Perl 6 - the future is here, just unevenly distributed

IRC log for #openframeworks, 2014-02-04

| Channels | #openframeworks index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:06 danomatika Ive done it many times so far and it's worked
00:07 danomatika there were only issues if/when there was a big difference in version, aka trying to run something compiled on a new version of OF on say a Mac OSX 10.5 machine
00:07 danomatika the libs etc are packaged in the app bundle so portability should work ok in general, of course YMMV
00:13 dantheman joined #openframeworks
01:00 bakercp joined #openframeworks
01:11 robot joined #openframeworks
01:17 dantheman joined #openframeworks
01:55 kureta_m joined #openframeworks
02:00 bakercp joined #openframeworks
02:02 danomatika bakercp
02:02 danomatika see the latest commit
02:03 bakercp ok
02:03 danomatika checkout cairo, it's a lot cleaner now
02:03 kureta_m joined #openframeworks
02:03 bakercp nice.
02:03 danomatika I also added a section on dependencies in the readme
02:03 bakercp super.
02:03 danomatika should be easy to implement now
02:04 finger563 joined #openframeworks
02:04 finger563 hello!
02:05 danomatika you can build depends by name and there is a "depends" target to run commands on everything in the _depends folder
02:05 danomatika it automatically calls dependency commands in the background when running a formula, but you can override that if needed and call things manually, there are convenience functions for this now
02:06 finger563 can anyone here help me with this: I've got this question:  I'm writing image data from the GPU to the SD card (on the rasberry Pi) and i've found that the bottleneck is actuallly writing the image data to disk (copying back to CPU is fast)
02:06 danomatika hopefully this is not too complicated
02:06 danomatika that will always be slow
02:06 danomatika thats the nature of computer architectures
02:06 bakercp @finger563  are you using a class10 sd card?
02:06 finger563 yea
02:06 danomatika best bet is to have as fast an sd card as possible
02:07 finger563 and i can only write ~2 frames a second
02:07 bakercp that's about all you can do as far as i know ...
02:07 danomatika yup
02:07 bakercp that seems pretty fast :)
02:07 bakercp ;)
02:07 finger563 :/
02:07 danomatika rip != new laptop
02:07 danomatika rpi
02:07 finger563 hah
02:07 finger563 i know
02:07 finger563 i'm an embedded systems guy, just never dealt with image processing
02:07 bakercp one option is to make images smaller, compressed
02:07 finger563 yea
02:07 danomatika yeah
02:07 finger563 i was thinking about doing image compression/encoding on gpu
02:08 bakercp could speed up a tiny bit.
02:08 finger563 how feasible you think that is?
02:08 bakercp that sounds like fun ...
02:08 finger563 haha
02:08 danomatika what format are you using now?
02:08 danomatika what about crapping out tiffs
02:08 finger563 well the gpu is supposed to have hardware encoding iirc
02:08 finger563 hmm
02:08 finger563 havent tried tiffs
02:08 finger563 tried all qualities of jpeg, and then png, pbm, ppm
02:08 bakercp yeah, i think there is hardware acceleration for jpeg
02:08 danomatika no compression might be faster ... but then you might be slowed done by writing a much larger file
02:09 danomatika worth trying
02:09 finger563 so you dont think OF uses hardware encoding on RPi?
02:09 finger563 i didnt find anything saying either way
02:10 danomatika he should know, he wrote a good part of the rpi backend :)
02:10 finger563 and i'd love to do it in the pipeline to avoid the extra copies
02:10 finger563 haha excellent :)
02:10 finger563 (also, good job!)
02:10 bakercp you might check in w/ jvcleave
02:10 bakercp he got video recording working the other day via openmax.
02:10 finger563 yea?
02:11 finger563 that sounds awesome
02:11 bakercp depending on your needs you might be able to use that
02:11 finger563 i've been doing image processing for feature detection in shaders, but writing it back to disk and transmitting is a problem now
02:11 bakercp @jvcleave ring ring
02:11 finger563 yea writing video would be good too
02:11 danomatika o bakercp I'm out
02:11 bakercp thanks for all your work @danomatika
02:11 danomatika I have a Mars show to fnknish up for Sat
02:11 bakercp i know it wasn't your plan today :)
02:12 danomatika release early, release often ... sometimes you just have to crap it out and move on
02:12 danomatika my problem is I have to make it as good as I reasonably can before moving on
02:12 finger563 haha "good" = enemy of "done" ;)
02:13 danomatika edison was right, he should know
02:13 danomatika there's also, "better = enemy of good"
02:13 danomatika as software is never done
02:13 finger563 also true
02:13 danomatika good night
02:14 finger563 g'night
02:14 danomatika you should have no problems now bakercp
02:14 danomatika left #openframeworks
02:20 finger563 is there a way I can enable H/W encoding of JPEG for imge writing?
02:20 finger563 image*
02:29 bakercp @finger563 i don't know - i know there is hardware jpeg decoding, but not sure about encoding.
02:30 finger563 hmmm
02:30 bakercp i just pinged @jvcleave, so perhaps he'll show up and set us all straight.
02:30 finger563 oh, thanks!  I'll take a look around some code to see what i can figure out
02:30 bakercp you might best find feasibility realted answers on the RPI forum.
02:31 finger563 good point, i'll check there too
02:31 finger563 :)
02:31 bakercp and check check this
02:31 bakercp https://github.com/jvcleave/ofxRPi​CameraVideoGrabber/commits/develop
02:32 finger563 oh, that does look promising.  I've been working my code off the develop branch over there but not messed with the actual recording stuff cause i wasn't sure how mature it was yet (i.e. functional)
02:34 bakercp it's fresh -- like 5 hours
02:34 bakercp jason is super responsive, so he'll probably get you hooked up.
02:34 finger563 :)
02:40 jvcleave what up - finger563?
02:40 finger563 hey jvcleave!
02:40 bakercp see -- i told you he was responsive :)
02:40 finger563 (also excellent work on the video grabber!)
02:40 bakercp g'night all. ;)
02:41 finger563 g'night
02:41 jvcleave thanks bakercp :)
02:41 finger563 so jvc, i'm using your develop branch currently to capture picam data, process it with frei-chen edge detection, and then write it to an image file
02:42 finger563 with everything but the actual writing i get good response even at higher resolutions
02:44 finger563 but the writing is a huge bottleneck for me
02:44 jvcleave yeah - develop is in a very work in progress/experiemental stage
02:44 jvcleave hmm - how much of a bottleneck?
02:44 finger563 so everything up to and including copying data back to CPU i get at least 30 fps for 1280x720 and 60 fps for 640x480
02:44 finger563 but image writing can only be done at about 1-3 frames a second
02:44 jvcleave yeah - anything that hits the CPU is tough. you using a class 10 card?
02:44 finger563 yea i am
02:45 jvcleave that helps a lot - there is image decoding/encoding on the GPU. I haven't messed with it much but the first place to start is https://github.com/raspberrypi/firmware/t​ree/master/opt/vc/src/hello_pi/hello_jpeg
02:46 jvcleave I would feed it one of your source images and see if it is any faster
02:46 finger563 thanks for that link, i was pretty sure there was, but haven't done much like this before so i'm not as familiar with it
02:46 finger563 sadly this is my first image processing project
02:48 dantheman joined #openframeworks
02:48 jvcleave well - that is actually a decoder - won't help you with writing speed
02:48 finger563 yea :/
02:48 finger563 but might point me towards the encoding
02:48 finger563 though i'll have to see if i can easily get the shader output to the encoder
02:49 finger563 he said you're doing video writing on the develop branch now? how are you pulling that off?
02:49 dantheman joined #openframeworks
02:49 jvcleave I can at least tell you it won't be easy :/
02:49 finger563 yup
02:49 finger563 well on that site theres some code for mjpeg encoding
02:49 finger563 maybe i can work along those lines
02:50 finger563 i'm not restricted to using images, i can write video
02:50 finger563 i just figured images would be easier and less intensive
02:51 jvcleave It all uses OpenMax underneath the hood which is pretty hard to get used to - the ilclient stuff is a bit of a wrapper for it
02:52 finger563 ah i see
02:52 finger563 yea when reading through your code i found a lot of openmax calls and such that i wasnt familiar with so i've been sorta learning those
02:52 jvcleave my addons don't use the wrapper - just "raw" OpenMax
02:53 jvcleave the biggest problem is that the RPi uses a custom implementation so even if you write the code to the OpenMax spec it may not work
02:53 finger563 is openmax doing the conversion from YUV?
02:53 jvcleave yeah - it can
02:53 finger563 yea, i've found thats true for a few other things as well
02:53 finger563 i do wish we had a newer version of OGL & OGLSL :P
02:54 finger563 you've configured it to do so by default, correct?
02:54 jvcleave so it is a lot of stabbing in the dark/trial and error
02:55 kureta_m joined #openframeworks
02:55 finger563 yea that's how i've been going the past few weeks on this project
02:55 finger563 took me a while to find your code even :P
02:55 finger563 (thankfully i did though!)
02:55 jvcleave I am just capturing the camera output to a raw h264 stream
02:56 jvcleave the other mode can capture to a texture that OF can use with shaders
02:56 jvcleave textured mode isn
02:56 finger563 yea that's the mode i'm using
02:56 jvcleave 't recordable tho
02:56 finger563 yea, i'm trying to make it recordable ;)
02:57 jvcleave I doubt if it is possible but I have been often suprised
02:57 finger563 what use is computer vision if you can't record the data or process it after the GPU handles it? :P
02:57 jvcleave well computer vision is all CPU based
02:57 finger563 unless you do it in the shaders
02:57 finger563 it's all matrix math anyway
02:57 jvcleave so until they make that GPU optimized
02:57 finger563 lol
02:58 finger563 i've already got the CV working on gpu
02:58 finger563 i just need a way to save the data XD
02:58 jvcleave how is that working?
02:58 finger563 extremely fast
02:58 finger563 i can do some nice edge detection and distance transform without noticing much performance hit
02:59 finger563 maybe 1-2 FPS out of 60 FPS,
02:59 jvcleave yeah - you can get pixels via an FBO
02:59 finger563 i.e. you code runs with the camera at 60 FPS, i get about 58-59 FPS
02:59 jvcleave but that is a big hit as well
02:59 finger563 not really
02:59 finger563 the copying back to CPU isnt that slow
02:59 finger563 i don't notice a performance hit
03:00 finger563 it's *just* the writing to the SD card
03:00 jvcleave at what resolution?
03:00 finger563 i've only run both 640x480 and 1280x720]
03:00 finger563 havent tried any others
03:01 finger563 or rather let me put it this way
03:01 finger563 and this may be where my issue lies
03:01 finger563 when i perform a readPixels from the fbo into an image i've defined
03:01 finger563 does that happen entirely on the GPU?
03:01 finger563 or is it doing the copying then from GPU to CPU
03:02 jvcleave the latter
03:02 finger563 then yea
03:02 finger563 i can get data onto CPU quickly
03:02 finger563 but i can't write to disk quickly
03:02 jvcleave you may be able to thread that
03:02 finger563 sadly it is already :/
03:03 jvcleave you using preallocated pixels?
03:03 finger563 granted
03:03 finger563 nope, which is part of my issue definitely
03:03 finger563 my threading has been the most recent attempt to speed it up
03:03 finger563 so i havent done all possibilities yet
03:03 finger563 but a few different thread scenarios i've tried all produce the same results
03:03 jvcleave yeah - threading and maybe a circular buffer
03:03 jvcleave of preallocated pixels
03:04 finger563 yea that should cut down a good bit i agree
03:04 finger563 and my final implementation will have that
03:04 finger563 so i should start is early :P
03:04 finger563 it*
03:05 jvcleave the RPI also lacks NEON which is what a lot of the other optimizations use
03:05 finger563 NEON?
03:05 jvcleave http://www.arm.com/products/pr​ocessors/technologies/neon.php
03:06 finger563 ah
03:07 finger563 yea i do wish we had that
03:07 finger563 though i'm thankful for the fast interface between CPU/GPU
03:09 jvcleave I've only used it with the video player - not the camera yet
03:09 finger563 i'm looking through the mjpeg encoder sample they have to see if it might help too
03:09 finger563 you taken a look at that?
03:10 jvcleave that is essentially what my recording is doing
03:11 finger563 well then worst case scenario, i can copy the data back from GPU then use the OMX mjpeg encoding to encode it
03:11 finger563 though i'm not sure what GPU cores that will use
03:12 finger563 so hopefully that won't impact my image processing as well
03:13 finger563 have you found performance hit when you've done recording?  or does your test code not use the gpu?
03:15 jvcleave it's been fine but it is actually all on the GPU so I am not doing any CPU operations to be able to notice
03:16 jvcleave the main thing I am trying to overcome now is being able to preview the camera AND record over 720p
03:16 finger563 i see; so with the OMX camera modules you can pass that data directly to the OMX core for encoding?
03:16 jvcleave I can record at 1080p but not preview at the same time
03:16 jvcleave yeah - the camera is OMX based
03:16 finger563 haha man, that's pretty high resolution
03:17 finger563 i'd be happy with this workinng at 720p :P
03:17 finger563 (not that i'd mind a resolution bump lol)
03:17 jvcleave here is a 1080p recording https://www.youtube.com/watch?v=RF6cPHo_yb0
03:18 finger563 man that is good quality
03:18 finger563 what's the fps range you get with it?
03:18 jvcleave there is a hardware limitation of 30fps in all resolutions of the camera - the encoder seems to be limited to 25fps
03:18 finger563 ah ok
03:19 finger563 i'm assuming you're using a model B?
03:19 finger563 (rev 2 or w/e)
03:20 jvcleave yeah - the 512 - don't think it matters tho
03:20 finger563 probably not
03:20 finger563 though trying to do the full record while preview might be limited based on your RAM allocated to gpu
03:20 jvcleave yeah - that doesn't seem to matter much
03:21 finger563 well that's good ;)
03:21 jvcleave the lowest I go is 128 but usually develop on 256
03:21 finger563 yea that's what i run at
03:36 finger563 thanks for pointing me to this mjpeg encoding though
03:36 finger563 hopefully i can get that working with my image data and it will be *faster*
03:36 finger563 :)
03:36 finger563 it's good for me too cause I don't care about preview, since I'll be sending the image data over wifi anyway
03:36 jvcleave yeah - good luck - would love to see it happen :)
03:36 finger563 haha thanks :)
03:38 kureta_m joined #openframeworks
03:51 finger563_ joined #openframeworks
03:55 finger563 joined #openframeworks
04:02 boltR joined #openframeworks
04:11 finger563 looking through this omx encoding sample code, it doesn't seem like it will be too hard to adapt it into a thread that writes videos of every N frames
04:11 finger563 the only downside is i cant seem to get the sample code to work
04:12 finger563 having built it the output video seems to have nothing
04:12 finger563 (rather, it doesn't want to play)
04:12 jvcleave you have to compile the llclient libs if you haven't
04:13 finger563 yea i ran the build.sh which recompiles the libs and the sample apps
04:13 finger563 and it creates a video file according to argv[1]
04:13 jvcleave oh - thats new :)
04:13 finger563 but i can't seem to get vlc to play it
04:13 jvcleave ah - change the name to .h264
04:13 finger563 haha go figure
04:14 finger563 i read i should use .mp4
04:14 finger563 :P
04:14 jvcleave yeah - i just ran into that this week - now familar with what an h264 stream is
04:14 finger563 nice
04:15 jvcleave I am now doing a post process using mkvmerge - that quickly writes a wrapper around it
04:15 finger563 so you can use .mp4 or why?
04:15 jvcleave omxplayer yourfilename <-will work as well
04:15 jvcleave its not the same thing
04:15 finger563 ah thanks, good to know
04:16 finger563 (yea i'm quite new to image/video encoding/decoding)
04:16 jvcleave my ofxOMXPlayer addon used to be able to play them back as well but I removed all that code as it required abunch of conditionals
04:17 finger563 ah ok
04:17 jvcleave I may have been trying to put it back but it was more difficult than I expected
04:17 finger563 question: are you still writing using YUV or have you switched to writing a different type of image?
04:18 finger563 cause i think that for the userland stuff YUV from camera was faster, but i'm not sure if that's still the case for OMX
04:18 jvcleave i think it's still YUV as I am not changing it
04:22 jvcleave I think these are the options
04:23 jvcleave https://github.com/jvcleave/ofxRPiCameraV​ideoGrabber/blob/develop/addons/ofxRPiCam​eraVideoGrabber/src/TextureEngine.cpp#L55
04:25 finger563 yea that seems right
04:26 finger563 i just wondered cause the data i'm getting in my shaders is already coded rgb
04:26 finger563 so i wasnt sure (hadn't checked) what format you're telling OMX to provide the camera data in
04:26 jvcleave that is set up in the omx component
04:26 jvcleave so - yes - with textures it is RGB
04:27 finger563 which is good, since it means any extra processing required for the conversion (if it's required) is quite minimal
04:55 jvcleave just tried forcing it to OMX_COLOR_Format24bitRGB888 but it's failing - not sure what options are compatible
05:00 finger563 yea i was worried about that; which options would be supported
05:01 finger563 i'm currently writing my image circular buffer code for my thread, hopefully in a bit or tomorrow i can test out writing
05:01 jvcleave the inputs have to line up with the outputs - some of it is self-configuring which is what I try and change as few params as possible
05:01 finger563 yea that makes sense
05:02 finger563 worst case scenario i'll have my shaders output YUV had YUV passed from OMX
05:02 finger563 but i don't think that'll be necessary
05:04 finger563 for your recording, do you care which colorspace it's in?
05:05 kureta_m joined #openframeworks
05:05 jvcleave I don't specify the colorspace
05:06 jvcleave fyi - beware of this bug - it still applies https://github.com/openframeworks/openFra​meworks/issues/2593#issuecomment-26657402
05:08 finger563 ah thanks, thats good to know
05:08 finger563 thankfully i'm doing all my shader code on an fbo inside of update, so i should be clear of that if i understand correctly
05:08 jvcleave yes - you lucked out :)
05:09 finger563 though i'm not sure I understand why you'd *need* to use shader code inside of draw instead of update
05:10 jvcleave so people prefer it that way - it does work on other platforms
05:11 finger563 haha yea i figured it was something along those lines
05:46 finger563 so for your camera copy code, you're creating a tunnel to pass the omx camera data directly to the omx encoder?
05:46 finger563 so that it doesn't leave the gpu?
06:09 LCID_Fire joined #openframeworks
06:34 oneirona1t joined #openframeworks
08:41 LCID_Fire joined #openframeworks
08:44 ncls joined #openframeworks
08:49 nylki_ joined #openframeworks
08:55 ofarturo joined #openframeworks
09:55 neue joined #openframeworks
10:08 nylki_ joined #openframeworks
10:40 robot joined #openframeworks
11:16 robot joined #openframeworks
11:35 boltR joined #openframeworks
12:39 nylki_ joined #openframeworks
12:39 kureta_m_ joined #openframeworks
13:21 Pando_ hello
13:21 Pando_ does anyone here know how to do mouse picking in a 3D scene with oF ?
13:22 Pando_ I've already seen some examples but it seems pretty hard
13:23 Pando_ I've tried to do like in the pointPicker example, but for some reason the results are kind of random, whereas it works well in the example ...
13:34 Pando_ actually, I'm still working on my solar system. So I have planets which are orbiting around the sun and rotating on themselves
13:36 Pando_ I think that's why the pointPicker method (using ofEasyCam::worldToScreen with the meshes vertices) won't work, vertices are always moving. And moreover the planets are textured so I can't really do color recognition
13:39 Pando_ I wonder if that's even possible at all in this case. There's no such example on the web, all are very simple to demonstrate the concept
13:55 LCID_Fire joined #openframeworks
14:13 robot joined #openframeworks
14:17 jacres joined #openframeworks
14:24 ofarturo _Pando if you are using ofNodes for each planet then you can multiply the center of the planet by the it's global matrix before applying worldToScreen and that should give you the position in screen coords of the center of the planet, you can do the same for each vertex but i would just use a distance to the center or something like that so it's not so computationally expensive
14:24 ofarturo Pando_
14:24 Pando_ hi ofarturo
14:24 Pando_ yes my planets are all ofSpherePrimitives
14:26 Pando_ so I have to retrieve the matrix with getGlobalTransformMatrix() ?
14:26 ofarturo yes
14:27 ofarturo then multiply the point by that and it'll give you it's transformed positon in 3d
14:27 ofarturo like
14:27 ofarturo planet.center * planet.getGlobalTransformMatrix()
14:27 ofarturo where center is mostly sure 0,0,0, right?
14:28 ofarturo the order of multiplication matters i think that's correct but i always have to check :)
14:29 Pando_ ok, actually I first tried to do it with the planets center, but since I didn't use the matrix, it didn't work either
14:29 Pando_ gonna try that
14:30 Pando_ well at the moment the sun's center is at 0,0,0, and all other planets are around it
14:33 ofarturo yes but i guess they are around it by using ofNode.orbit or something similar no? the original position of the ofSpherePrimitive is 0,0,0?
14:39 Pando_ hm... actually no, they orbit around the sun on the X axis, so I put all of them at a different distance on X towards the sun
14:39 Pando_ and then I use rotate to make them rotate on themselves, and rotateAround to make them orbit the sun
14:40 Pando_ at first I tried to use orbit, but it didn't work, so I tried something else. but maybe I was not using it well
14:41 Pando_ also, in some examples, people set the mouse coordinates' Z to ofCamera::getImagePlaneDistance. is it important ?
14:49 lukaszw joined #openframeworks
14:55 neue_ joined #openframeworks
14:57 ofarturo Pando_ i think as long as you don't have camera offset it's ok to use 0 but not sure
14:59 laserpilot joined #openframeworks
15:27 admsyn joined #openframeworks
15:31 Pando_ ofarturo: even after multiplication with the matrix, doesn't seem to work
15:32 Pando_ the planet reported the nearest is never the one the mouse is on :(
15:33 dantheman joined #openframeworks
15:33 Pando_ oops, did a little mistake
15:36 ofarturo also you can probably just use getPosition() which should return the transformed position and then use worldToScreen
15:37 Pando_ well that's what I already used... I didn't know it returned the already transformed position
15:38 lukaszw joined #openframeworks
15:38 Pando_ but, actually, what I do is something like
15:38 Pando_ ofVec3f pos = cam.worldToScreen(planet.g​etGlobalTransformMatrix() * planet.getPosition());
15:41 Pando_ hm... also, it maybe comes from the fact that my application uses multiple cameras, so I have an array of viewports, and maybe I must explicitly set the viewport of worldToScreen to  my main viewport
15:42 ofarturo no, the position you get from the node should be already transformed so you don't need to multiply it by the matrix
15:42 ofarturo ofVec3f pos = cam.worldToScreen(planet.getPosition());
15:42 ofarturo should do
15:42 ofarturo for the vertices is different cause they only get transformed once in the gpu
15:43 ofarturo and yes you need to set the viewport if you are using other than the full screen
15:43 Pando_ well, that's what I tried before, but it wasn't working either then
15:43 Pando_ yep, gonna try that
15:45 Pando_ I also have to check if the click actually was on a planet or just near it. I think I'm going to compare the nearest distance obtained and the radius of the nearest sphere
15:46 Pando_ if the distance is superior than the sphere radius, then it was not on it
15:52 Pando_ yes, seems to work ! :-)
15:52 Pando_ not very accurate at the moment, I'm going to have to do some more checks
15:53 Pando_ thanks for the help ofarturo ! apparently, the only problem here was just that I didn't pass my main viewport to worldToScreen
15:56 Pando_ maybe I'll try to use orbit instead of rotateAround for the planets orbits. but I'm not sure what to set for longitude/latitude
15:59 Pando_ isn't it for geostationary orbit ?
16:03 ofarturo both are just degrees but i remember having some problems with orbit and that particular example so perhaps better to stick to some displacement in x and then rotate
16:04 Pando_ ok, I guess that you just update longitude/latitude in the app update function...
16:04 Pando_ maybe I'll give it another try. but it works well at the moment, so...
16:05 Pando_ (my orbits are not realistic, they are all circular)
16:08 Pando_ I was also wondering how to make the planets cast shadows on others. but as I understood it, the default light system doesn't compute that for you
16:08 Pando_ maybe have to use a shader
16:16 dantheman joined #openframeworks
16:30 lukaszw joined #openframeworks
16:46 finger563 joined #openframeworks
16:53 ncls joined #openframeworks
17:20 dantheman joined #openframeworks
17:25 nylki_ joined #openframeworks
17:41 boltR joined #openframeworks
17:46 roxlu joined #openframeworks
17:49 tgfrerer joined #openframeworks
17:50 pizthewiz joined #openframeworks
18:03 LCID_Fire joined #openframeworks
18:19 pizthewiz joined #openframeworks
18:20 nicedice joined #openframeworks
18:52 lukaszw joined #openframeworks
18:56 dantheman joined #openframeworks
19:36 starPause joined #openframeworks
19:39 kureta_m joined #openframeworks
19:45 dantheman joined #openframeworks
19:50 pizthewiz joined #openframeworks
20:03 dantheman joined #openframeworks
21:26 robot joined #openframeworks
21:39 ofarturo joined #openframeworks
21:47 bert_ joined #openframeworks
21:57 nicedice joined #openframeworks
22:24 Razorien joined #openframeworks
22:47 GuidovanPossum joined #openframeworks
23:27 nylki joined #openframeworks
23:39 GuidovanPossum joined #openframeworks
23:57 Razorien joined #openframeworks

| Channels | #openframeworks index | Today | | Search | Google Search | Plain-Text | summary