GEM-dev March 2008

gem-dev@lists.iem.at

12 participants
16 discussions

2textures + gemframebuffer
by marius schebella 02 Apr '08

02 Apr '08

hi, I can't get gemframebuffer working with 2 shaders. see my files here: http://www.parasitaere-kapazitaeten.net/files/2tex+gemframe.zip I would like to take the output of a shader with 2 textures and use it in another pix_texture. but I don't find the right way to do this... anybody with help? thanks over'n out (no more mail today) marius.

4 52

[ pd-gem-Feature Requests-1930037 ] texunit to gemframbuffer
by SourceForge.net 31 Mar '08

31 Mar '08

Feature Requests item #1930037, was opened at 2008-03-31 10:04 Message generated for change (Tracker Item Submitted) made by Item Submitter You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=507082&aid=1930037&group_… Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: None Group: None Status: Open Priority: 5 Private: No Submitted By: Cyrille Henry (nusmuk) Assigned to: Nobody/Anonymous (nobody) Summary: texunit to gemframbuffer Initial Comment: I would be very handly to add a texunit message to gemframebuffer, just like pix_texture. This could make patch using multi shader less complex, and easier to understand. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=507082&aid=1930037&group_…

1 0

gl: stack overflow
by marius schebella 30 Mar '08

30 Mar '08

hi, I am sending texunit 1 to a texture and get a stack underflow/stack overflow error and don't know why. can someone test this and see if it also happens on another machin? or is this only with my gemversion? please see the attached patch for details. marius.

1 0

texunit 8 (bug)
by marius schebella 29 Mar '08

29 Mar '08

hi, I wonder if the limitation to 8 texunits is correct. I know that there is a limitation of 8 textures per glsl shader, but I can load the same shader 9 times and have 9 different textures (one for each shader). if I send a texunit 0-7 to pix_texture that only allows me to reference 8 different textures total. I think it should work like Mytex 17 where 17 is the texture ID that is spit out of the right outlet of a pix_texture object. with the texunit method I am running into bad limitations. any solution for that? marius.

1 0

shader test
by marius schebella 29 Mar '08

29 Mar '08

hi, I figured out a more extreme way to chain shaders and would like to hear your comments on this. http://www.parasitaere-kapazitaeten.net/files/prerelease.zip there are 3 glsl.* abstractions. they don't pass gemstates, but only ID numbers of textures. not intuitive to use, but very convenient. please tell me also if glsl is an appropriate name for such abstractions. I don't know yet if this will work with any aspect ratio and I still have to figure out how to support multiple textures, but not now. oh, and thanks to vade for the v001 shaders. (more to come...) marius.

1 0

vertex objects
by marius schebella 25 Mar '08

25 Mar '08

hi, there is a vertex folder within gem src (hans pointed me to it), and it compiles, but I am not sure if the objects are working? I can create them, but handling and syntax seems not quite clear to me. is this in development? I tried [gemhead] | [vertex_grid] | [vertex_draw] which gave me a white square in the right top corner and [gemhead] | [vertex_model] | [vertex_draw] which did not draw anything. marius.

1 0

interlaced output from Gem?
by Claude Heiland-Allen 12 Mar '08

12 Mar '08

Hi all, Does Gem (or OpenGL in general) have any support for interlaced output? The extra temporal resolution would be nice when rendering for display on a TV, as the spatial resolution would suck anyway. Here's an IRC log from #dataflow : 09:56 < ClaudiusMaximus> anyone got any tips for generating interlaced video in Gem? i think i need to set the gemwin to 2x the framerate and 1/2x the vertical resolution, and do some small vertical translations up/down depending on which field is the current one 09:56 < ClaudiusMaximus> it's the last step that i'm not sure exactly how to do 09:57 < ClaudiusMaximus> rather i know how, but just not what values i should put in.. 10:29 < rdz> ClaudiusMaximus, i once thought about the same.. 10:29 < rdz> ClaudiusMaximus, actually it would suffice to have only half of the vertical resolution (the scan lines) 10:30 < ClaudiusMaximus> rdz: i'm not sure, because the different fields are displayed at slightly different vertical positions.. 10:30 < rdz> ClaudiusMaximus, yeah exactly.. 10:30 < rdz> ClaudiusMaximus, this could be accomplished by shifting half a line 10:32 < ClaudiusMaximus> rdz: what about OpenGL line widths being measured in pixels? wouldn't that cause problems like "fat horizontal lines"? 10:35 < rdz> ClaudiusMaximus, sorry.. i didn't understandyour last line 10:35 < rdz> what do you mean by fat lines? 10:36 < ClaudiusMaximus> rdz: well, if you do "draw line, width 1" you get a 1pixel line, right 10:36 < rdz> yeah 10:36 < rdz> understood 10:36 < ClaudiusMaximus> and 1pixel at half the vertical resolution is 2x as fat as 1pixel at the full horizontal resolution 10:36 < rdz> hm... 10:36 < rdz> yo........ 10:36 < rdz> seems like there should be native support for that in gem/openGL 10:38 < ClaudiusMaximus> i'm currently leaning towards rendering 720x576 @ 50fps and using a field-aware downscaler 10:39 < ClaudiusMaximus> especially as i want to blur everything slightly to avoid horrible compression artifacts 10:39 * ClaudiusMaximus been reading this page http://www.adamwilt.com/Tidbits.html#CGs 10:40 < rdz> ClaudiusMaximus, are you producing for tv ? ;-) 10:40 < ClaudiusMaximus> rdz: well, i want to make any dvds i produce look better than they do currently... 10:40 < rdz> ClaudiusMaximus, i see 10:40 < ClaudiusMaximus> rdz: so that's a sorta-yes ;) 10:40 < rdz> ClaudiusMaximus, have you already asked gem-dev about it? 10:41 < rdz> ClaudiusMaximus, interlaced 3d renderings....... sounds very promising... 10:41 < rdz> i'd love to see a way to be able to do that 10:41 < ClaudiusMaximus> rdz: not yet, will post something today I think Thanks, Claude -- http://claudiusmaximus.goto10.org

3 3

Re: [GEM-dev] [ pd-gem-Bugs-1771966 ] texturing movies pegs CPU using fglrx/ATI Radeon X1x00
by Roman Haefeli 11 Mar '08

11 Mar '08

yo.. i updated and compiled Gem with NO configure flags (such as --with-glversion) nor did i delete any files like Manips/glsl_* compiling went fine. loading was no problem. i tested [pix_film] with following files: alea.mpg anim-1.avi homer.avi http://romanhaefeli.net/pix_film-test.mov all of them worked fine. at least for me (and probably for other ati users?) things have definitely improved since the switch to GLEW. thank you very much for the efforts you put into making Gem work better. roman On Wed, 2008-03-05 at 11:49 -0800, SourceForge.net wrote: > Bugs item #1771966, was opened at 2007-08-10 22:08 > Message generated for change (Settings changed) made by zmoelnig > You can respond by visiting: > https://sourceforge.net/tracker/?func=detail&atid=507079&aid=1771966&group_… > > Please note that this message will contain a full copy of the comment thread, > including the initial issue submission, for this request, > not just the latest update. > Category: Pixes (pix_ objects) > Group: None > >Status: Pending > >Resolution: Fixed > Priority: 8 > Private: No > Submitted By: Hans-Christoph Steiner (eighthave) > Assigned to: Nobody/Anonymous (nobody) > Summary: texturing movies pegs CPU using fglrx/ATI Radeon X1x00 > > Initial Comment: > > Whenever loading a movie with pix_movie on specific machines with an ATI, the CPU gets pegged. The example patch has notes on to isolate when it happens. It happened before and after the [mode 0( bug was fixed. > > root@puesto18:~# pd -stderr -nogui > GEM: Graphics Environment for Multimedia > GEM: ver: 0.91-cvs > GEM: compiled: Aug 5 2007 > > It is compiled like this: > > ./configure --without-ffmpeg --disable-NV --with-glversion=1.5 > > "ATI Technologies Inc RV515 [Radeon X1300]" with Ubuntu binary driver > > > It works fine on these machines: > > > Ubuntu Studio Feisty > "nVidia Corporation G70 [GeForce 7600 GS] (rev a1) > Ubuntu binary driver > > Mac OS X 10.4.10 > MacBook Pro > ATY,RadeonX1600 > > > > > > ---------------------------------------------------------------------- > > >Comment By: IOhannes m zmlnig (zmoelnig) > Date: 2008-03-05 20:49 > > Message: > Logged In: YES > user_id=564396 > Originator: NO > > this bug might be fixed in the current CVS > (at least i don't get the CPU-pegging on re-load of a movie like i used > to) > especially with romans test-movie i now have acceptable results. > > it still takes quite a lot of CPU here (compared to other hardware i am > using), but this might well be related to my test laptop's memory bus > > > could anyone please test. > > > ---------------------------------------------------------------------- > > Comment By: IOhannes m zmlnig (zmoelnig) > Date: 2007-10-26 11:18 > > Message: > Logged In: YES > user_id=564396 > Originator: NO > > confirmed with ati-x2300 und ubuntu/feisty (using fglrx drivers) > > OpenGL info > Vendor: ATI Technologies Inc. > Renderer: ATI Mobility Radeon X2300 > Version: 2.0.6334 (8.34.8) > > > ---------------------------------------------------------------------- > > Comment By: Roman Haefeli (reduzent) > Date: 2007-10-08 05:36 > > Message: > Logged In: YES > user_id=1895440 > Originator: NO > > unfortunately, the 'create, 1, 0, 1' does not always work. i am working on > a patch,where it doesn't. though, in pix_film-help.pd the trick works. > couldn't find the reason of the different behaviour. > > > ---------------------------------------------------------------------- > > Comment By: Roman Haefeli (reduzent) > Date: 2007-10-08 05:33 > > Message: > Logged In: YES > user_id=1895440 > Originator: NO > > the cpu hunger seems to correlate (proportionally?) with the displayed > size of the geo, that is textured with the movie. > > ---------------------------------------------------------------------- > > Comment By: Roman Haefeli (reduzent) > Date: 2007-09-29 07:03 > > Message: > Logged In: YES > user_id=1895440 > Originator: NO > > yo.. i could find a difference between anim-1.mov and my selfencoded movie > files, that seem to trigger the cpu boost: > > anim-1.mov uses colorspace: Planar 422P > > my encodes use: Planar YV12 > > i tried to encode a movie with mencoder, that was also using colorspace > 'Planar 422P', but it seems, that it is not possible with mencoder. > > ---------------------------------------------------------------------- > > Comment By: Roman Haefeli (reduzent) > Date: 2007-09-29 05:24 > > Message: > Logged In: YES > user_id=1895440 > Originator: NO > > i investigated this a bit further. i made some test, which hopefully help > locate the bug. > > a) loading pix_film-help, loading a mov-file [1], 'create, 1' to gemwin -> > 100% cpu / very slow > > b) same, but sending 'create 1, 0, 1' instead -> no cpu boost / plays > fine. it seems, that only the first time rendering is turned on, the > problem can be triggered. > > c) test only [gemhead]-[pix_film] (without [pix_texture]-[pix_rectangle]) > -> no cpu boost / cpu consumption is similar to case b) > > d) i noticed, that when cpu is boosted, the texturing is done wrong / > image is stretched [2]. when playing fine, also the image is textured > correctly [3]. > > [1] http://romanhaefeli.net/pix_film-test.mov > [2] http://romanhaefeli.net/pix_film_test_wrong.png > [3] http://romanhaefeli.net/pix_film_test_correct.png > > > ---------------------------------------------------------------------- > > Comment By: Roman Haefeli (reduzent) > Date: 2007-09-21 06:18 > > Message: > Logged In: YES > user_id=1895440 > Originator: NO > > similar behaviour on my machine: > > specs: > GEM: ver: 0.91-cvs > GEM: compiled: Sep 20 2007 > pd 0.40.2 > ubuntu dapper > fglrx 7.0.0 > > fglrxinfo: > OpenGL vendor string: ATI Technologies Inc. > OpenGL renderer string: MOBILITY RADEON 9600 Generic > OpenGL version string: 2.0.5814 (8.25.18) > > what i tested with pix_film-help.pd: > - anim-1.mov seems to work fine > - homer.avi and alea.mpg peg the CPU and run very slowly. > - any other selfmade mov was also slow and ate my cpu. > > > > ---------------------------------------------------------------------- > > Comment By: Hans-Christoph Steiner (eighthave) > Date: 2007-08-29 18:08 > > Message: > Logged In: YES > user_id=27104 > Originator: YES > > Here's the lspci string for my card: > > ATI Technologies Inc M56P [Radeon Mobility X1600] > > > ---------------------------------------------------------------------- > > Comment By: Hans-Christoph Steiner (eighthave) > Date: 2007-08-29 18:06 > > Message: > Logged In: YES > user_id=27104 > Originator: YES > > > Ok, another machine that's affected (my machine actually, so I can test > things for you). Here's fglrxinfo: > > OpenGL vendor string: ATI Technologies Inc. > OpenGL renderer string: ATI Mobility Radeon X1600 > OpenGL version string: 2.0.6334 (8.34.8) > > I am running Ubuntu Studio with the binary drivers from Ubuntu. > anim-1.mov works fine, homer.avi and alea.mpg both trigger the problem. > The texturing works fine after sending [mode 1( but it does not affect the > CPU pegging or slowness of video playback. > > > ---------------------------------------------------------------------- > > Comment By: Hans-Christoph Steiner (eighthave) > Date: 2007-08-28 19:59 > > Message: > Logged In: YES > user_id=27104 > Originator: YES > > > Here's Kubuntu machine that has the problem: > > olsen@ll-928:~$ fglrxinfo > display: :0.0 screen: 0 > OpenGL vendor string: ATI Technologies Inc. > OpenGL renderer string: ATI Mobility Radeon X1400 > OpenGL version string: 2.0.6334 (8.34.8) > > 04.pix/05.movie.pd > > - anim-1.mov seems to work fine > - homer.avi and alea.mpg peg the CPU and run very slowly. > > > Also, we tested using a much older ATI that was not supported by fglrx, > and it seemed to work fine. It seems the bug is related to fglrx. > > > > ---------------------------------------------------------------------- > > You can respond by visiting: > https://sourceforge.net/tracker/?func=detail&atid=507079&aid=1771966&group_… ___________________________________________________________ Der frühe Vogel fängt den Wurm. Hier gelangen Sie zum neuen Yahoo! Mail: http://mail.yahoo.de

3 6

GLEWified
by IOhannes m zmoelnig 10 Mar '08

10 Mar '08

so i since today Gem depends on glew, the "openGL Extension Wrangler", a runtime openGL extension checker. this basically means, that Gem now does runtime checks, whether your graphics card supports a certain feature (e.g. GLSL shaders) instead of making decisions at compiletime. this should magically solve all the problems people kept having with certain openGL-functions not being found when trying to load Gem. it should also make the "--disable-ARB", "--disable-NV" and "--with-glversion=" flags to configure obsolete. Gem has the sourcecode of glew included (so you don't need to install it separately) and links statically to it. for now, all (or most?) objects that had compile-time openGL checks should now have runtime checks (probably i have forgotten some, who knows) i have successfully compiled it on linux/debian and osx-10.4. i have successfully tested it on debian(lenny/sid) with nvidia (GeForce 7800 GTX; proprietary drivers v169) and on debian(etch) with ati (Rage Mobility P/M; old "ati" xorg drivers; mesa-6.5.1) please test! (especially specialized functions, like texture, glsl and othe shaders, framebuffer) mfga.ds IOhannes PS: i have for know purged pix_texture support for outdated openGL-versions (like openGL<1.1)

5 9

[ pd-gem-Bugs-1908218 ] texture-sharing crashes on ATI
by SourceForge.net 05 Mar '08

05 Mar '08

Bugs item #1908218, was opened at 2008-03-05 19:49 Message generated for change (Tracker Item Submitted) made by Item Submitter You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=507079&aid=1908218&group_… Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: Pixes (pix_ objects) Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: IOhannes m zmölnig (zmoelnig) Assigned to: Nobody/Anonymous (nobody) Summary: texture-sharing crashes on ATI Initial Comment: running any patch that involves sharing of textures between 2 [pix_texture] objects (e.g. 07.texture/09.sharedTextures.pd) crashes on here with a Radeon-X2300 on ubuntu. the error i get is: socket receive error: Connection reset by peer (104) Segmentation fault (core dumped) a backtrace in gdb reveals that it crashes in pix_texture::render() this is card info as reported by GLEW: Reporting capabilities of display :0.0, visual 0x2b Running on a ATI Mobility Radeon X2300 from ATI Technologies Inc. OpenGL version 2.0.6473 (8.37.6) is supported ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=507079&aid=1908218&group_…

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

GEM-dev March 2008