrmph. does it always have to be the hard way ? :(12:26
kristianpaulthats the video projector dint allowme to demo M1 in pasto :(12:27
wolfspraulkristianpaul: excellent, thanks a lot!12:28
kristianpaulAnd btw i may demo M1 and nanonote in Ecuador in mid feb12:30
wolfsprauldaredevil12:30
wolfspraul:-)12:30
lekernelthose VGA problems are incredibly pesky12:32
kristianpaulindeed12:32
wpwrakinteresting. the emp-1765 is unknown to google. what kind of device is this ? portable, home cinema, big screen, ... ?12:32
lekernelwith another projector I had the issue that it detected 640x480 as 1280x1024 and only displayed a zoomed part of the picture12:32
lekernelI wonder what causes that12:32
lekernelmaybe using 25MHz instead of 25.175MHz as pixel clock?12:33
wpwrakmaybe some pulses the wrong length ? you'd think they could count the hsyncs between vsync, though12:33
wolfspraulone guy I met asked me whether m1 supports 'resolution detection'12:34
wolfspraulI don't know vga well enough - what 'resolution detection' is even possible?12:34
lekernelyes, with DDC12:34
lekernelthe hardware supports it, it's even part of the test routine12:34
lekernelFN does not12:34
wolfspraulcan m1 detect what resolution the other side wants/supports? or can the other side detect m1's resolution?12:34
lekernelM1 can detect12:34
lekernelif someone writes the right software12:35
kristianpaulwpwrak: big screen i guess, it was used to project like 6m from wall12:35
wolfspraulwait, who can detect what?12:36
wolfspraul(in vga)12:36
wolfspraulm1 can detect what the other side wants/supports? or the other way round? or both?12:36
lekernelthe M1 can probe the connected devices for resolutions it supports12:36
lekernelthen it output a signal in some of these resolutions, and the device will display it12:36
wolfspraulalso unusual/special resolutions, say 640x200 ? (random example)12:37
lekernelwith proper software support, which doesn't exist today, yes12:37
wolfspraulok good12:37
wolfspraulhe said that led walls often have unusual resolutions... (that's where the question came from)12:37
lekernelLED walls are often DMX... also possible, also lacks software12:38
wolfspraulhis last event had a led wall with a 13:2 aspect ratio12:38
wolfspraulyou mean all pixels of the entire wall are controlled by dmx?12:38
lekernelyes12:38
wolfspraulwow12:38
wolfspraulinteresting12:38
wolfspraulis dmx fast enough for that?12:38
lekernelusually you have several DMX connectors for that12:38
lekernelso using the single DMX connector on the M1 will limit the resolution12:39
wolfspraulok12:39
lekernelhowever it shouldn't be too hard to add up to 12 DMX outputs on the M1 using the expansion connector12:39
wolfspraul:-)12:39
lekernelthe hardware is trivial, but it will - again - need copious amounts of software supports12:40
wolfspraulthanks for helping me understand the vga resolution detection situation, that's helpful12:40
wolfsprauloh sure, one by one12:40
wpwrakmy wish would be more USB ports. a lot more :) usb-midi may already add 1-2 devices.13:06
kristianpaulsome: fedback from a guy wich work in visual arts and assisted to christopher adam sharism talk13:39
wpwraks/assisted/attended/ :)13:40
kristianpaulah yes yes...13:40
kristianpaulok:13:40
kristianpaulhe mention the lack of a video mixer13:40
kristianpaulusing a simple slider to mix two video inputs13:41
wpwrakah, he wants more inputs13:41
kristianpaulis expensive and dont allow live coding13:41
wpwraki think the mixer is kinda there13:41
kristianpaulall the expensive vj statations have video mixer13:41
kristianpaulalso the effects still not impresive at all13:41
wolfspraulthat is good feedback. how can we bind him a little more into the milkymist community?13:42
lekernelkristianpaul, did they use the camera? what patch did they show?13:42
wolfspraulalways point to the twitter account, maybe this channel13:42
wpwrakfor really good effects, i think we need better controls and also image support. that way, other design tools can be used to draw images and they can then be mixed into the patch13:42
kristianpaulwolfspraul: i already did point him to the channel and prefer stay where he is afaik13:43
wolfspraulsure, perfect13:43
kristianpaulhe just know i own a M1 and wanted to share the comments13:43
wolfspraullet's bring people together13:43
wolfspraulyes, perfect!13:43
kristianpaullekernel: i think he want to use two cameras at the same time13:43
wolfspraulwe want to hear it, even if we heard it before etc. we need to bring those people together13:43
kristianpaulor two video sources perhaps?13:44
lekernelkristianpaul, but did they use the camera during the demo? and what patches did they show?13:44
wpwraki'm not sure yet how good the current mixing features of M1 are. just made my first experiment with video on yesterday, in a bit of a hurry (to impress some girls ;-)13:44
wolfspraulone thing you can always reply today is that for sure, m1 is *different* from everything else13:44
lekernelI'm asking about the source of the "effects not impressive"13:44
wolfspraulthe entire architecture13:44
kristianpaullekernel: dont know13:44
wolfspraulthat difference alone should justify (to some people at least) to give it the benefit of the doubt today and follow where it's going, with whatever level of commitment13:45
wpwraklekernel: probably the usual: M1 running without sensitivity tuning - "screensaver mode"13:45
wolfspraulfrom that difference, I believe over time it will have features that people really love and will not find with the other solutions. but right now I'd have a hard time pointing to these features myself.13:45
wolfspraulkristianpaul: just saying this in general, I think people will be willing to listen and follow, m1 is something new :-)13:46
wpwraklekernel: i'm actually beginning to like the built-in microphone because it provides a setup-free audio input. that helps to bring at least some life into the patches.13:47
kristianpaulwolfspraul: sure sure, he knows that, i told him about M1, he saw it working at labsurlab for the dorkbot transmition and now from this christopher adam and made the comments13:47
kristianpaulof course i asked to reply this here, he said is OKAY13:47
kristianpaulhe is aware of copyleft hw, just when something dont seems to work for him, he just said it :)13:48
lekernelvideo mixer is so boring... but everyone's asking for that13:49
lekernelphew13:49
kristianpaulwell i just telling what _people_ _want_ :-)13:50
wolfspraulkristianpaul: another thing to keep in mind, whenever you talk to someone who already has VJing experience, is that they will naturally compare it to the stuff they already know13:50
kristianpaulyes13:50
wolfspraulit may be worth pointing that out even to them13:50
wolfspraulwe do not in any way want to claim that m1 can 'beat' all these other vj setups and software13:51
wolfspraulthat's also why we must be open about potential feature shortcomings13:51
lekernelkristianpaul, for the last weeks I have not been complaining about people who are on this channel, I've been complaining about people who are _not_ here13:51
kristianpaul;)13:51
wolfspraulnot that someone feels cheated or so once they find out the 640x480 resolution, for example13:51
wolfspraulso we clarify that13:51
wolfspraulbut then we say - m1 is DIFFERENT13:51
kristianpauli havent heard complaing about resolution yet13:52
wolfspraulsome people have patience and energy to follow things that are different, and some don't13:52
wolfspraulsure I just tell you how I think we can convert some people13:52
wolfspraulwhen they start comparing to their existing solutions, the end result is already clear13:52
wolfspraulit will be: don't switch13:52
wolfspraulwe don't want people to switch from resolume/modul8 etc on a notebook to m113:52
wolfspraulinstead, it's a new/different tool13:53
wolfspraulVERY DIFFERENT13:53
wolfspraul:-)13:53
wolfsprauleither they appreciate that difference, or m1 is not for them, at least today13:53
wpwrakoh. by the way, how difficult would it be to support RGB in, i.e., use all three video connectors ? is this just a question of setting the right registers in the codec or would there be more ?13:53
wolfspraulsee my point?13:53
kristianpaulno wait, he dint own a VJ stuff, but he already saw working some of then, i guesss prety similar situation if he already owned one13:53
lekernelwpwrak, in theory, just write a couple of adv7181 regs13:53
wolfspraulhighlight the difference of m1, the fact that it's entirely new and different13:53
wolfspraulbecause I think it's true13:54
wolfspraulwe compare features to a notebook/software solution, but in reality it's a completely different architecture13:54
wpwrakkewl. we should make that work at some point in time, if only because the presence of these three connectors strongly suggests you can use split RGB. and maybe it'll also help with video in quality.13:55
wpwrakwolfspraul: i'd emphasize the interactivity aspect. after all, people also buy gaming consoles :)13:56
kristianpaulanyway, i still asking people buy a new a very different M1 :-)13:57
wolfspraulinteractivity yes, but you are a little ahead in your use of controllers :-)13:58
wolfspraulbut of course, I agree13:58
kristianpaullekernel: if got your message right, you mean this people out there should join the channel so we can convert him ;) ?13:58
kristianpauli already realized midi is way to avoid get a boring M1 :-)13:58
kristianpaulwpwrak: had you already tried midi over osc?13:59
wpwrakkristianpaul: no, i haven't touched OSC so far. i picked the conservative approach and went all-MIDI14:00
wpwrakso yes, maybe you could also do USB-MIDI -> PC -> MIDO-to-OSC -> Ethernet -> M114:01
lekernelyes, that should work14:01
wpwrakand save USD 5-15 for the USB-MIDI dongle :)14:01
wpwrakwhat would be a program (on linux) that implements such a transport ?14:02
kristianpaulperhaps puredata14:03
wpwrakfhew. sounds very high-level14:04
kristianpaulyeah..14:07
wpwrakthis looks promising: http://controsc.sourceforge.net/14:09
wpwraklet's try ...14:09
wpwrakgrmbl. irc server dropped the connection14:17
wpwrakokay, how do i use OSC on the M1 ? i.e., how do i make things received end up in one of the oscN variables ?14:29
wpwrakor is this somethign that doesn't work yet ?14:31
wpwrakah .. midi-over-osc ought to end up in the midi subsystem. good14:35
wpwrakgrmbl. probably easier to write my own forwarder ...14:57
wpwrakdoes anyone have an example of a working oscsend ... /midi command ? i'm getting a little bit of reactions, but things are a tad weird14:57
lekerneloscsend /midi m ...15:04
wpwraki'm trying things like  oscsend m1 4444 /midi m b001020015:05
wpwrakwhat happens is that every once in a while. "Latest active controller" shows 1 (1)15:06
wpwrakleeet's debug this then ...15:16
lekernelpacket losses? OSC is based on UDP15:28
lekerneltry connecting the M1 directly to the computer15:28
kristianpaulhe said potencial is no bad, and people ask about live coding because thing fpga can do on that15:30
kristianpaulso, last qute from him "pero ver que deviene en un screensaver es unpoco triste para tanto harware power :)"15:31
kristianpaulwhat i understand as hardware been really powerfull but underdeveloped?15:31
wpwrakoh ;-)15:32
wpwraknaw, the problem is elsewhere :)15:32
lekernelmore bugs? phew15:32
lekernelthis is sooo painful15:33
lekerneland it works for me, as always15:33
wpwrakare you using OSC for midi messages ?15:33
lekernelyes, I tested it15:34
wpwrakso how would you send a controller message, like a  b0 01 02  sequence ?15:35
wpwrakoscsend m1 4444 /midi m b0010200 ?15:36
wpwrakor maybe 00b00120 ?15:36
lekernelsomething like that... but that's something I did one year ago15:36
lekerneland no one touched it since then15:36
wpwrakokay15:36
wpwraki think i figured out what's wrong. but it depends a bit on how oscsend is really supposed to work15:37
wpwrakah yes, 00b00102 is right15:38
wpwrakgood. then it's a one-line fix :)15:38
wpwraknow, let's test it ...15:39
wpwrakperfect15:39
lekernelah, no bug this time :)15:42
wpwrakoh yes, a bug ;-)15:43
wpwraki guess you were cleaning up some code and made a typo15:43
wpwrakcontrOSC doesn't see to want to do what we need. (uses some special-purpose identifiers) hmm. seems funny that there wouldn't be a tool for simple MIDI->OSC forwarding15:46
wpwrakwell, let's hack one ...15:47
lekernelah, yes probably15:48
lekernelthanks for fixing15:48
wpwraknp. i like an easy one from time to time :)15:49
mumptaimoin16:18
GitHub78[flickernoise] sbourdeauducq pushed 1 new commit to master: http://git.io/8NxJWw16:19
GitHub78[flickernoise/master] input.c: an OSC-MIDI message contains a complete command - process accordingly - Werner Almesberger16:19
GitHub22[flickernoise] sbourdeauducq pushed 1 new commit to stable_1.0: http://git.io/oJP2Fg16:19
GitHub22[flickernoise/stable_1.0] input.c: an OSC-MIDI message contains a complete command - process accordingly - Werner Almesberger16:19
lekernelhi mumptai16:21
Action: mumptai needs some coffee16:23
kristianpaulwolfspraul and lekernel: there is  sdram on the LX9 board, 64Mb LPDDR16:35
wpwrakoddly enough, i can't hang the M1 with MIDI-over-OSC17:23
lekernelif my "message queue from ISR broken" theory is right, it can make sense17:29
wpwrakyes, it's a good theory17:40
wpwrakthat would explain all the weirdness i'm seeing, plus it would explain why this kind of problem exists in such fundamental functionality17:41
wpwrakbut before i try to pry apart that code, the hangover from yesterday needs a bit more subduing :)17:41
qi-botThe Firmware build was successfull, see images here: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-11072011-1742/19:19
lekernelwpwrak, if you want to dive into RTEMS code, I'd recommend you use the upstream version (with the lag bug)... not the old one19:21
sh4rm4lekernel: i was looking at http://lists.milkymist.org/pipermail/llhdl-milkymist.org/2011-October/000011.html 30 minutes ago... now it returns a 40419:38
sh4rm4as if that single file had been deleted in the meantime... 00012 works19:38
sh4rm4nm, it works with another browser. seems i hit a random bug in chrome19:41
kristianpaulis okay here19:42
wpwraklekernel: hmm, i rather start with a basis i know a bit. else it's very easy to end up chasing ghosts. but i'm looking at diffs between CVS and our tree (from the SDK) now22:01
wpwrakinteresting detail: the rb trees are used in files that have "scheduler" in their names (and nowhere else). so maybe this is actually what causes the infamous lag.23:01
--- Tue Nov 8 201100:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!