May 26, 2008 
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | >
May 26, 2008 [00:01:24] <JKnife> \o/ i did it[00:01:37] <JKnife> iz got my tirang;e[00:01:40] <JKnife> ...[00:01:43] <JKnife> triangle*[00:02:24] <JKnife> http://nopaste.com/p/aiiN7J2R8[00:02:46] <MatthiasM> you can also call begin() :DD[00:03:19] <JKnife> I like doing this the hard way?[00:03:49] <MatthiasM> the easy way would be to read the docu :)[00:04:59] *** scy has left ##opengl[00:07:06] *** juanmabc has joined ##opengl[00:16:19] *** stringfellow_ has quit IRC[00:18:38] <SANiK> Handy tip JKnife, for future use: separate your code into: "think()" and "render()"; while(program running){ think(); render(); }[00:19:04] <SANiK> I mean - you can separate things into threads if you want[00:19:21] <MatthiasM> and use a lib to get rid of the system dependant stuff[00:19:48] <SANiK> or a wrapper[00:20:26] <SANiK> I mean - I've separated my renderer into a .dll; thing is - if I want to port to the DS, what use is a .dll[00:20:44] <SANiK> So I had to remove the DLL hooks and code-in a wrapper[00:28:27] <JKnife> SANiK: >_> it is.. (if you notice SetUpVertexs and RenderLoop are different classes.. >_> I should rename RenderLoop to Render or Draw...[00:34:40] *** Xmas| has joined ##OpenGL[00:35:57] *** Ragnarok has joined ##opengl[00:36:10] <Ragnarok> can you turn on lighting through the GLSL[00:38:17] <MatthiasM> you can write equvalent code :)[00:42:49] <Walt> Makegho, why would you write a ray tracer on the CPU anyway O_O[00:43:11] <Walt> if it's only for graphical purpose[00:43:20] <Makegho> first this way of course[00:44:01] <Makegho> you suggest coding it with glsl or something? :)[00:44:10] <RTFM_FTW> sure[00:44:29] <RTFM_FTW> GPU accelerated RT is pretty fun[00:44:31] <Makegho> Why not, when it just works first :)[00:44:45] <Makegho> I don't know GLSL yet though, but I'll need to learn it anyway[00:44:57] <RTFM_FTW> the performance can be significantly better on consumer level hardware[00:44:58] <Walt> Cg all the way bh :)[00:45:01] <Walt> tbh*[00:45:13] <RTFM_FTW> that I'd disagree with[00:45:34] <RTFM_FTW> since GLSL is just fine for this... or for anything else[00:45:41] <Makegho> RTFM_FTW: sure... are not the modern GPUs like five times as powerful as the modern CPUs or something :)[00:45:46] <Walt> well, I love Cg[00:45:47] <Makegho> in simple floating point arithmetics[00:46:06] <RTFM_FTW> add another zero onto that and you are getting close[00:46:27] <Makegho> not that much probably :)[00:46:52] <RTFM_FTW> easily that much if the workload is right[00:47:09] <RTFM_FTW> if the workload isn't then performance will suck (no surprise here)[00:47:52] <RTFM_FTW> in any case having hundreds of ALUs at your disposal in the land of massively data independent / parallel tasks is always a good thing[00:48:10] <SANiK> Guys, ph33r: http://sanik.hacking-cult.org/ph33r.jpg[00:48:22] <SANiK> *plays terminator theme*[00:49:16] <MatthiasM> lol[00:49:18] <Makegho> sure[00:49:45] <Makegho> RTFM_FTW: a win-win move would be to use both cpu and gpu for rendering :)[00:49:51] <Makegho> I suppose[00:50:04] <RTFM_FTW> using whatever you have at your disposal is a "win win"[00:50:09] <Makegho> yep[00:50:24] <RTFM_FTW> in any case at the high end ASICs are rapidly approaching the 1 TFLOP mark[00:50:28] *** TheLorax has quit IRC[00:50:36] <RTFM_FTW> in terms of raw (and accessible) compute performance[00:51:10] <RTFM_FTW> and this will continue to go up so its a good idea to start trying to take advantage of it[00:51:23] <Makegho> heh[00:54:14] <Makegho> RTFM_FTW: "As of 2007, the fastest PC processors (quad-core) perform over 30 GFLOPS.[9] GPUs in PCs are considerably more powerful in pure FLOPS. For example, in the GeForce 8 Series the nVidia 8800 Ultra performs around 576 GFLOPS on 128 Processing elements." - http://en.wikipedia.org/wiki/FLOPS[00:54:37] <Makegho> So I was really wrong :) that's a huge difference![00:55:20] <RTFM_FTW> and the GF8 series is pretty low in the realm of raw compute[00:55:40] <SANiK> Funny how GPUs are like ... putting CPUs to shame[00:55:43] <RTFM_FTW> ATI's ASICs are much higher in that department[00:56:40] <RTFM_FTW> yes GPUs put CPUs to shame in this business because of the comments I made above related to the problems we deal with... "massively data independent / parallel workloads"[00:56:44] <RTFM_FTW> that is key[00:57:19] <RTFM_FTW> basically we see wonderful speedups in this business by simply throwing more and more ALUs (and hence compute threads) at our collective problem-sets[00:57:43] <RTFM_FTW> this is *not* the case for the vast majority of "typical CPU codes" out there'[00:58:27] <Makegho> So I suppose that real time raytracing is somewhat reality nowadays. :)[00:58:40] <Makegho> At least with some kind of graphics.[00:58:51] <SANiK> I read an article where they used the GPU to help with some calculations[00:59:21] <SANiK> e.g., the RGBA of the framebuffer contains an address of where the previous light came from - or something like that[01:00:06] <RTFM_FTW> which can help to give you an idea as to how the GPU and CPU approach these problems[01:00:30] <Makegho> thanks[01:01:23] <RTFM_FTW> note slides 11 and 12 specifically[01:06:16] *** kaotrix has joined ##OpenGL[01:07:29] <Makegho> good night[01:09:47] *** prophile has quit IRC[01:13:57] *** LordMetroid has joined ##OpenGL[01:19:36] *** Antibuddha has joined ##OpenGL[01:20:00] <Antibuddha> what is opengl[01:21:41] <HuntsMan> an API for graphics processing and display[01:22:01] *** Dew420 has joined ##OpenGL[01:24:15] *** charlie55 has joined ##OpenGL[01:28:42] *** Antibuddha has left ##OpenGL[01:30:09] <Ragnarok> I still can't find any decent tutorial on VBO. ones I keeep finding are too complex and cryptic[01:36:40] *** cupe has quit IRC[01:37:24] *** MatthiasM has quit IRC[01:37:30] *** MatthiasM has joined ##opengl[01:38:20] *** charlie5 has quit IRC[01:40:50] *** charlie55 is now known as charlie5[01:44:59] *** Burga has quit IRC[01:48:10] <SANiK> If it wasn't for DirectX - OpenGL should have been made part of ANSI-C - seriously =3[01:48:31] <SANiK> In my own little world, it is *sniff*[01:48:39] <SANiK> Right up there with sdtio.h and stdlib.h[01:53:27] <Tenac> Can OpenGL load and handle .x files as well as DirectX can?[01:53:43] <SANiK> Depends on how well you code the .x handler ;)[01:54:01] *** neoneye has quit IRC[01:54:02] <SANiK> GL's a graphics library - that's it - nothing more[01:54:05] <Tenac> has anyone done it before, seems like a pretty good idea[01:55:24] <Weiss> Ragnarok: the one in the Blue Book is concise and clear to me[02:04:52] *** elite01 has quit IRC[02:15:15] *** Tenac has left ##opengl[02:17:08] *** juanmabc has quit IRC[02:18:47] <Ragnarok> Weiss, what page iurc?[02:20:25] <Ragnarok> I miss DirectX a little[02:20:50] <Ragnarok> Why would u want to use *. files?[02:20:56] <Ragnarok> *.x*[02:22:04] *** florian has quit IRC[02:22:05] <Ragnarok> The only reason I liked DirectX, was the simplicity of it being C++ based and OpenGL is C *shudders*[02:22:39] <HuntsMan> well i don't think DirectX is simple[02:22:44] <charlie5> hi folks ... i've been using C unsigned's for indices ... would there be much advantage in using unsigne short's, when possible ?[02:23:01] <HuntsMan> just those all uppercase class names and macro namings make me yell[02:23:40] <Ragnarok> HuntsMan, you gotta do that with extensions if u don't use glee or glew[02:23:52] <HuntsMan> do what?[02:24:02] <Ragnarok> uppercase extension names[02:24:27] <HuntsMan> that's on Windows[02:24:41] <Ragnarok> you sure not on *nis or mac?[02:24:45] <Ragnarok> also*[02:24:55] <Ragnarok> nix* goddamn can't type correctly[02:25:26] <Ragnarok> and also, searching the documentation on DirectX is a pain also[02:25:34] <HuntsMan> on UNIX and Mac you don't need to use extension pointers[02:25:45] <Ragnarok> did not know that[02:25:58] <HuntsMan> you need to be careful only to what you link against with[02:26:21] <Ragnarok> one good thing about OpenGL. I can run 2.1 on nVidia 6150 LE xD[02:26:30] <HuntsMan> for example, libGL.so provides 2.1 and below entry points[02:26:34] <Ragnarok> can't with DirectX[02:27:19] <HuntsMan> yeah[02:27:32] <HuntsMan> well i can use SM 4.0 features without using Windows Vista :D[02:27:33] *** dvoid__ has quit IRC[02:27:39] <Ragnarok> OpenGL iirc version 3 will run on Windows XP (for those that hate Vista which I disgree but anyways)[02:28:11] <Ragnarok> You see the funniest things working at a computer repair store[02:28:19] <Ragnarok> People still use windows 98 and me[02:28:39] <HuntsMan> OpenGL runs on vista also, is'nt specific to a Windows version BTW[02:28:48] <Ragnarok> yeah[02:28:59] <Ragnarok> And people having more problems with XP than Vista[02:30:37] <Ragnarok> I use Linux so im happy. The only thing I hate is that not many games for Linux, and the ones that are available are garbage[02:31:14] <HuntsMan> you call ETQW garbage? :P[02:31:25] <Ragnarok> can't play it xD[02:31:32] <HuntsMan> i can :P[02:34:42] <hibread> Ragnarok: thats a bit harsh. Those guys spent millions of hours making those games and for what? For someone like you to come along and call them garbage :)[02:34:54] <HuntsMan> yeah[02:35:40] <HuntsMan> i'm gonna buy some games from the LGP[02:35:58] <Ragnarok> hibread, im talking about the user made ones you never will buy on the streets(stores)[02:36:00] <hibread> although maybe its a context of what you class as garbage. Someones garbage is someone elses treasure[02:36:17] <Ragnarok> Nexiuz sucks, Treminious sucks[02:36:36] <hibread> :) just having a dig for no real reason[02:36:56] <Ragnarok> One mans pain is the next mans pleasure[02:37:11] <Ragnarok> one says infinity the next says forever[02:37:12] <Ragnarok> xD[02:37:20] <Ragnarok> its off a song xD[02:38:29] <LordMetroid> How much is a new PC game in USA?[02:38:44] <hibread> 100 million euros[02:38:50] <Ragnarok> $49?[02:38:58] <LordMetroid> hm[02:39:03] <RTFM_FTW> mmm ETQW is awesome[02:39:10] <Ragnarok> i hate u guys[02:39:17] <Ragnarok> I need a beetter graphics card[02:39:23] * Ragnarok loves UT 2004[02:39:30] <hibread> 49$ ? jeeez. Maybe ill start importing games since the us -> aus dolar is nearly 1:1 these days[02:39:37] <hibread> games are like 90-100 here for pc i think[02:39:47] <Ragnarok> hey is UT3 really that bad?[02:39:48] <LordMetroid> $49 is way cheaper than Swedish prices :([02:41:14] <Ragnarok> LordHavoc, go pirate then xD[02:44:14] <hibread> Im sure LordHavoc has more heart than that[02:45:11] <Ragnarok> lol[02:45:38] <Ragnarok> i bet everyone pirated when napster first came around[02:45:50] <hibread> and then metallica had a good point?[02:45:59] <Ragnarok> they did?[02:46:04] <hibread> or rather, Lars Ulric[02:46:19] * Ragnarok shrugs[02:46:42] <Ragnarok> i know someone who pirates songs then when his paycheck comes around he buys the cd's[02:47:06] <hibread> i think its pretty funny. Metallica were the big guns fighting napster. But now it seems they may be joining the bandwaggon for free downloads online like who was it.. radio head (and others i assume) ?[02:47:13] *** Xmas| has quit IRC[02:47:32] <Ragnarok> iono i don't pay attention to the news[02:47:39] <Ragnarok> im too busy on my computer[02:48:27] *** dv_ has quit IRC[02:49:14] *** dv_ has joined ##OpenGL[02:50:02] <Ragnarok> whats a good way to pass a custom vector to a glVertex(). instead of glVertex3f(x.x, x.y, x.z); etc[02:50:50] <hibread> glVertex3fv ?[02:51:25] <hibread> glVertex3fv( &x);[02:52:13] *** LordMetroid has quit IRC[02:52:28] <Ragnarok> so like Vec3<float> x(0.0f, 0.0f, 0.0f); glVertex3fv(&x)l; ah okay[02:53:33] <hibread> Ragnarok: thats assuming that the memory consumed by a Vec3 object doesn't have any other data members[02:54:01] <Ragnarok> yeah it only have T x, y, z has its private members[02:54:15] <Ragnarok> well proteected[02:54:53] <hibread> err. protected and private you'd need a get() function...[02:55:46] <Ragnarok> heres the code I know it has some errors at places haven't fixed it yet[02:55:47] <Ragnarok> http://rafb.net/p/BAVKpg67.html[02:57:56] <hibread> hmmm[02:58:07] *** dv_ has quit IRC[02:58:39] <hibread> whats your reasoning behind Vec3 inheriting from Vec2... what benifit do you think that offers?[02:59:01] <hibread> merely a question, i dont have a reason why its "bad" yet[02:59:35] <hibread> and the data member is public by the way[02:59:40] *** replor has quit IRC[02:59:47] <Ragnarok> inheritance. Vec3 is a is relationship with Vec2.[02:59:52] <Ragnarok> yeah[03:01:06] <Ragnarok> so I can easily do like Vec3<float> x(...) Vec2<float>(x) even though is loses the z info[03:01:12] <Ragnarok> it*[03:01:46] <Ragnarok> if u don't think it should i can rewrite it[03:02:29] <Ragnarok> the only thing is I haven't figured out a way to optimize those Vec classes yet[03:04:08] <hibread> you could awesome create a Vec3 constructor that takes a Vec2, extracting out the elements[03:04:16] <Ragnarok> awesome?[03:05:41] <hibread> hmm, but im not sure how that works with templates... Should be the same deal?[03:06:00] <Ragnarok> yeah its the same, nothing differeent[03:07:20] <Ragnarok> im trying to figure out which isn't is needed to make it possible to use Matrix multi array with OpenGL.[03:08:02] <Ragnarok> extension*[03:08:54] <hibread> I'd personally be tempted to not worry about the inheritance heirachy between say vec4, vec3 and vec2. You don't appear to be using any functions up the tree to assist with clearing up details etc.. But again, im no expert[03:10:37] <hibread> but you'd best get a real opinion from someone on ##c++. They'll surely have a better one[03:15:59] <Ragnarok> says I should do template<float, dimension> x[03:19:21] <hibread> Ragnarok: hmm. How would that work with functions that work on explicitly 1, 2, 3, 4 number of memebers? Normalize() for instance? Need a switch or something?[03:20:02] <Ragnarok> i guess arrays?[03:20:58] <hibread> To me, this is getting too complex with no reward that i can see. I'd just use the template for type, and explicitly define the vec2, vec3, vec4, mat4x3, mat3x3, mat4x4 or what ever you need[03:21:09] <hibread> without inheritance[03:21:25] <hibread> if you need a conversion between vec4 and vec2.. create a constructor[03:21:32] <Ragnarok> okay[03:21:38] *** Amorphous has quit IRC[03:21:47] <Ragnarok> i wish i could benchmark this to see the speed difference[03:21:57] <Ragnarok> but i don't know how to do that[03:22:02] <hibread> there'll be no speed difference at all[03:23:02] *** Amorphous has joined ##opengl[03:23:16] <Ragnarok> I want to use double over float for double precision, but I know it has a performance impact but how much?[03:24:00] <hibread> thats beside the point dealing with this inheritance business. So where does you real question lie with performance? Lies with the inheritance or?[03:24:24] *** m4ggus has joined ##opengl[03:24:24] <Ragnarok> yeah i wanna try and gain better performance because thats critical in this stuff[03:24:40] <Ragnarok> thats why I went to templates and reduce reduntacy(sp?)[03:25:47] <hibread> if you use inheritance with virtual functions you'll decrease performance. Inheritance how you've got now would be probably identical to not using inheritance since the compiler would just merge the two object types to create one anyway (along those lines i believe)[03:25:54] <hibread> Ragnarok: nah it doesn't[03:26:28] <hibread> if ouy create an object of type Vec3<float>, it'll allocate space for 3x sizeof(float)[03:26:31] <Ragnarok> http://www.flipcode.com/archives/Faster_Vector_Math_Using_Templates.shtml[03:26:54] <hibread> well looks like im wrong then, ill have read[03:27:49] <Ragnarok> :)[03:27:57] <Ragnarok> the way he has it, its cryptic XD[03:28:22] <hibread> whats the over view of what he's doing here? How's he gaining performance?[03:28:45] <Ragnarok> using 6-8 classes[03:35:51] <hibread> Ragnarok: this article is back in 2001... is there any chance that modern compilers could overcome these limitations?[03:37:35] <Ragnarok> i can't compile it, because i think it uses some windows specific stuff[03:37:37] <Ragnarok> u can try[03:38:21] <Ragnarok> http://rafb.net/p/PQrVIN32.html[03:38:28] <Ragnarok> heres a sample of the new class[03:39:04] <hibread> nah dont worry about testing the implimentation at this stage. I just ask whether compilers these days might be able to add this level of optimization without having to deal with templates etc. Interesting read though, ill keep reading[03:39:42] <Ragnarok> hehe[03:39:48] <Ragnarok> what do u use? C?[03:39:53] <hibread> ++[03:40:00] <Ragnarok> ah okay[03:40:05] *** Ademan has joined ##OpenGL[03:40:17] <Ragnarok> i wish SGI would have a C++ version of opengl with a C version. it should be that hard[03:40:43] <Ragnarok> shouldn't*[03:40:44] <hibread> why does it matter?[03:41:00] <SANiK> I always found it idiotic to call gl.vertex3f...[03:41:10] <SANiK> glVertex3f FTW[03:41:40] <Ragnarok> because u can achieve much more stuff, there would be no more need for glVertex2[d,v,fv] stuff[03:42:12] <Ragnarok> SANiK, u can have operator overload, and plus C teaches bad habits[03:42:16] <hibread> Ragnarok: you shouldn't be using that function anyway[03:42:27] <Ragnarok> hibread, i know[03:42:33] <Ragnarok> im just saying[03:42:34] <SANiK> Ragnarok, sec - let me post your statement in #C =3[03:42:44] <Ragnarok> :D[03:42:50] <Ragnarok> it does teach bad habits[03:43:24] <hibread> Read "Effective C++" Ragnarok?[03:43:36] <Ragnarok> don't have the money :([03:43:52] <hibread> fair enough. It's alittle old by thelooks, but its a very good book imo[03:44:55] <Ragnarok> okay ill pick it up one day[03:45:25] <Ragnarok> iirc C functions don't do any parameter checking, you can pass parameters to a void parameter function[03:45:39] <Ragnarok> and it compiles when it shouldn't[03:46:03] <SANiK> void blah(void) ...?[03:46:12] <Ragnarok> C even has const and stupid programmers won't even use it[03:46:15] <Ragnarok> yeah[03:46:34] <Ragnarok> const is good[03:46:36] <Ragnarok> not bad[03:46:52] <Ragnarok> C programmers should go play in traffic[03:46:58] <Ragnarok> =3[03:48:31] <Ragnarok> i think imma stick with Vec2,3,4 and not use <t,n>[03:48:52] <Ragnarok> because i have to do alot of checking on the dimensions[03:49:20] <SANiK> Good luck with that and stay clear of the 5th dimension[03:49:42] <Ragnarok> good luck with ur bad habits with C =3[03:50:09] <SANiK> Tests have shown that C programmers have higher IQ ;)[03:50:34] <SANiK> We have to pay attention to more details - makes our brains bigger[03:50:41] <Ragnarok> pfft[03:50:58] <SANiK> It's true - C++'s is lazyman's programming[03:51:03] <Ragnarok> you write more code to acheive the same thing in C++[03:51:06] <SANiK> With with, passing by reference made easy[03:51:33] <hibread> SANiK: easy to use badly, hard to use well (c++)[03:52:33] <Ragnarok> you have to write a function billion different functions for different data types in C when C++ can just use template<typename T>void T foo(T x){ return x; }[03:52:48] *** ajww has joined ##OpenGL[03:52:55] <SANiK> <orbitz> and C functiosn *do* do parameter checking, as long as you prototype them[03:53:24] <Ragnarok> yeah if u write ugly ass prototypes[03:53:35] <SANiK> Whoa whoa - look at the C++ header files[03:53:37] <SANiK> :: < wtf is that[03:53:46] <Ragnarok> atleast is structured[03:53:50] <Ragnarok> its*[03:53:53] <Ragnarok> unlike C[03:54:14] <Ragnarok> C++ can also acheive the same speed as C so don't use that as an argument[03:54:52] <hibread> I can't believe you guys are even fighting a war on C vs C++... does it really matter? :)[03:54:52] <SANiK> <Ragnarok> atleast is structured[03:55:01] <Ragnarok> SANiK, tell <orbitz> how to achieve this in C better than C++ template<typename T>void T foo(T x){ return x; }[03:55:02] <SANiK> hibread - Java sucks[03:55:23] <LordHavoc> <Ragnarok> hey is UT3 really that bad?[03:55:23] <LordHavoc> <LordMetroid> $49 is way cheaper than Swedish prices :([03:55:23] <LordHavoc> <Ragnarok> LordHavoc, go pirate then xD[03:55:23] <LordHavoc> <hibread> Im sure LordHavoc has more heart than that[03:55:27] <LordHavoc> I own ETQW and UT3[03:55:35] <LordHavoc> ETQW looks great, gameplay sucks though :P[03:55:43] <LordHavoc> UT3 looks okay, gameplay is good[03:55:44] <Ragnarok> LordHavoc, and how does this fit here?[03:55:45] *** Maerz has quit IRC[03:55:55] <Ragnarok> fighting over C vs C++[03:55:56] <Ragnarok> xD[03:56:08] <LordHavoc> Ragnarok: only fits because I was mentioned (by accident) in the discussion earlier[03:56:10] <LordHavoc> Ragnarok: so I'm replying[03:56:13] <hibread> LordHavoc: someone had a whoopsy on the auto complete[03:56:26] <Ragnarok> LordHavoc, I think I asked in #gentoo-chat xD[03:56:27] <LordHavoc> hibread: that was obvious, I'm still commenting[03:56:36] <hibread> LordHavoc: good[03:56:42] <LordHavoc> but I don't think either game is worth $50[03:56:59] <hibread> Even if they only sold one copy?[03:57:05] <LordHavoc> hahah[03:57:18] <LordHavoc> I can't even imagine that[03:57:42] <LordHavoc> I think publishers count on the fact that even a horrible game would sell more than one copy[03:58:25] <Ragnarok> LordHavoc, have u tried Nexiuz?[04:00:33] <LordHavoc> Ragnarok: I wrote Nexuiz[04:00:47] <Ragnarok> oops[04:00:54] <Ragnarok> Well I find it bad[04:01:14] <LordHavoc> I find the player models and weapon models bad[04:01:17] <Ragnarok> good graphics but bad control etc[04:01:25] <LordHavoc> and I disagree with the player physics[04:01:30] <LordHavoc> (which is why I made the pro mode)[04:01:38] <Ragnarok> the ai moves too quick, and by the time you shoot ur dead[04:01:52] <hibread> Ragnarok: you are banished from ##opengl, and even freenode.org in general[04:01:57] <LordHavoc> the ai is dumb, I know, I wrote it :P[04:02:02] <Ragnarok> xD[04:02:06] <Ragnarok> hibread, why[04:02:16] <hibread> hehe, no reason[04:02:18] <LordHavoc> maybe for using "ur" :)[04:02:56] <LordHavoc> the campaign isn't easy, but it's meant to be practice[04:03:00] <Ragnarok> hibread, imma throw a 500 pound woman on u[04:03:01] <Ragnarok> xD[04:03:04] <LordHavoc> online it's much harder :P[04:03:27] <Ragnarok> LordHavoc, the user interface is ugly[04:03:35] <Ragnarok> =3[04:05:51] <SANiK> "the user interface is ugly"[04:05:57] <SANiK> "the user is ugly" *fixed*[04:05:58] <SANiK> =P[04:06:04] <Ragnarok> SANiK, ur a troll xD[04:06:50] <SANiK> I just worked on a lot of AI + neural nets, so I tend to speak from the program's perspective[04:09:01] <SANiK> Lindsey Lohan is such a bitch *watches Herbie: Fully Loaded*[04:10:17] <Ragnarok> i'd totally hump her dry[04:10:31] <SANiK> She's really stunning though[04:10:46] <SANiK> Esp. in person - but her attitude is what kills it[04:10:57] <Ragnarok> mmmm amy lee and adriana lima are better ;)[04:11:35] <Ragnarok> or avril lavigne that bitch is hot[04:12:45] <SANiK> avril lavigne looks hot but it's the makeup[04:12:50] <SANiK> Without the makeup she's crap[04:13:04] * Ragnarok wishes i was her husband[04:13:12] * Ragnarok wishes he can knock her up any day[04:16:15] <SANiK> http://img2.timeinc.net/people/i/2007/gallery/avril_lavigne/avril_lavigne5.jpg << HS photo[04:18:01] *** TheLorax has joined ##opengl[04:19:09] <TheLorax> I'm includeing glew, but it says undefined reference to glewInit()...little confused[04:19:41] <SANiK> Did you link the library[04:19:55] <TheLorax> no[04:20:15] <SANiK> That's your problem[04:20:25] <SANiK> What compiler - visual studio? Mingw?[04:20:26] <TheLorax> strange, I would have expected a link time error[04:20:33] <TheLorax> gnu gcc[04:20:39] <TheLorax> it's ok, works now[04:20:42] <TheLorax> thanks[04:20:52] <SANiK> Yay - no we can go back to talking off-topic shit[04:21:03] <SANiK> Channel rule is - we can talk off-topic if no one needs help[04:21:25] <TheLorax> eehh, I've still got a problem though, sorry[04:21:35] <SANiK> ...[04:21:38] <SANiK> go on[04:21:45] <TheLorax> glXSwapIntervalSGI undefined[04:22:06] <Ragnarok> SANiK, wow[04:22:30] <SANiK> Try the latest SDL package TheLorax[04:22:48] <TheLorax> SDL_opengl.h right?[04:23:39] <LordHavoc> Ragnarok: user interface ugly??[04:23:48] <LordHavoc> Ragnarok: I'd completely agree with you about nexuiz 2.3 and below[04:23:52] <LordHavoc> Ragnarok: but 2.4 looks nice in my opinion[04:24:01] <SANiK> TheLorax - I think ;)[04:24:23] <SANiK> TheLorax - but why are you using such weird ass functions O_O[04:24:39] <SANiK> That's a Nvidia only function I assume[04:25:12] <hibread> [ebuild N ] games-fps/nexuiz-2.4.2 USE="alsa opengl sdl -dedicated -maps" 384,551 kB[04:25:24] *** TheChuckster has joined ##OpenGL[04:25:29] <TheLorax> it's just for me...I thought I could quickly limit the FPS, but I think I'll just do it manually[04:25:30] <hibread> if only my net connection wasn't shaped[04:25:55] *** TheChuckster has quit IRC[04:26:31] <Ragnarok> http://www.craigslist.org/about/best/sfo/66795671.html[04:26:36] <Ragnarok> oops wrong paste[04:26:42] <SANiK> Ragnarok - Danica McKellar[04:26:43] <Ragnarok> [ebuild N ] games-fps/nexuiz-2.4.2 USE="alsa opengl sdl -dedicated -maps"[04:26:46] <SANiK> http://i1.ebayimg.com/01/i/06/5b/64/76_1.JPG[04:26:59] <SANiK> She was Winnie on the Wonder years[04:27:20] <hibread> hehe.. and marilyn manson is that other dude[04:27:25] <Ragnarok> SANiK, i'd bone her[04:28:15] <Ragnarok> if Miley Cyrus was 18+ i'd bone her xD[04:28:45] <SANiK> Miley Cyrus has an annoying voice[04:28:50] <SANiK> I'd be like, "Please - don't talk"[04:29:00] <Ragnarok> just shut up and lets get busy[04:29:01] <SANiK> Like one can hear her vocal chords crackling[04:29:16] <SANiK> I guess some 'grease' down the mouth should fix that ;)[04:30:13] <SANiK> And she's 16 - never knew that =o[04:30:20] <Ragnarok> 15 iirc[04:30:44] <SANiK> 16 in November =3[04:31:10] <Ragnarok> im 22 :([04:31:32] <Ragnarok> 23 next year[04:31:39] <SANiK> I'm 20 =([04:31:39] <Ragnarok> w00t i got this chicks number xD[04:31:40] <LordHavoc> I thought this was an informal channel about OpenGL :P[04:31:51] <SANiK> Emma Watson turned 18 (the harry potter chick)[04:31:55] *** braahyan has joined ##OpenGL[04:31:57] <Ragnarok> no one asks, if they do we will stop[04:32:04] <Ragnarok> SANiK, omfg I want her xD[04:32:24] <Ragnarok> that small chick on Gilmore girls <3[04:32:31] <Ragnarok> she was born in my town[04:32:45] <SANiK> http://farm1.static.flickr.com/164/426352429_2a5f1eef50.jpg << she's a party girl[04:32:54] <Ragnarok> damn[04:32:59] <Ragnarok> shes underage also[04:33:03] <SANiK> 18 now though[04:33:08] <SANiK> (that's an old photo)[04:33:10] <Ragnarok> but 21 to legally drink[04:33:15] <SANiK> True[04:33:31] <SANiK> You see the SNL skit where Lindsey Lohan plays Hermione?[04:33:42] <Ragnarok> hey what about ashlee Simpson[04:35:24] <SANiK> Not bad[04:36:15] <Ragnarok> remember the lip signing shit on Leno i think then she started dancing to cover it up[04:36:44] <SANiK> That's what did her in[04:36:51] <SANiK> After that - I haven't heard her name pop up much[04:37:32] <SANiK> Interestingly enough - she's the only one that kind of stayed still "innocent" inside[04:37:33] <Ragnarok> yeah[04:37:49] <Ragnarok> wow this chick I started talking to today, I already got her wanting me[04:37:50] <Ragnarok> xD[04:37:50] <SANiK> Like one hears on the news how Britney fucks up, and Avril,[04:37:56] <Ragnarok> no[04:37:59] <Ragnarok> tell me[04:38:13] <Ragnarok> Britney Spears is a slut and haha her kids got taking awayu[04:38:39] <SANiK> But that's the thing - Ashlee is the only one that still has that innocence in her[04:39:48] <SANiK> All the others have fallen and become sluts in a way[04:39:53] <Ragnarok> yup[04:40:13] <SANiK> Emma Watson (hermione) is already on that path =o[04:40:30] <Ragnarok> i wonder if she will step out of the car next with no undies[04:41:42] <SANiK> It could happen ;)[04:41:50] * Ragnarok wishes badly[04:42:02] <hibread> this is pretty sad stuff[04:42:02] <Ragnarok> i've seen the uncensor britney one[04:42:10] <Ragnarok> hahaha @ hibread[04:42:24] *** Plagman has quit IRC[04:42:48] *** X-Scale has joined ##OpenGL[04:43:29] <SANiK> X-Scale, need help?[04:46:24] <SANiK> It makes me sad how clueless Paris Hilton is[04:46:25] *** braahyan has quit IRC[04:46:26] <Ragnarok> i ggota get off irc becvause this girl says its rude to be on pc while on phone[04:46:39] <SANiK> Get off the phone[04:46:40] <SANiK> then =3[04:46:52] *** ectropy_ has quit IRC[04:46:53] <Ragnarok> naw imma try and get with this chick ;)[04:47:06] <SANiK> She's not worth it man - her cooch is infested with VD[04:47:10] <SANiK> Stay clean - the PC way =o[04:47:13] <SANiK> ok aight peace[04:47:23] <Ragnarok> not yet xD[04:47:33] <Ragnarok> i wish i could have a silent keyboard[04:47:51] <SANiK> I g2g eat anyways - peacers *gone*[04:47:54] *** SANiK has left ##OpenGL[04:47:55] <Ragnarok> later[04:51:28] <hibread> Ragnarok: ive briefly looked over that article, and i can't see the optimizations including inheritance say in the order vec2 to vec3 or what ever. The optimization exists by reducing the size of the temporaries for addition, multiplication or what ever; and then template-ifying everything. I could be wrong ofcourse[04:52:02] <hibread> temporaries in registers[04:52:13] *** braahyan has joined ##OpenGL[04:52:42] <RTFM_FTW> let the optimizer handle this for you[04:53:03] *** TheLorax has quit IRC[04:53:49] <RTFM_FTW> unless you have a damn good reason why the optimizer in your compiler of choice isn't reasonable in this department[04:54:23] <hibread> RTFM_FTW: you're assuming that the optimizer will do a better job than you can. It looks like (atleast back in 2001) you can do better than the compilers optimizer[04:54:38] <hibread> for this specific task, atleast[04:55:11] <hibread> but its not like its an order of magnitude faster...[04:56:00] <RTFM_FTW> that depends upon the coder[04:56:30] <RTFM_FTW> furthermore you aren't going to beat the compiler keeping it all serial[04:56:51] <hibread> we're talking about optimizing say Vec3 someVec = someVec1 + someVec2 + someVec3;[04:56:58] <RTFM_FTW> and if you are vectorizing this then you probably aren't running into this situation in the first place[04:57:17] <hibread> optimizing that expression[04:57:31] <hibread> http://www.flipcode.com/archives/Faster_Vector_Math_Using_Templates.shtml[04:58:28] <RTFM_FTW> in any case for what I've mentioned above something like macstl (http://pixelglow.com/macstl/) can come in handy[04:59:24] <RTFM_FTW> hibread get yourself a better compiler[04:59:27] <RTFM_FTW> heh[05:00:13] <RTFM_FTW> honestly properly using inline and const should go w/o saying AFAIC[05:02:26] *** replor has joined ##OpenGL[05:04:41] <hibread> im not the one trying to optimize here.. it was Ragnarok. But yeah, modern compilers shouldn't realize a performance increase with what that article demonstrates[05:05:34] <RTFM_FTW> yep[05:06:44] <hibread> and who knows.. with all the SSEx instructions these days, the optimization path is probably different[05:09:59] *** ginoman has joined ##OpenGL[05:11:09] <RTFM_FTW> correct... if you are going to optimize look into vectorizing your code[05:11:48] <hibread> RTFM_FTW: what does that mean exactly?[05:12:03] <RTFM_FTW> since the speedup(s) for these sorts of workloads going from serial (even optimized serial) into parallel are far beyond what you are going to get going from serial -> optimized serial[05:12:08] <RTFM_FTW> SIMD[05:12:20] <RTFM_FTW> i.e. AltiVec, SSEx, MMX, ...[05:12:26] <hibread> so you have to do inline assembly for that?[05:12:47] <RTFM_FTW> i.e. mutate your algorithms from serially based into data parallel[05:13:22] <RTFM_FTW> the STL replacement I linked to is one such (platform specific) example of this[05:13:36] <RTFM_FTW> furthermore this doesn't require directly writing assembly code[05:13:49] <RTFM_FTW> there *are* tools available to accomplish this[05:14:14] <hibread> so the tools will use assembly code to realize the benefits?[05:14:23] <hibread> ie, someone has to do the assembly.. the compiler wont?[05:15:15] <RTFM_FTW> yes certain tools will generate SSEx, AltiVec, ... opcodes for you[05:16:03] *** rnx has left ##opengl[05:16:28] <RTFM_FTW> and while there are autovectorizing compilers available (in fact the link I posted does a performance comparison between macstl and Intel's ICC exploiting autovectorization) they frequently fall short of other approaches[05:16:36] <RTFM_FTW> concerning performance[05:17:46] <RTFM_FTW> http://pixelglow.com/macstl/benchmark/ http://pixelglow.com/stories/macstl-intel-autovectorization/[05:18:52] <RTFM_FTW> of course on the desktop AltiVec is all but dead so the performance comparisons noted there are somewhat less useful for most people[05:18:54] <RTFM_FTW> heh[05:22:23] *** braahyan has quit IRC[05:22:38] <hibread> yep no worries then. Ill check all that out at some stage[05:23:49] <Ragnarok> well how do i get sse3 instructions for C++[05:25:51] <hibread> Ragnarok: yep, it looks like that is the place to look for optimization anyway. Just make your inheritance hierachy there as easy to use as possible. Which may contain non at all for vec2, vec3 etc. Templates will be handy though[05:26:06] <hibread> *none[05:26:19] <hibread> im outta ere. Catchyas[05:26:39] <RTFM_FTW> as I've statded above you either (a) use a library capable of generating these for you, (b) use a autovectorizing compiler or (c) convert your code manually... generating them yourself[05:26:40] <Ragnarok> lata[05:27:49] *** braahyan has joined ##OpenGL[05:27:50] <RTFM_FTW> the macstl library I linked to is an example of option (a)[05:28:02] <RTFM_FTW> Intel's ICC is an example of option (b)[05:28:05] *** bfrog has joined ##OpenGL[05:28:32] <RTFM_FTW> GCC also offers support for autovectorization but its quite limited (effectively pathetic in the cases I've looked at)[05:29:12] <bfrog> you have to write your code in a very specific way for it to work good[05:30:07] <RTFM_FTW> yep even then its pathetic[05:30:16] <RTFM_FTW> I've looked into it a few times over the years[05:30:23] <bfrog> its still not that great then, your right[05:30:24] <RTFM_FTW> never found it worthwhile[05:30:44] <bfrog> better off writting out the gcc intrinsics yourself[05:30:53] <RTFM_FTW> always ended up going with option (a) or (c)[05:31:18] <RTFM_FTW> autovectorization is a *very* hard problem :D[05:32:08] <bfrog> who's got the better card out there at the moment, ati or nvidia?[05:32:21] <RTFM_FTW> that depends upon what you are looking for[05:32:33] <bfrog> fragment shader speed[05:32:38] <RTFM_FTW> for anything ALU limited I'd go with ATI[05:33:05] <RTFM_FTW> since ATI stomps Nvidia in that realm right now[05:34:12] <bfrog> what about upload and download speeds?[05:34:27] <bfrog> doesn't ati use gddr4 or something too?[05:34:41] <bfrog> I was seriously thinking the ati cards were looking mighty fine[05:35:37] <RTFM_FTW> yep[05:36:46] <bfrog> they stepped up their linux support I hear quite a bit too[05:36:48] <bfrog> which is nice[05:36:49] <Ragnarok> ati sucks on linux[05:37:12] <bfrog> not from what I've read about the latest cards[05:37:18] <bfrog> I don't have one though so...[05:37:29] <bfrog> got a mac... but its got the measly intel card in there[05:37:36] <bfrog> the ones with real video cards were like double the price[05:37:38] <bfrog> stupid...[05:38:03] *** BahamutZERO has quit IRC[05:40:49] <RTFM_FTW> yes the R6xx series does very well in the realm of commodity image processing[05:41:09] <bfrog> whats it suck at then[05:41:22] <RTFM_FTW> AFAIC nothing[05:41:27] <bfrog> cause the game benchmarks all show the 9800 doing significantly better at the moment[05:41:47] <RTFM_FTW> what games? what platforms?[05:42:02] <RTFM_FTW> and in any case games != image processing codes[05:43:26] <bfrog> the card itself would be what, 3870 ?[05:43:37] <RTFM_FTW> yep and the X2[05:43:40] <bfrog> I never keep track of chip<-> card names[05:43:49] <RTFM_FTW> which is effectively two RV670s on a single board[05:43:53] <bfrog> the x2 being 2 chips on one board? nice :-)[05:44:02] <bfrog> maybe I'll get that then since those are so much cheaper than that nvidia 9800[05:44:11] <bfrog> I was planning on forking over like $350 for a card[05:44:27] <bfrog> at least I thought they were... time to look[05:44:51] <bfrog> what a deal[05:44:58] <RTFM_FTW> indeed[05:44:59] <bfrog> $370 for that thing? thats ridiculous[05:45:09] <bfrog> nvidia's equiv is like 2x as much isn't it?[05:45:32] <bfrog> well $500 at least[05:46:39] <bfrog> oh wait[05:46:44] <bfrog> I didn't see the additional mail in[05:46:49] <bfrog> $329 after all the rebates[05:53:08] * JKnife has a RV610 >_>[05:54:35] *** poseidon has joined ##OpenGL[05:54:48] *** nplus has joined ##OpenGL[05:55:50] *** nplus has quit IRC[05:57:22] *** nathan_ has joined ##OpenGL[06:00:42] *** braahyan has quit IRC[06:04:41] *** obmij_ has joined ##OpenGL[06:05:02] <obmij_> would glGetError() produce something other than GL_NO_ERROR if there wasn't an error[06:05:03] <obmij_> ?[06:05:21] <obmij_> apparently on nVidia it does .... :S[06:06:26] *** braahyan has joined ##OpenGL[06:07:05] <poseidon> Can I pass a SDL_Surface to glBindTexture?[06:07:54] <obmij_> hmm I guess it produces results on ignorable errors too[06:08:00] <obmij_> nevermind[06:09:31] <HuntsMan> poseidon: no[06:10:14] <poseidon> HuntsMan, does glBindTexture just take an array of rgb?[06:10:50] <HuntsMan> hell no![06:11:06] <HuntsMan> it takes a texture target and a texture ID[06:11:19] <HuntsMan> to upload texture data, you use glTexImage2D()[06:11:28] *** rutski has quit IRC[06:12:09] <obmij_> alright, who here loves Bioshock?[06:16:59] <HuntsMan> :crickets:[06:18:50] *** amz has quit IRC[06:22:34] *** mm765 is now known as mm765^away[06:26:01] *** poseidon has quit IRC[06:45:27] *** mediogre has joined ##OpenGL[07:00:19] *** cdleonard has joined ##OpenGL[07:00:38] *** mm^away has joined ##opengl[07:00:48] <cdleonard> fixed-function texgen is per-vertex; right? So I need tesselation for it to be useful[07:05:32] *** ginoman has left ##OpenGL[07:17:56] *** mm765^away has quit IRC[07:21:33] *** JKnife has quit IRC[07:31:56] *** m4ggus has quit IRC[07:40:56] *** HuntsMan has quit IRC[07:44:34] <obmij_> just out of curiosity, do the pixels that don't pass zbuffer check go into the fragment shader?[07:45:55] <obmij_> I guess they actually should[07:45:59] <obmij_> nvm[07:48:29] <cdleonard> AFAIK depth check, alpha test etc are after the fragment shader. But the card can try to optimize the depth test if you don't change the output depth.[07:49:32] <obmij_> Not a good idea actually[07:50:08] <obmij_> because it leaving fragment effects for the end would make repeating the linear interpolations etc. done almost impossible[07:51:32] <obmij_> unless the card has a way of keeping the lerp values, which I doubt[07:51:55] <cdleonard> what lerps are you talking about?[07:52:29] <cdleonard> I don't understand what you're saying[07:53:27] <obmij_> cdleonard, have ever tried vertex colors?[07:53:33] <obmij_> have you*[07:53:40] <obmij_> or do you know how texture mapping is done?[07:54:27] <cdleonard> stuff that goes out of vertex processing is lerped across the triangle; I know that :)[07:54:43] <obmij_> yes, and those 'lerp' values need to be tracked :)[07:55:08] <obmij_> if it leaves fragment processing for end, it needs to go 'back' to do the 'effect' for every pixel[07:55:23] <obmij_> for for every pixel, it needs to bookkeep those 'lerp' values[07:56:38] <obmij_> which is extra memory consumption[07:56:46] <cdleonard> those are separate pipelined units; there is no going back and forth[07:57:08] <obmij_> yep, that's why keeping fragment shading till after zbuffer check is most probably done :D[07:57:23] <obmij_> not done([07:57:27] <obmij_> not done* ^^[07:57:36] <obmij_> my spelling and grammar is failing me tonight[07:59:07] <cdleonard> the pipeline semantic is that stuff which interacts with the framebuffer is done "after" the fragment program[08:00:21] <obmij_> err I think we have an understand gap here[08:00:26] <obmij_> understanding*[08:00:30] <obmij_> nevermind about it[08:00:40] <cdleonard> oh?[08:02:25] <obmij_> just wondering, have you ever gave a shot at designing a 3d engine?[08:02:34] <obmij_> design+implement[08:03:50] <cdleonard> http://ati.amd.com/developer/SDK/AMD_SDK_Samples_May2007/Documentations/Depth_in-depth.pdf[08:05:17] <obmij_> I've coded an entire 3d engine from scratch in Java before mind you[08:05:41] <obmij_> thanks for the link, but I think I already know how depth checks work[08:06:23] <obmij_> there's many ways for hidden surface removal anyway[08:06:36] <cdleonard> Avoid shader depth output: Outputting depth in the fragment shader should be avoided if possible. If a fragment shader modifies the Z value to be used in the depth test, it is not available until after the fragment shader has executed, meaning that no Z optimizations are possible.[08:07:15] <obmij_> still the 'optimizations' they're talking about, I doubt it really leaves out fragment processing[08:08:45] <obmij_> it'd be a waste to attempt such a thing[08:08:53] <cdleonard> it's a very difficult optimization; but probably worth it[08:09:21] <obmij_> but then they'd have to bookkeep things[08:09:47] <obmij_> hmmm, depends on how costly the fragment shader itself is too[08:09:52] <obmij_> maybe they do it?[08:10:38] <cdleonard> the paper says they do it in certain conditions[08:10:53] <obmij_> then again 'ATI'...[08:11:00] <obmij_> wonder if nVidia does that too[08:11:35] <obmij_> I guess they wouldn't worry, if they don't mind saving 2 floats per pixel[08:11:58] <obmij_> 2 extra floats*[08:12:05] *** scy has joined ##opengl[08:12:20] <cdleonard> what 2 extra floats?[08:12:47] <obmij_> heh, have you 'done' a 3d engine before?[08:13:26] <obmij_> you need 2 lerp'd floats that kind of given the relative position of current pixel relative to your other triangle edges[08:13:33] <cdleonard> no, I've never done low-level rasterizing if that's what you mean[08:13:35] <obmij_> s/given/gives you[08:13:50] <obmij_> you need it for texture mapping[08:14:15] <obmij_> based on that you can figure out many things such as fragment color, fragment's texel u,v coords etc.[08:14:30] <obmij_> anything you want lerp'd across the triangle or whatever surface you're rendering[08:15:10] *** Roderic has joined ##OpenGL[08:15:11] *** Roderic is now known as Ingenu[08:18:03] <obmij_> but hmm that won't be the only thing they'll need to book keep[08:18:37] <obmij_> there's numerous things, like texture IDs etc.[08:18:50] <obmij_> damn near impossible[08:19:15] <obmij_> gah I don't know, maybe their 'optimzation' routines just give up 99% of the time[08:19:21] <obmij_> optimization*[08:19:34] <cdleonard> hardware is highly parallelised; it doesn't render one pixel at a time. It is probably VERY different from how software rendering works.[08:19:53] <obmij_> ugh[08:20:05] <obmij_> I don't think hardware can do voodoo magic[08:20:12] <obmij_> there's certain algorithms that we have to this date[08:20:17] <obmij_> and they're widely used[08:20:21] <obmij_> and pretty standard[08:21:16] <obmij_> they can try doing certain operations to speed up the process[08:21:29] <cdleonard> yes; but a hardware pipeline is probably very different from a software callstack[08:21:51] <obmij_> still, there's a place where we call 'human knowledge limit' ;)[08:22:15] <cdleonard> the first cards we're actually called "voodoo" :D[08:22:30] <obmij_> :D[08:22:33] <obmij_> I remember those times[08:22:36] <obmij_> good times[08:24:05] <obmij_> voodoo cards didn't have programmable graphics units, but they were a small steps towards taking some load of the CPU[08:24:12] <obmij_> and people were all hyped up[08:24:34] <obmij_> I almost wonder if you recall the game 'Blood 2: The Chosen'[08:24:41] <obmij_> I also remember nVidia TNT cards[08:24:52] <obmij_> 64 megs of video ram was considered as godly[08:25:26] *** higgsfett has joined ##OpenGL[08:25:36] <higgsfett> hello[08:25:44] <obmij_> hello higgsfett[08:26:48] <higgsfett> well, what are the advantages of a nvidia quadro compared with a nvidia geforce?[08:26:59] <Ingenu> none for gaming[08:27:05] <higgsfett> are the quadros still faster in "line rasterisation"?[08:27:08] *** andikr has joined ##OpenGL[08:28:05] <higgsfett> Ingenu: what else?[08:28:31] <higgsfett> No Quadbuffer Stereo on Geforce?[08:28:42] <higgsfett> and slower line rasterisation?[08:28:47] <higgsfett> for GL_LINE[08:30:50] <Jupp3> obmij_: There's 64 megs on voodoo aswell :)[08:30:50] *** braahyan has quit IRC[08:31:24] <obmij_> I never owned a Voodoo[08:31:25] <Jupp3> At least my Voodoo 5 5500 has[08:31:32] <obmij_> but managed to get my hands on a RIVA TNT 32 Megs[08:31:35] <obmij_> I was a kid[08:31:36] <Jupp3> obmij_: Well just wanted to point that out[08:31:40] <obmij_> so not 'my hands' but my 'dad's'[08:31:47] <Jupp3> And it has 2048 texture size aswell[08:31:53] <obmij_> hah nice[08:32:06] <obmij_> Remeber when 3DMark was actually pretty popular? :D[08:32:12] <obmij_> what was your favorite game?[08:32:13] <Jupp3> Which is required by f.ex. compiz[08:32:40] <Jupp3> Voodoo 3 had... 256, was it?[08:32:44] <Jupp3> max texture size[08:32:47] <obmij_> wait, don't tell me you're using that history artifact to this date ... are you?[08:33:01] <Jupp3> Heh, I remember on Amiga when they made 3D accelerated version of Payback (GTA clone)[08:33:19] <obmij_> omg[08:33:21] <Jupp3> And it refused to run on Voodoo boards, but worked just fine on "much worse" Permedia 2 accelerators :)[08:33:24] <obmij_> which Amiga did you have?[08:33:28] <Jupp3> A1200[08:33:29] <obmij_> I had an Amiga 600[08:33:32] <obmij_> bah[08:33:35] <obmij_> <-- more oldschool ;D[08:33:54] <Jupp3> That was all becouse it just happened to support more detailed textures[08:33:55] <higgsfett> <- A1000[08:34:03] <obmij_> higgsfett, now that's older :D[08:34:10] *** bfrog has quit IRC[08:34:11] <Jupp3> Even when it had only 8 megs of memory :)[08:34:14] <Jupp3> <-C64[08:34:15] <higgsfett> ;)[08:34:18] <obmij_> Man love the old days[08:34:23] <obmij_> My brother had an A500[08:34:32] <obmij_> let's leave that out[08:34:33] <Jupp3> Hey, I had only C64 untill rather late in the 90's![08:34:42] <obmij_> because I could go on rants about how Workbench kicked ass[08:34:53] *** DobosCake has quit IRC[08:34:55] <obmij_> hey Jupp3, my brother had a Sinclair ZX-Spectrum[08:34:58] <obmij_> beat that ;D[08:35:11] <Jupp3> I have Vic-20 in closet aswell[08:35:15] <obmij_> grrr[08:35:19] <obmij_> Ok you win[08:35:21] <obmij_> XD[08:35:22] <Jupp3> And C64 with stereo sound[08:35:42] <obmij_> Oh my dad had an 8086 JR at his office[08:35:48] <obmij_> now you can't beat THAT[08:35:50] <Jupp3> obmij_: AND I am using an irc client that's also available for amiga[08:35:56] <Jupp3> obmij_: On Amiga compatible OS[08:35:58] <obmij_> For the new amiga?[08:36:01] <obmij_> I don't really care[08:36:04] <obmij_> AmigaOne?[08:36:09] <Jupp3> No, I said "compatible"[08:36:11] <obmij_> pffft, Amiga Inc. is just raping the trademark[08:36:12] <Jupp3> MorphOS[08:36:17] <obmij_> oh nvm[08:36:21] <obmij_> don't really care[08:36:29] <Jupp3> obmij_: I have good-as-new Schneider EuroPC[08:36:34] <Jupp3> 8088 system[08:36:39] <obmij_> eh? :D[08:36:43] <obmij_> 8086 JR baby[08:36:48] <obmij_> We still have it[08:36:57] <Jupp3> obmij_: Here's a photo of it: http://www.amiga.org/modules/myalbum/photo.php?lid=2420&cid=5[08:37:05] <obmij_> wait a minute[08:37:11] <obmij_> are you working for Amiga inc.?[08:37:13] *** braahyan has joined ##OpenGL[08:37:22] <Jupp3> obmij_: Sorry, but I rather get paid :-)[08:37:34] <obmij_> oh nvm[08:37:45] <obmij_> I thought it was Amiga.com related[08:37:50] <Jupp3> You know it's funny how they are all broke when they would need to pay ex-employees[08:38:07] <obmij_> Jupp3, name favorite demo/game[08:38:17] <obmij_> Game: Turrican3/Demo: Hardwired[08:38:19] <Jupp3> BUT when it comes to fighting in court, they don't seem to have big problems paying their lawyers[08:38:30] <obmij_> Jupp3, they're just raping Amiga[08:38:36] <Jupp3> obmij_: Have you seen TBL - Starstruck?[08:38:43] <obmij_> honestly I could get a retard to do a better website[08:38:51] <obmij_> naw but I know The Black Lotus[08:38:56] <obmij_> I saw their PSP demo, suicide barbies[08:38:59] <Jupp3> It was shown on UAE as the authors system broke on their way to assembly[08:39:06] <Jupp3> obmij_: Check it out, it's pretty nice[08:39:11] <obmij_> Assembly 08?[08:39:25] <obmij_> btw I was talking to InsaneTTM a while ago[08:39:30] <obmij_> do you know who that is ? :)[08:39:33] <Jupp3> And already at the partyplace there were these rumours, that "it was shown on UAE becouse it wouldn't run properly on any real amiga"[08:39:48] <Jupp3> obmij_: Seriously... There hasn't been any "real" assembly this year yet[08:39:54] <Jupp3> Only the "gamers only" event[08:39:58] <obmij_> pfft[08:40:09] <Jupp3> It was 06 I think when TBL won[08:40:10] <obmij_> I can't really make it to EU demo events anyway[08:40:11] <obmij_> I'm too far[08:40:22] <obmij_> Assembly,breakpoint etc.[08:40:33] <obmij_> Anyway, do you know who InsaneTTM is? :D[08:40:48] <Jupp3> obmij_: But there were many people, some of which are actually good coders... All claiming that it couldn't run on amiga, becouse "it looked to good" :)[08:40:58] <Jupp3> And all that becouse they didn't realize how it was done :)[08:41:10] <obmij_> heh[08:41:24] <Jupp3> obmij_: But anyway, check it out, it's on youtube for example, and see if you can figure it out, how it's done :)[08:41:35] <obmij_> Just wondering though[08:41:40] <obmij_> do you know 'InsaneTTM' ? :D[08:42:16] <Jupp3> obmij_: http://pouet.net/prod.php?which=25778&howmanycomments=-1 - there's a better video there[08:42:21] <Jupp3> Hmm, not sure[08:42:28] <obmij_> do you know Razor1911?[08:42:30] <obmij_> :D[08:42:47] <Jupp3> Who doesn't?[08:42:59] <obmij_> Do you know who started it?[08:43:02] <obmij_> :>[08:43:11] <obmij_> In Norway 1985 :D[08:43:48] <obmij_> 3 friends by the names, InsaneTTM, Dr.No and $ector9... so yeah I was talking to InsaneTTM the other day :D[08:43:59] <Jupp3> Of course the demo runs better on UAE on really powerful system, but that doesn't make it "way too slow" on real amiga, especially on 060 overclocked to 66MHz[08:44:44] <obmij_> quite honestly my brother was more of Amiga/m68k coder, so he know a lot more about playing with AGA etc.[08:44:54] <obmij_> Not me :)[08:45:04] <obmij_> it had its own ways of dealing with sprites etc.[08:45:18] <obmij_> the 'lock mode' with 16bit color and all its implications[08:45:29] <obmij_> he knows better[08:45:30] <Jupp3> But check that video[08:45:34] <Jupp3> lock mode?[08:45:48] <Jupp3> Do you mean Hold-And-Modify?[08:45:53] <obmij_> yeah I think[08:46:06] <obmij_> again, I'm no Amiga coder :D[08:46:08] <obmij_> never was[08:46:11] <obmij_> my brother is though[08:46:15] <Jupp3> HAM8, which can "officially" do 272k colors in 8bpp[08:46:18] <obmij_> he's many many years older than me[08:46:31] <obmij_> I just loved my Amiga[08:46:32] <Jupp3> But if you do some maths in "best case scenario", it can do easily more[08:46:33] <obmij_> dearly[08:46:56] <Jupp3> But no point creating a picture that can "only" show lots of colors instead of showing something good looking :D[08:47:03] <obmij_> :)[08:47:16] <obmij_> <3 the old intermissions[08:47:17] <obmij_> with art work[08:47:24] *** nathan_ has quit IRC[08:47:25] <obmij_> almost always you had something witty[08:47:26] <obmij_> :D[08:47:35] *** kenws has joined ##OpenGL[08:47:38] <obmij_> But it's good to see EU guys still keeping it alive[08:47:49] <obmij_> Afterall they were the heart of the cracking/demo scene[08:47:53] <obmij_> FLT, RZR, etc.[08:48:47] <Jupp3> obmij_: But check that video :)[08:49:39] <obmij_> Jupp3, right now I'm a hurry to finish a study for an exam[08:49:41] <obmij_> that I have at 9 am :P[08:49:46] <obmij_> it's 2:46 right now[08:49:51] <Jupp3> It's got music by Olof Gustaffson btw[08:49:52] <obmij_> I'll watch that video when I get back home[08:50:15] <Jupp3> dl it anyway in advance[08:50:25] *** kenws has quit IRC[08:50:27] *** charlie55 has joined ##OpenGL[08:50:37] <obmij_> bookmarked[08:50:38] <obmij_> :)[08:57:38] <obmij_> btw Jupp3, you only win by 2 years ;D[08:57:45] <obmij_> VIC-20 was released in 1980/1981[08:57:54] <obmij_> Spectrum was releasd 1982 ;D[08:58:03] <Jupp3> Well I got it way after C64 anyway :)[08:58:29] * obmij_ runs to see if he can find a DEC PDP-7[08:58:43] <obmij_> then nothing can claim oldschool on me muahahaha }:))[08:58:49] <obmij_> ;D[09:01:21] *** braahyan has quit IRC[09:02:07] *** scy has left ##opengl[09:07:14] *** braahyan has joined ##OpenGL[09:09:14] *** charlie5 has quit IRC[09:17:48] *** vasoq has joined ##OpenGL[09:18:49] <obmij_> alright, still my early question remains ... who here loves Bioshock?[09:19:02] <obmij_> come on, one you guys must've played it :d[09:19:03] <obmij_> :D[09:19:12] *** Jupp3 has quit IRC[09:19:21] <vasoq> why does x1 = ((lastx / winw)*2.0f) - 1.0f; (where x1 is a GLfloat and the others are ints) always equal -1?[09:20:35] <vasoq> what i'm trying to do is map window coordinates to [(-1,-1), (1, 1)][09:20:49] <obmij_> that's about right[09:21:17] <obmij_> given x ranging from 0 to max and winw being max[09:21:39] <vasoq> yeah[09:22:01] <vasoq> i checked the data and it's all winw = 800 and x = 400 or so if i pick the middle of the screen[09:22:13] <vasoq> but x1 always equals -1[09:22:16] <obmij_> ok[09:22:29] <obmij_> do (float)lastx[09:22:46] <obmij_> int/int with no typecast gives int[09:22:48] <obmij_> and that's a 0[09:22:51] <obmij_> for your case[09:23:06] <vasoq> purty, thanks[09:23:09] <obmij_> especially since you put it in brackets[09:23:14] <obmij_> purty?[09:23:19] <obmij_> what is that?[09:23:22] <vasoq> i figured it was a type problem but i thought i might be misusing GLfloat since i have no idea what it is[09:23:32] <vasoq> purty is like pretty but for hicks[09:23:35] <obmij_> typedef float GLfloat;[09:23:48] <obmij_> wait are you a female? :D[09:24:45] <vasoq> no, females don't exist on the internet[09:24:51] <obmij_> lulz[09:24:54] <vasoq> i'm just strange[09:24:57] <obmij_> no they do :)[09:25:05] <obmij_> unless you mean the intarbuttz[09:25:09] <vasoq> well the whole damn assignment works now, hooray[09:25:13] <obmij_> which is completely different from the intarweb[09:25:20] <obmij_> therefore being different from the Internet[09:25:41] <vasoq> i hate when i finish an assignment and the thing they were trying to get me to make ends up sucking anyway, though[09:26:17] <obmij_> hmm they're asking you to do it in OpenGL?[09:26:18] <obmij_> what is it anyway?[09:27:47] <vasoq> rotating a bottle using quaternions[09:27:47] *** [AD]Turbo has joined ##OpenGL[09:27:54] <vasoq> mesh wire bottle that is[09:27:54] <obmij_> hmm[09:27:57] <obmij_> interestingly[09:28:02] <obmij_> everyone's using quaternions[09:28:02] <vasoq> but it turns out when it's all done it rotates way too damn fast[09:28:10] <obmij_> however[09:28:15] <vasoq> this assignement seems to be from 1997 or something[09:28:18] <obmij_> there is another method which is slightly easier to understand[09:28:23] <obmij_> with the same accuracy[09:28:25] <obmij_> and end results[09:28:57] <obmij_> map the vector onto a plane represented at the origin with the rotation axis as normal[09:29:05] <obmij_> rotate it as if it is in some local 2d coordinate system[09:29:16] <obmij_> and then add the 'previously taken out' z-component[09:29:23] <obmij_> to get rotated final vector[09:29:52] <obmij_> much easier to understand than a whole 4x4 matrix which looks like gibberish[09:30:11] <obmij_> and cost wise, it isn't really costly in comparison[09:30:16] <obmij_> I'd say it's even faster[09:30:17] *** groton has joined ##OpenGL[09:31:24] <obmij_> wow I just compared codes[09:31:29] <obmij_> and it is insanely less costly[09:31:35] <obmij_> the latter method (my method)[09:31:55] <obmij_> setting up the quaternion spatial rotation matrix is heavy in terms of mathematical operations[09:32:08] <obmij_> meh your choice[09:32:13] <vasoq> it's not my choice[09:32:40] <vasoq> it's their assignment[09:32:47] <obmij_> I guess :)[09:33:16] <obmij_> sometimes some concepts are over complicated for no good reason really :/[09:33:38] <vasoq> why do people like quaternarions if what you say is so good?[09:34:08] <vasoq> doesn't it produce "good" rotations, as in no defects?[09:34:13] <[AD]Turbo> yo all[09:34:35] <obmij_> it produces perfect rotations[09:34:47] <obmij_> the fact is, it does what quaternions do[09:34:54] <obmij_> without a matrix multiplication involved[09:35:19] *** kenws has joined ##OpenGL[09:35:26] <obmij_> as to why people don't use it... I don't know[09:35:39] <obmij_> maybe it's because it's not as simple as setting up a matrix[09:35:52] <obmij_> but overall it's faster by maybe a few clock cycles[09:36:05] <obmij_> the concept however is easier to grasp[09:38:35] *** cdleonard has left ##OpenGL[09:38:35] *** belou has joined ##OpenGL[09:38:54] <obmij_> and really using quaternions for axis rotations is kind of a waste[09:39:02] <obmij_> quaternions on their own have MUCH better applications[09:39:13] <obmij_> to use them for such a petty effect seems waste :/[09:43:54] <Ingenu> try to use quaternions with scaling and you'll see they aren't so great[09:46:28] *** braahyan has quit IRC[09:52:18] *** braahyan has joined ##OpenGL[09:56:17] <obmij_> wait, I'm using non power of two textures with sampler2D and GL_TEXTURE_2D and everything's fine[09:56:39] <obmij_> like u,v coords at range 0 to 1 work fine[09:57:01] <obmij_> am I missing something?[09:59:41] <obmij_> oh wait , gotta check to see NPOT is supported or not[10:03:23] *** blbmp has quit IRC[10:06:34] *** kaotrix has quit IRC[10:10:23] *** braahyan has quit IRC[10:11:24] *** servus has quit IRC[10:15:15] *** servus has joined ##opengl[10:16:42] *** braahyan has joined ##OpenGL[10:40:49] *** linxuz3r_ has joined ##OpenGL[10:43:33] *** X-Scale has left ##OpenGL[10:50:10] *** neoneye has joined ##OpenGL[10:53:31] *** sohail has quit IRC[10:55:58] *** scy has joined ##opengl[11:00:15] *** BahamutZERO has joined ##OpenGL[11:00:50] *** Burga has joined ##OpenGL[11:03:34] <Weiss> Ragnarok: page 450 (chapter 11)[11:07:06] *** predaeus has joined ##opengl[11:10:11] *** braahyan has quit IRC[11:16:26] *** braahyan has joined ##OpenGL[11:18:23] *** dvoid has joined ##OpenGL[11:21:46] *** braahyan has quit IRC[11:23:36] *** Eforen has joined ##opengl[11:25:33] <Eforen> hihi[11:27:02] <Eforen> anyone awake?[11:27:33] *** braahyan has joined ##OpenGL[11:28:32] <Eforen> well here is my question i am trying to get opengl working in msvs[11:29:08] <Eforen> and when i try to include #include <gl/glaux.h> it can't find it[11:32:05] *** Suprano has joined ##OpenGL[11:32:10] <belou> what is your compilation call ?[11:32:23] <belou> do you have that file on your computer ?[11:32:41] <predaeus> Eforen, as far as I know glaux is not used anymore these days.[11:33:28] <predaeus> http://opengl.org/resources/faq/technical/gettingstarted.htm tells you to use glut instead. So just use any other option that creates a context for you.[11:35:55] <Eforen> even http://nehe.gamedev.net/data/lessons/lesson.asp?lesson=01 uses it[11:36:22] <Eforen> how can i lurn opengl if all tutorials use a unsuported lib?[11:36:53] <Eforen> every tut i have seen uses glaux[11:41:46] <predaeus> Eforen, http://www.glprogramming.com/red/index.html includes GLUT examples.[11:42:03] <belou> http://www.pixyland.org/peterpan/pixyPromos.html[11:42:18] <belou> you can use theses textures for your fist cube[11:42:23] <belou> really nice :-)[11:54:42] *** braahyan has quit IRC[11:55:22] *** linxuz3r_ has quit IRC[11:55:31] *** Ademan_ has joined ##OpenGL[11:55:51] *** Ademan has quit IRC[11:56:00] *** Ademan_ is now known as Ademan[12:00:07] *** braahyan has joined ##OpenGL[12:02:45] *** nplus has joined ##OpenGL[12:03:11] *** nplus has quit IRC[12:04:42] *** Plagman has joined ##OpenGL[12:06:19] *** thesquib has joined ##OpenGL[12:12:10] *** braahyan has quit IRC[12:17:15] *** braahyan has joined ##OpenGL[12:22:20] *** braahyan has quit IRC[12:29:09] *** thesquib has quit IRC[12:35:24] *** Eforen has quit IRC[12:46:46] *** BadChoice has joined ##OpenGL[12:47:11] <BadChoice> hi![12:47:22] *** Xmas| has joined ##OpenGL[12:47:25] <BadChoice> do you know if is possible to run a video in a texture'[12:47:27] <BadChoice> ?[12:49:35] *** korff has joined ##OpenGL[12:52:04] <predaeus> BadChoice, sure is[12:52:31] *** scy has left ##opengl[12:53:12] <BadChoice> do you know how?[12:53:25] *** dvoid has quit IRC[12:53:56] *** dvoid has joined ##OpenGL[12:54:43] <predaeus> BadChoice, you can likely use some library to decode the video data and load that as a texture in OpenGL and display it i.e. on a fullscreen quad.[12:55:29] <predaeus> BadChoice, several video players provide OpenGL renderers[12:55:30] <BadChoice> mm, code a timer and change the texture every specific time ?[12:56:15] <predaeus> that's the basic idea, yes[13:00:12] *** Burga has quit IRC[13:06:43] *** Rangar has joined ##OpenGL[13:07:36] <BadChoice> mm[13:07:37] <BadChoice> ok[13:07:41] <BadChoice> I'll try it[13:07:44] <BadChoice> thanks![13:14:50] <obmij_> something that has never ceased to amaze me is why ... for +x,-x,+z,-z screen-to-cubemap surface rendering through frame buffer objects, I had to use (0,-1,0) up vector instead of (0,1,0) :/[13:16:05] <obmij_> maybe that's how it renders ... upside down... no idea[13:16:51] *** obmij_ has quit IRC[13:19:25] *** sanjin has joined ##OpenGL[13:20:40] *** LiQuiDninja has joined ##OpenGL[13:22:52] *** sin has joined ##OpenGL[13:24:42] <sin> hey there, I've got GLfloat *vertices; and I dynamically allocate memory for a set of vertices and then I am trying to use glVertexPointer and glDrawElements (using another set of faces as well) however I get weird results or sometimes my program crashes. If I use an array of vertices GLfloat vertices[] = { /* ... */ }; it works.[13:25:03] <sin> how can I use dynamically allocated arrays with glVertexPointer and glDrawElements ?[13:30:29] <andikr> sin: It should work either way. Perhaps you didn't allocate the correct size for the vertices?! You need "number of vertices" * sizeof (GLfloat) bytes.[13:33:12] <andikr> sin: I mean "number of vertices" * sizeof(GLfloat) * 3 - sorry[13:35:26] *** elite01 has joined ##opengl[13:36:16] *** Jernej has joined ##OpenGL[13:38:14] *** belou has quit IRC[13:39:04] *** belou has joined ##OpenGL[13:40:05] <sin> andikr, actually that's exactly what I am doing, sometimes I get a memory corruption failure... I've got vertices = (GLfloat *)malloc(nvertices * sizeof(GLfloat) * 3);[13:40:17] *** JernejL_ has joined ##OpenGL[13:43:44] <andikr> sin: No idea then. For debugging you could replace the glDrawElements call with a standard loop drawing the elements to check if the vertices are ok.[13:43:44] *** sin has quit IRC[13:47:47] *** Suprano has quit IRC[13:51:36] *** nplus has joined ##OpenGL[13:54:28] *** JernejL has quit IRC[13:55:49] *** Jernej has quit IRC[14:04:32] *** braahyan has joined ##OpenGL[14:05:03] *** LiQuiDninja has quit IRC[14:08:54] *** brutopia has quit IRC[14:12:17] *** brutopia has joined ##opengl[14:14:16] *** wizo has joined ##OpenGL[14:15:19] <wizo> hey, not sure if this is the right place to ask, but what do i have to search for if i wanna learn about "directions"?[14:15:55] <Weiss> wizo: vectors?[14:15:58] <wizo> like.. directions or headings[14:16:16] <wizo> vectors? alright, i'll try that[14:16:26] <wizo> what about the maths behind it?[14:17:52] <Ingenu> a vector is a mathematical object[14:21:04] <wizo> i see, but to use that to get a "direction" ? i would like ot understand it[14:22:46] <Weiss> a vector *is* a direction[14:22:52] <Weiss> are you looking for the angle between a direction and another direction?[14:23:34] <wizo> for example i would like to know which "direction" a car is facing[14:23:51] <wizo> http://www.gamedev.net/reference/articles/article435.asp <== is it this ?[14:24:57] <Weiss> yes, that looks like a good starting point[14:31:05] *** braahyan has quit IRC[14:33:44] <wizo> alright thanks[14:34:14] <Ingenu> I don't think we should let people learn about linear maths from the web[14:34:20] <Ingenu> it's too incomplete and innacurate[14:34:25] <Ingenu> books and teachers are better[14:34:50] *** groton has quit IRC[14:35:30] *** brutopia has quit IRC[14:35:32] <wizo> i wonder where is the "previous" section it was talking about..[14:36:47] *** BadChoice has left ##OpenGL[14:37:05] *** braahyan has joined ##OpenGL[14:39:37] *** brutopia has joined ##opengl[14:42:11] *** charlie55 is now known as charlie5[14:50:28] *** prophile has joined ##opengl[14:52:32] *** groton has joined ##OpenGL[14:56:29] *** brutopia has quit IRC[14:56:43] *** juanmabc has joined ##opengl[15:01:51] *** brutopia has joined ##opengl[15:02:02] *** mm^away is now known as mm765[15:05:29] *** wizo has quit IRC[15:06:16] *** KU0N has joined ##OpenGL[15:06:26] <KU0N> hello[15:13:39] *** rnx has joined ##opengl[15:16:09] *** bbeausej has joined ##OpenGL[15:22:30] *** fusion-35 has joined ##OpenGL[15:23:26] <fusion-35> anyone mind helping me with some shadow problem I'm having? i got shadows cast from solid objects on the ground, but im having trouble casting them onto another plane (another wall in this case). I have solid objects on walls and on ground, but its only casting them on ground[15:23:52] <fusion-35> I know its gotta do somethign with changing the planes around, and then push, matrixshadow, drawobjects, pop[15:23:57] <fusion-35> buti cna't seem to pull it together[15:24:14] <fusion-35> i'd probably need to send the main.cpp if someone was interested :)[15:25:05] *** Suprano has joined ##OpenGL[15:25:40] <hibread> fusion-35: what sort of shadows are these? Not shadow maps i gather?[15:25:53] <fusion-35> hmm don' tthink so, just standard shadows from a light source[15:26:20] <hibread> errr. hmmm. there is no such thing :)[15:26:45] <fusion-35> oh ok >_< umm let me copy a few lines of the code[15:26:50] <fusion-35> computeShadowVolume(L, O, M, N, Pbg, P);[15:26:51] <fusion-35> constructShadowVolumeMatrixd(P, matrix);[15:26:51] <fusion-35> shadowMatrix(shadowMat, Pg, Lf);[15:26:56] <fusion-35> does that show what type shadow it is?[15:27:09] <fusion-35> pg = plane ground, which was defined earlier in the code[15:27:18] <hibread> and where do these functions come from?[15:27:18] <fusion-35> pbg = plane below ground defined earlier in code[15:27:43] <hibread> are you using the accum buffer by any chance?[15:28:04] <fusion-35> they have been written into the main.cpp code. those like "constructshadowvolumematrix" etc[15:28:13] <fusion-35> hmm dont' know what accum buffer is >_< im kinda new to opengl[15:28:33] <fusion-35> maybe learnign shadows at this stage is a bit advanced for me, but i can do shapes quite easily and moving things etc[15:28:44] <fusion-35> i successfully managed to reflect the scene on the floor too :)[15:29:34] *** Walt has quit IRC[15:30:25] <fusion-35> i think doing the shadows on differnet planes has the logic of changing the shadow matrix for each plane, and then drawing the objects, and then doing a popmatrix, then starting again wtih push, change shadowplane, then draw objecxts again, then pushmatrix[15:31:11] <fusion-35> i've got the ground plane defined and works fine, but when i try for a differnet plane, it just doesn't really do anything..[15:31:29] <hibread> I can recall a tutorial i read many years ago that did shadowing like this. I would personally forget this completely and use either shadow maps or stencil shadows[15:31:40] <hibread> they are far more robust than the ones you mention here[15:31:49] <fusion-35> are they easy to implement?[15:31:55] <fusion-35> for a beginner[15:32:28] *** dv_ has joined ##OpenGL[15:32:56] <hibread> ummm... the concept is relatively straight forward but to be quite honest, i am yet to impliment them myself[15:33:17] <fusion-35> ahh ok >_<[15:34:21] <hibread> shadow maps entail having to render the scene from the point of view of the light source into a depth buffer... and then when rendering the parts of the scene relevant to that light source, you use that depth shadow map to select which pixels to light or not light[15:37:15] <hibread> stencil shadows use the stencil buffer along with silhouette polygons to mask particular areas of the scene where the light will reach or not[15:37:57] <fusion-35> hmm stencil shadows is the one im doing, the stencil buffer stuff rings a bell now that you mention it[15:38:17] <fusion-35> im still getting accostomed to these terms[15:39:21] <hibread> hmm, but these shadow planes you're talking about.. doesn't sound like the true stencil shadow maps[15:39:21] <fusion-35> which one do you think si the better one for shadows of solid shapes? :)[15:40:51] <hibread> stencil shadows will create very hard edges (unless some more advanced techniques are used). Shadow maps are less perfect in that way, but are more robust when drawing objects that may have transparent parts[15:41:04] <hibread> stencil shadows strictly follow the geometry[15:42:07] <hibread> where as say if you were rendering a fence with one quad, but had the texture of the wire etc (with the gaps being tranparent), stencil shadows would not work, whether as shadow maps will[15:42:18] <fusion-35> hmm interesting, so i suppose having a tree image mapped to a quad coudlnt' be shadowed with stencil shadows[15:42:23] <hibread> i think generally speaking shadow maps are the more desired method[15:42:56] <hibread> yeah thats right.. if the billboard tree used the alpha channel to create transparencies or what ever[15:43:54] *** nplus has quit IRC[15:44:07] *** nplus has joined ##OpenGL[15:46:59] *** nplus has quit IRC[15:47:05] *** nplus has joined ##OpenGL[15:47:32] <fusion-35> cool. well thanks for the info hibread, i'd be interested to look this up in a few days (about the shadow maps). im gonna get back to figuring out my shadow thing again[15:47:56] <hibread> no worries[15:49:08] *** fusion-35 has quit IRC[15:49:17] *** harrisg has joined ##OpenGL[15:57:54] *** sanjin has left ##OpenGL[16:01:03] *** deniz_ has joined ##OpenGL[16:02:11] <deniz_> Hi, I want to make a drawing application, simple one, where the drawing is like the pen brush. Suggestions?[16:02:26] *** korff has quit IRC[16:04:12] <groton> deniz_, do not call it Paint[16:05:07] <deniz_> hehe[16:05:10] <deniz_> alright[16:10:37] *** braahyan has quit IRC[16:13:17] <Rangar> http://www.foxnews.com/story/0,2933,358029,00.html for those interested in the Titanic mystery :)[16:13:33] <neric> hy[16:14:03] <neric> someone knows if x600 is able to deal with non-power-of-two textures ?[16:15:41] *** braahyan has joined ##OpenGL[16:16:14] <Weiss> neric: don't think so (at least the linux driver for my mobility radeon X600 doesn't advertise it)[16:16:41] <neric> oh really ... :|[16:17:45] *** alyawn has quit IRC[16:18:06] <neric> thx Weiss[16:19:38] <hibread> neric: maybe GL_TEXTURE_RECTANGLE_ARB might be of use?[16:20:19] <neric> apparently .. it doesn't ..[16:21:18] *** ajww has quit IRC[16:25:05] *** belou has quit IRC[16:25:08] *** neoneye has quit IRC[16:25:13] *** alyawn has joined ##opengl[16:26:44] *** belou has joined ##OpenGL[16:29:03] <Weiss> neric: however, i often wonder how much of the (admittedly not very great) abilities of my GPU aren't exposed by the driver..[16:30:42] <Ingenu> what GPU do you have ?[16:31:11] <neric> actually, it's not for me .. i usually nvidia :p[16:31:30] <Ingenu> if you use intel, then it expose features you don't have ;p[16:31:43] *** poseidon has joined ##OpenGL[16:31:43] <Weiss> Ingenu: radeon X600 mobility[16:31:46] <Ingenu> nvidia and ati are fine, except ATi is currently a bit behind with OpenGL due to GL3.0[16:32:15] <poseidon> Whats a good image library for loading textures into openGL?[16:32:33] <neric> poseidon: OpenCV ;)[16:32:38] <Ingenu> devil maybe[16:32:42] <Ingenu> I just load things mysefl[16:33:10] <Ingenu> or I used to anyway, now I'm working for video game firms, it's worth than whatever you can imagine[16:33:19] <Ingenu> worth <= HERM[16:33:20] <Ingenu> WORST[16:33:58] <neric> devil looks better than opencv ..[16:34:58] <Ingenu> opencv might be newer though no clue[16:35:05] <poseidon> Ingenu, what types of games are you making? (game development is a career path I'm thinking of).[16:35:21] <Ingenu> I've been working on a number of games[16:35:46] <poseidon> Ingenu, do you like the job?[16:36:06] <Ingenu> I don't like working on pieces of shit that student could have written[16:36:20] <Ingenu> written by so called "professionnals" that code first, think later[16:36:25] <Ingenu> then tell you it's not their fault[16:36:28] <Ingenu> like they didn't write it[16:36:35] <poseidon> lol[16:36:48] <Ingenu> they always have an excuse[16:36:54] <Ingenu> management is often pretty lame[16:37:01] <Ingenu> if it's average you're lucky[16:37:46] *** higgsfett has quit IRC[16:37:51] <Ingenu> that said it may be due to having been ordered to work on week-ends and holidays[16:37:58] <poseidon> Ingenu, are your games usually win32 specific or are they cross-platform? (or are you making games for consoles)[16:38:00] <Ingenu> and not being paid for it neither overtime[16:38:03] <Ingenu> from my current employer[16:38:12] <Ingenu> that makes me so fucking upset[16:38:26] <Ingenu> and of course management finds a 3 months crunch time 'normal"[16:38:27] <Ingenu> my ass[16:38:28] *** higgsfett has joined ##OpenGL[16:38:29] *** huperniketes has quit IRC[16:38:52] <Ingenu> games are always written the worst way possible, so of course they are not portable and require a lot of work[16:38:59] <Ingenu> (to get to run on another target)[16:39:09] <poseidon> You should make like the guy in office space.[16:39:20] <Ingenu> don't tempt me ;p[16:39:28] <harrisg> Ingenu you don't have a red stapler do you?[16:39:52] <Ingenu> no[16:40:02] *** huperniketes has joined ##OpenGL[16:40:07] <harrisg> whew[16:40:58] <poseidon> Ingenu, Does the company allow you to use openGL?[16:43:05] *** stringfellow has joined ##opengl[16:44:23] <Ingenu> using DX9[16:44:24] <Ingenu> of course[16:45:09] *** harrisg has left ##OpenGL[16:45:42] <neric> =)[16:48:32] *** huperniketes has quit IRC[16:50:21] *** huperniketes has joined ##OpenGL[16:53:20] *** LiQuiDninja has joined ##OpenGL[16:53:45] *** prophile has quit IRC[16:54:12] <hibread> Ingenu: looks like im a masochist. Well.. a wannabe masochist[16:56:05] <Ingenu> you'll change your mind[16:56:37] <Ingenu> with time I become less and less merciful when it comes to aweful code, I know it can be done right, I know it's easy[16:57:14] <Ingenu> I'm just leaving the office now anyway, see you people tomorrow[16:57:35] <hibread> later[16:57:48] *** Ingenu has quit IRC[16:59:01] *** Walt has joined ##opengl[17:03:25] *** stringfellow_ has joined ##opengl[17:04:49] *** andikr has quit IRC[17:07:59] *** scy has joined ##opengl[17:10:01] *** braahyan has quit IRC[17:11:16] *** Walt has quit IRC[17:11:30] *** sohail has joined ##OpenGL[17:16:26] *** braahyan has joined ##OpenGL[17:16:42] *** rutski has joined ##OpenGL[17:17:16] *** stringfellow has quit IRC[17:19:27] *** braahyan has quit IRC[17:25:42] *** braahyan has joined ##OpenGL[17:28:16] <Ragnarok> I wanna go into the gamign industry but I don't have college experience :([17:28:30] <Ragnarok> it sucks hibread[17:29:32] <hibread> Ragnarok: i wouldn't let that stop me[17:30:11] <Ragnarok> but whats wrong with it?[17:30:26] <hibread> whats wrong with what?[17:30:57] <Ragnarok> massochist?[17:31:02] <Ragnarok> iono wth that is[17:31:07] <Ragnarok> xD[17:33:05] <Ragnarok> hmmm[17:36:35] <vasoq> how hard is the gaming industry to get in to even with a college degree?[17:38:26] <poseidon> Wheres the best place to get a free opengl documentation or reference?[17:38:59] *** BadChoice has joined ##OpenGL[17:39:18] <BadChoice> Hi! is libmpeg3 good to use with sdl and opengl?[17:39:19] <poseidon> vasoq, That depends on your idea of the "gaming industry." ;)[17:39:20] *** mattn2|home has joined ##OpenGL[17:40:21] *** oogaw is now known as Wagoo[17:40:43] <hibread> poseidon: www.opengl.org[17:40:46] *** gus4n0 has quit IRC[17:40:47] <poseidon> BadChoice, Not sure about openGL (I'm new to it). But I used SDL_mixer when I was using strictly sdl[17:41:16] <BadChoice> SDL_mixer can load mpgs?[17:41:36] <vasoq> my idea of the gaming industry is the place where i can contribute to the creation of a video game[17:43:06] <BadChoice> i try to use smpeg but I cant get i work if the surface is other than screen..[17:44:47] *** Walt has joined ##opengl[17:46:40] *** kenws has quit IRC[17:46:45] <rnx> mixer does audio ... i assume he wants video[17:49:16] <BadChoice> yes[17:49:27] <BadChoice> I'm dealing with smpeg...[17:52:10] <BadChoice> but all he examples I found uses it with the main screen.. and I want it on a texutre that I can rotate scale etc..[17:54:20] <rnx> ages that i used smpeg ... ffmpeg is the thing to use these days faik[17:55:32] *** gusano has joined ##OpenGL[17:58:57] *** nytejade has quit IRC[17:59:58] <BadChoice> ffmpeg?[18:00:07] <BadChoice> I'll take a look[18:00:09] <BadChoice> thanks![18:00:17] <BadChoice> I have my project in a very good way![18:00:22] <BadChoice> the quicklook for linux[18:00:33] <BadChoice> I can preview images mp3 pdfs and text files,[18:00:40] <BadChoice> next step.. videos![18:01:03] *** nytejade has joined ##OpenGL[18:02:58] <Ragnarok> are bsp trees outdated?[18:03:19] *** [AD]Turbo has quit IRC[18:03:41] <hibread> Ragnarok: i dont think outdated is the right word. They're probably not used as offen as they used to be[18:04:25] <BadChoice> mm[18:04:33] <BadChoice> is there a ffmpeg library?[18:04:38] <BadChoice> or it uses libmpeg4?[18:04:47] <Ragnarok> wow we had 2 A10's fly over where i live[18:05:35] *** dvoid has quit IRC[18:06:10] *** dvoid has joined ##OpenGL[18:06:57] <BadChoice> or is this library :libavcodec-dev ?[18:09:08] *** groton has quit IRC[18:10:34] *** nplus_ has joined ##OpenGL[18:11:53] *** nplus has quit IRC[18:16:04] *** Yustme has joined ##opengl[18:17:13] *** mattn2|home has quit IRC[18:20:15] <poseidon> Isn't there a method to the names of openGL functions?[18:21:26] *** LiQuiDninja has quit IRC[18:23:13] <belou> poseidon, yes[18:23:28] <poseidon> belou, where can I find out what it is?[18:23:30] <belou> but to know the method you have to pass 3 steps[18:24:56] <belou> 1- learn Native Americans langage[18:25:13] <belou> 2- prove us that you can fly without using your legs[18:25:51] <belou> 3- stole a candy to a child and throw it in a garden with a big bad dog, then fight the dog, save the kid, date the mum[18:26:39] <belou> or there is a lot of naming functions method[18:26:43] *** bfrog has joined ##opengl[18:27:02] <bfrog> have the ati drivers improved significantly since like the radeon 9000 days[18:27:07] <bfrog> especially for linux?[18:27:09] <belou> mine is an action naming format[18:27:28] <belou> the function's name describes what it does[18:27:37] <belou> with _ between words[18:27:47] *** belou has quit IRC[18:28:48] *** jann has joined ##OpenGL[18:29:06] <jann> hi[18:29:12] *** Jupp3 has joined ##OpenGL[18:32:57] <poseidon> bfrog, do you indicate what it returns or takes?[18:33:18] <poseidon> btw, #3 is a good idea[18:34:10] <poseidon> *belou[18:35:04] *** brutopia has quit IRC[18:40:03] *** brutopia has joined ##opengl[18:41:07] *** braahyan has quit IRC[18:41:16] *** brutopia has quit IRC[18:46:28] *** brutopia has joined ##opengl[18:46:58] <poseidon> How do I set up devIL so I can include it in my cpp files (no matter their directory)?[18:47:00] *** Yustme has quit IRC[18:47:05] *** wersf has joined ##opengl[18:47:05] <poseidon> linux[18:48:34] <bfrog> add a -I flag[18:48:46] <bfrog> gcc blah blah -I/path/to/devil/headers blah blah[18:49:37] <bfrog> or put devil in /usr or /usr/local so that all the headers are in include, the lib in libs etc[18:49:41] *** dv_ has quit IRC[18:49:45] <bfrog> should be standard if you do make install or whatever[18:49:48] <bfrog> or install from a package[18:50:39] <poseidon> bfrog, I did install it from a package (libdevil-dev) but I get the compiler error of devil.h not found[18:50:44] *** dv_ has joined ##OpenGL[18:50:54] <poseidon> *and that all the devil functions are out of scope of course[18:51:02] <bfrog> look in /usr/include for a folder called devil then[18:51:13] <bfrog> I bet it shoved all the headers in to a directory there[18:51:22] <bfrog> just add the -I flag then[18:52:13] <poseidon> bfrog, i don't see it there[18:52:47] <bfrog> if your using debian/ubuntu they apparently shove them in to /usr/include/IL for whatever reason[18:52:50] <bfrog> I have no idea why...[18:52:53] <poseidon> Just saw that[18:53:16] <bfrog> when in doubt... do a find / -name "nameoffile"[18:53:33] <poseidon> O, should I include il.h and ilu.h (I thought I had to include devil.h[18:53:55] <Ragnarok> its probably in /usr/local/include[18:54:05] <bfrog> no.......[18:54:06] <bfrog> its not[18:54:10] <Ragnarok> ok[18:54:17] <bfrog> did you just read anything we were talking about... at all?[18:54:31] <Ragnarok> bskimmed[18:54:34] <Ragnarok> skimmed*[18:54:43] * Ragnarok wonders off[18:55:40] <poseidon> bfrog, is there not a devil.h?[18:55:59] <bfrog> I dunno, I don't use devil[19:06:51] *** DobosCake has joined ##OpenGL[19:20:51] <poseidon> Anyone here know how to set up devil in linux?[19:21:13] <Ragnarok> what distro?[19:21:20] <poseidon> kubuntu[19:21:32] <Ragnarok> check their package manager[19:21:40] <poseidon> Ragnarok, I installed it[19:21:47] <Ragnarok> oh okay[19:22:29] <Ragnarok> check /usr/include/IL /usr/include/devIL or just check /usr/include itself[19:23:01] <poseidon> config.h devil_internal_exports.h il.h ilu.h ilu_region.h ilut.h il_wrap.h are in /usr/include/IL/[19:24:12] <Ragnarok> poseidon, use il.h and ilu.h[19:24:47] <poseidon> Ragnarok, should I just include /usr/include/IL/il.h[19:25:06] <Ragnarok> checking[19:25:34] *** TheLorax has joined ##opengl[19:30:21] <Ragnarok> poseidon, it seems ilu.h includes il.h[19:30:30] <Ragnarok> so just use ilut.h and ilu.h[19:30:37] <Ragnarok> the documentation for it sucks >.<[19:30:59] <poseidon> Ragnarok, It does. Heres my problem http://codepad.org/OCrOCRaN[19:31:43] <Ragnarok> look at the examples in the source package on its site[19:31:54] <Ragnarok> iono i don't use devIL. I use Qt[19:32:17] <Ragnarok> wow[19:32:20] <Ragnarok> hmmm[19:32:44] <Ragnarok> paste ls -l /usr/include/IL[19:32:59] <Ragnarok> and ls -l /usr/include/GL[19:33:23] <Ragnarok> and how are you compiling, makefile, cmake etc?[19:34:14] *** stringfellow_ has quit IRC[19:34:22] *** stringfellow_ has joined ##opengl[19:34:27] <Ragnarok> im looking at that comment block at the bottom, your not including the headers in the compilation[19:35:17] <poseidon> Ragnarok, http://codepad.org/Pf3p9qNM[19:36:05] <poseidon> Ragnarok, the output is what the compiler on the server got. (it doesn't have the headers). Look at the one I wrote.[19:36:45] <Ragnarok> im looking[19:36:58] <juanmabc> poseidon: you need to link with -lILU -lILUT too ?[19:37:06] <Ragnarok> yes[19:37:14] <poseidon> juanmabc, I think I've tried. but I'll try again[19:37:40] <Ragnarok> too bad i can't get remote assistance ;)[19:37:48] <Ragnarok> or shell[19:37:54] <Ragnarok> ssh*[19:38:57] <poseidon> I get the same error if I include -lILU -lILUT[19:39:03] <juanmabc> poseidon: I think Ogl changed to GL as in ilutGLLoadImage, check your ilut.h[19:39:41] <Ragnarok> poseidon, is your devIL includes in ur path?[19:40:24] <Ragnarok> PATH=$PATH:/usr/include/IL if its not the you don't need to do -I/usr/include/IL[19:40:39] <Ragnarok> but i wouldn't recommand that[19:41:02] *** Walt_ has joined ##opengl[19:43:40] *** Nayena has joined ##OpenGL[19:43:48] *** fsvend_ has joined ##OpenGL[19:43:53] *** scy has left ##opengl[19:45:58] <poseidon> Got it to work :) it is now ilutGLLoadImage as juanmabc said![19:47:35] <Ragnarok> w00t[19:48:41] <Ragnarok> poseidon, any words of advice don't do #include </usr/include/> etc[19:49:07] <Ragnarok> g++ test.cxx -lglut -lIL -lILU -lILUT -L/usr/include/GL -I/usr/include/[19:49:12] <Ragnarok> thats what i did to compile[19:49:39] <Ragnarok> well remove the GL include part[19:49:53] <Ragnarok> b ut u get the idea[19:51:14] <Ragnarok> i love this movie[19:55:00] *** Obfuscate has quit IRC[19:56:40] *** m4ggus has joined ##OpenGL[19:58:12] *** eggauah has joined ##OpenGL[19:59:33] *** neoneurone has joined ##OpenGL[20:00:47] *** Obfuscate has joined ##opengl[20:00:56] *** WhitAnglAtWork has joined ##OpenGL[20:01:43] *** braahyan has joined ##OpenGL[20:03:08] <WhitAnglAtWork> hi! .. I'm starting with VBO... I was wondering: from a .obj file, I have a vertex array and a normals array, and 2 index arrays corresponding to references to these vertex and normals arrays. I am now wondering how I can use glDrawElements specifying that the indices to the normal list are not the same as the indices to the vertex list...[20:03:35] <WhitAnglAtWork> I guess glDrawElements assumes that the ordering in both list is the same[20:04:29] <Xmas|> yes, you can only use one index buffer[20:04:52] <WhitAnglAtWork> then how should I do ? re-order my normals ? use another function ?[20:05:03] *** stringfellow_ has quit IRC[20:06:29] <braahyan> what are the best resources for beginning opengl? anything other than nehe?[20:07:29] <hibread> WhitAnglAtWork: a fella was here the other day trying to do the exact same thing[20:07:39] <hibread> There is some work involved[20:07:57] <WhitAnglAtWork> mmm... like ?[20:08:08] <hibread> but there are 2 approaches, the naive approach and the performance critical approach[20:08:21] <WhitAnglAtWork> let's say the performance critical ?[20:09:09] <hibread> have you got texture coordinates aswell?[20:09:19] <WhitAnglAtWork> there will also be, yes[20:09:22] <hibread> does the .obj file type have smoothing groups? i can't recall[20:09:52] <WhitAnglAtWork> I won't use smoothing groups anyway[20:10:01] <WhitAnglAtWork> but there are some groups yes[20:10:20] <WhitAnglAtWork> my vertices are shared between multples submeshes[20:10:22] <hibread> i guess we'd assume that the index list is correct to start with[20:11:54] <WhitAnglAtWork> what do you mean by correct ?.. there are no negative numbers, no overflow etc.. yep[20:12:24] <hibread> WhitAnglAtWork: what you need is a "UNIQUE" list of vertex data; position, normal, texture coordinate, and tangent and bitangent if you want normal mapping.[20:12:48] <WhitAnglAtWork> ok, I know how to do that... then ?[20:13:06] <hibread> when i say unique, i mean that if you have a vertex that is shared by a few triangles, but a triangle needs a different normal for said vertex than another triangle for teh same vertex, you need to duplicate[20:13:22] <WhitAnglAtWork> mmm :s[20:13:28] <hibread> same with texture coordinates[20:13:56] <hibread> if the mesh is continuous at a poinrt, then generally a vetex will share the same normal, texture coordinates etc[20:14:12] *** TheChuckster has joined ##OpenGL[20:14:33] <Jupp3> braahyan: The Red Book[20:14:34] <hibread> but if there is a discontinuity (say the seam down the face of an animal), then you'll need to duplicate the vertex to allow for different texture coordinates[20:14:47] <WhitAnglAtWork> :s ok[20:15:03] <WhitAnglAtWork> and is the naive version much simpler ?[20:15:49] <hibread> well the naive approach would just be to walk over the index list which points to various positions/texcoords/normals, and create a list directly from that. The list will contain mountains of redundant data[20:16:12] <hibread> and you wont need the index list anymore. You'd just use glDrawArrays() instead of glDrawElements()[20:16:22] *** korff has joined ##OpenGL[20:17:07] <WhitAnglAtWork> mmm... arg[20:17:31] <Jupp3> hibread: Or you can just do glBegin(); go through both lists and do corresponding glNormal and glVertex calls :)[20:17:33] * Jupp3 runs[20:17:41] <Jupp3> But seriously, I have seen that done[20:17:42] <WhitAnglAtWork> lol[20:18:07] <Jupp3> You know, "wow, this tutorial actually uses vertex arrays!" and then "W T F..."[20:18:10] <hibread> Jupp3: well ofcourse you can do that :)[20:18:39] <hibread> but thats the same as the naive approach above, but with immediate mode as opposed to using vbo's[20:18:48] <WhitAnglAtWork> and there is just no way of using 3 index buffers ? :s[20:18:50] <Jupp3> And they actually have glArrayElement() function too[20:18:57] <hibread> WhitAnglAtWork: nope :)[20:19:04] <Jupp3> hibread: No, that's way worse approach[20:19:11] <Jupp3> Incompatible and way slower[20:19:17] <WhitAnglAtWork> damn... ok.. thanks, I'll work on that :)[20:19:32] <Jupp3> WhitAnglAtWork: With my way it would work but it sucks, don't do that. m'kay?[20:19:47] <WhitAnglAtWork> yeah yeah of course ;)[20:20:29] *** HuntsMan has joined ##opengl[20:22:53] <Jupp3> But do like vertexarray2[i]=vertexarray[vertexindex[i]][20:22:57] <hibread> Jupp3: hmm.. you are suggesting walking over the index list and issueing the glVertex()/glTexcoord()/glNormal() per the lists?[20:23:00] <Jupp3> and same for normals and what else you had[20:23:10] <Jupp3> hibread: No I'm not, it was a joke :)[20:23:30] <hibread> hehe yeah i know :) but thats the same as what i proposd with the VBO's naive approach[20:23:31] <Jupp3> And you would be doing it in wrong order anyway[20:23:46] <hibread> order?[20:23:52] <Jupp3> vertex first[20:24:03] <hibread> the correct vertex would correspond to the correct texcoord etc...[20:24:22] <WhitAnglAtWork> mm.. do you know if I can find a small piece of code splitting the redundant attributes ? (like the first method hibread proposed)[20:24:32] <Jupp3> hibread: Your solution would be 1) Faster, 2) Compatible with implementations lacking direct mode (such as OpenGL ES)[20:25:35] *** mattn2|home has joined ##OpenGL[20:25:48] <hibread> WhitAnglAtWork: actually what i said above about the seam along the center of the head is wrong. The same texture coordinates, but its the mirrored texture coordinate about that seam which stuffs around with tangent/bitangent creation :)[20:26:34] <WhitAnglAtWork> I won't need tangents/bitangents at the moment[20:27:34] <hibread> WhitAnglAtWork: Ive gotta run, but have a good hard think about what needs to be done. The easy solution is to use an O(n^2) algorithm to do the searching of duplicates etc. I ended up using a hash container which ended being pretty fast (near on O(n) time iirc)[20:27:57] <WhitAnglAtWork> ok... so thanks for your ideas :)[20:27:59] <hibread> no...[20:28:00] <hibread> hmm[20:28:01] <Jupp3> Run, Forrest, run![20:28:03] <hibread> anywho[20:29:52] <hibread> WhitAnglAtWork: and onces you've got the new lists created (index, vertices, texturecoords, normals), you then need to optimize the sorting of the index list to best take advantage of post vertex processing cache[20:30:05] <hibread> nah im joking, but it is an optimization[20:30:50] <WhitAnglAtWork> yep ;)[20:31:03] <WhitAnglAtWork> I hope my obj exporter already optimized that ;)[20:31:22] *** bfrog has quit IRC[20:31:30] <hibread> once you create the new list, it'll be ruined :)[20:31:34] *** TheLorax has quit IRC[20:31:44] <WhitAnglAtWork> hem yep[20:32:10] <hibread> im off. Seeyas![20:32:25] <WhitAnglAtWork> see ya, thks :)[20:41:30] *** aalex has joined ##OpenGL[20:47:07] *** stringfellow has joined ##opengl[20:53:23] *** blbmp has joined ##OpenGL[20:53:55] *** blbmp has quit IRC[20:54:02] *** blbmp has joined ##OpenGL[21:00:03] *** Eforen has joined ##opengl[21:00:27] <Eforen> are there any up to date opengl tutorials?[21:00:45] *** mediogre has quit IRC[21:00:59] <Ragnarok> read the red book[21:01:02] <Ragnarok> its up to date[21:01:05] <Ragnarok> and free online[21:01:40] <Ragnarok> so is the shader book and the reference book[21:02:05] <Eforen> thank you becose this is driveing me crazy trying to lurn from nehe becose i don't know what is my falt and what is becose its out of date[21:03:38] <Ragnarok> nehe's are oka btu they are 10 years out of date almost, he is making newer ones up to date. and it shows some bad habits[21:03:53] <Ragnarok> you cfan learn from them just don't let it sink in all the way[21:04:37] *** predaeus has quit IRC[21:04:54] *** m4ggus has quit IRC[21:05:01] <Eforen> yha[21:05:33] <Eforen> i was just getting to dang confused from it becose i have never coded at this low a level[21:05:47] <Eforen> i am comeing from C# and xna[21:05:51] <Eforen> so yha[21:06:06] <Eforen> i am trying to work my way down to opengl[21:06:13] <Eforen> and C++[21:06:30] <Eforen> but also C# commenting is soooo handy at times[21:06:32] <Ragnarok> Eforen, haha[21:06:33] *** m4ggus has joined ##opengl[21:06:50] <Eforen> well i started at php[21:06:58] <Ragnarok> i started with C[21:08:21] <HuntsMan> C# commenting?[21:08:30] <Ragnarok> xml commenting convention[21:08:44] * HuntsMan prefers Doxygen[21:09:02] <Ragnarok> hehe[21:09:11] <Ragnarok> this movie was awesome[21:09:29] <Eforen> i started with basic then html then php very much then vb then C++ a little bit then a little DX and ogl with C++ back out to C++ and ogre and then C# and xna and now i am back to C++ with OpenGl[21:09:48] <Ragnarok> so hows ogre?[21:09:54] <Ragnarok> i've never messed with it[21:09:57] *** Xmas__ has joined ##OpenGL[21:10:24] *** Walt_ has quit IRC[21:10:52] <Ragnarok> I've done X/HTML, XML, PHP, C#,VB.NET,Python,IronPython,Java,C,C++, Perl, Bash, Lisp[21:11:00] <Ragnarok> DirectX 9 and OpenGL[21:11:04] <Eforen> its prity good for prototypeing and if you don't whant to do all the code your self but if your gona whant anything with fast prosessing speed you need to make it your self in C++[21:11:04] <Ragnarok> and XNA[21:11:13] *** m4ggus_ has joined ##opengl[21:11:17] *** m4ggus_ has quit IRC[21:11:35] <Eforen> o i forgot the java and javascript in there[21:11:45] <Eforen> hows lisp?[21:11:46] <Ragnarok> and javascript xD[21:11:51] <Ragnarok> didn't like it[21:12:02] <Ragnarok> it was a bitch getting it set up in linux[21:12:11] <Ragnarok> and still could never figure out how[21:12:22] <Eforen> me eather but its almost crushal to have when your doing websites these days[21:12:45] <Ragnarok> so i've basically done almost every language, long language i stayed with was C# for 9 months[21:12:54] <Ragnarok> currently i've been with C++ for 6 months[21:12:56] *** stringfellow has quit IRC[21:13:00] <Ragnarok> give or take[21:13:22] <Eforen> i realy liked some of the stuff in C#[21:13:25] <Ragnarok> C3 imho was slow[21:13:29] <Ragnarok> C#*[21:13:45] <Ragnarok> oh and i've done MASM, and a little NASM[21:13:54] <Eforen> like the xml commenting and the intelisence with that commenting but that is exacly why i am back here to slow[21:14:17] <Ragnarok> i'd enjoy C++ more than any other language[21:14:22] * keitsi <3 C#[21:14:25] <Eforen> i don't reconise MASM and NASM...[21:14:42] <Ragnarok> its microsoft assembly and the other is cross assembly[21:14:50] <keitsi> Ragnarok: it all depends if C# is "fast enough" for your project[21:14:51] <Eforen> o ewww[21:15:17] <Ragnarok> i want to learn assembly but every one i found hasn't let me work with it :([21:15:40] <Eforen> i enjoy C++ more then any othere lang becose of the power it lest you have over the code but its debuging sucks sometimes[21:15:40] <Ragnarok> i want to embed assembly into my opengl project xD[21:16:01] <Ragnarok> C++ in VS was awesome but not in linux[21:16:07] *** m4ggus has quit IRC[21:16:33] <Ragnarok> the only debugging i hated in VS was it would like to open up standard headers which i never cared to go through and just wanted to go through mine only[21:16:50] <Eforen> when i make the linux port of my game i will be codeing in vs on win or wine[21:17:19] <Ragnarok> just use linux[21:17:28] <Ragnarok> u can compile windows apps on linux :D[21:17:42] <Eforen> yha but why?[21:17:43] <Eforen> lol[21:17:48] *** TheLorax has joined ##opengl[21:17:53] <Eforen> better to just do it here in windows[21:18:27] *** m4ggus has joined ##opengl[21:18:28] <Eforen> i dual-boot vista and ubentu with a gentoo network server[21:18:46] <Ragnarok> well not really because using the VS C++ compiler doesn't produce standard coding unlike g++[21:19:11] <Ragnarok> Linux Narf 2.6.26-rc3-zen0 #1 SMP Fri May 23 18:20:22 CDT 2008 i686 AMD Athlon(tm) 64 X2 Dual Core Processor 3800+ AuthenticAMD GNU/Linux[21:19:12] <Eforen> realy how so?[21:19:44] <Ragnarok> you'll see when you do it in vs and then try and compile it in g++[21:20:16] <HuntsMan> that's the code YOU produce, not the code VS produces :P[21:20:24] <Ragnarok> one thing i miss about C# I wish C++ was full OO and not here and there[21:20:41] <Ragnarok> VS lets you by with alot of shit unlike g++[21:20:43] <Ragnarok> xD[21:20:50] <Eforen> Vista Ultimet Intel Core2 Quad CPU Q6700 @ 2.66GHz Memory 8GB[21:20:58] *** prophile has joined ##opengl[21:20:59] <Ragnarok> lucky clit[21:21:15] <Eforen> lol[21:21:17] <Eforen> thank you[21:21:24] <bobbens> 65 nm eww :P[21:21:32] <Ragnarok> imma be getting me a Q9550 soon[21:21:46] <bobbens> already need an upgrade? >_>[21:21:47] <Ragnarok> 45nm[21:21:48] <Ragnarok> xD[21:21:50] *** stringfellow has joined ##opengl[21:21:52] <bobbens> i know the naming scheme :P[21:22:03] <Eforen> i forgot to say the dual sli nvidia 8800[21:22:17] <Ragnarok> i really don't care about sliing xD[21:22:22] <Ragnarok> just one card is enough[21:22:30] <bobbens> sli is pretty inefficient[21:22:37] <bobbens> powerwise and moneywise :)[21:22:37] <Ragnarok> unless u want ur electricity bill sky rocket xD[21:23:22] *** Xmas| has quit IRC[21:23:39] <Eforen> true but is nice to have 4 monitors and have both sets have a 8800 between them[21:24:17] <Eforen> i use them unsli and that way i have multipul monitors and extream power for each[21:24:23] *** hibread_ has joined ##opengl[21:24:37] <Eforen> i have never laged on a properly coded game[21:24:43] *** hibread has quit IRC[21:24:45] <bobbens> i need more desktop so I can actually fit a second monitor :P[21:24:57] <bobbens> aren't there some cards with 4 outputs already each?[21:25:17] <Eforen> i need more desktop so my mouse can move mor[21:25:18] <Eforen> e[21:25:43] <Ragnarok> do u really need 4 monitors?[21:26:03] <Eforen> i think theres a fuw but there nowere near as powerfull as a 8800 or atlest the ones in the price range[21:27:09] <Ragnarok> i want the 9800 GX2[21:27:17] <Eforen> you know whats cool is there are new ram cards that are 4g each![21:27:31] <Eforen> that means a total on most boards of 16g![21:27:31] <bobbens> more ram then my box >_>[21:27:42] <Eforen> i know i whant one of thos too[21:27:49] <bobbens> i have 2 GB[21:27:53] <bobbens> don't really need more :)[21:27:54] <Eforen> hehe[21:28:01] <Eforen> i am running vista[21:28:03] <Eforen> lol[21:28:15] <Eforen> but when i am booted into linux i am fine[21:28:21] <Eforen> i screem[21:28:23] <Eforen> lol[21:28:35] <Eforen> 8gb ram in linux is crazy[21:28:58] <bobbens> 2 GB is overkill for everything I do[21:29:01] <bobbens> but i'm not much of a gamer[21:29:06] <bobbens> nor do I have gigantic databases :)[21:29:09] <Eforen> i am ^.W^[21:29:25] <bobbens> i'm a coder :)[21:29:30] <bobbens> you just need a decent CPU for compiling :)[21:29:35] <Eforen> i am both[21:29:40] <bobbens> and E6600 is more then good enough :)[21:29:59] <Ragnarok> i only have 1.5GB RAM[21:30:01] <Eforen> i have a Q6700 @ 2.66GHz[21:30:07] <Ragnarok> I need 1GB more and ill be satisfy[21:30:08] <Eforen> that works well[21:30:17] <Ragnarok> but would love to have 8GB RAM[21:30:26] <Eforen> lol[21:30:33] <Eforen> it is very nice[21:30:34] <Ragnarok> but that means i gotta move to 64bit[21:30:35] <bobbens> quad just seems like a waste of processors, i don't think you actually get to use them all much together :)[21:30:46] <Eforen> and the way i brouse in firefox i need it hahah[21:30:47] <bobbens> i wouldn't use 8 GB ever[21:30:55] <bobbens> firefox is my only hog[21:30:56] <Ragnarok> bobbens, some games use all of them[21:30:59] <Ragnarok> i wish i knew how[21:30:59] <Eforen> i have used all of that in firefox befor hahaha[21:31:01] <bobbens> it gets up to 900 MB on bad days[21:31:34] <Ragnarok> mine gets up to 600MB max i think[21:31:40] <Ragnarok> on gentoo[21:31:40] <Eforen> i will have like 20-30 firefox windows with around 30-80 tabs in each[21:31:49] <Ragnarok> o_O[21:32:04] <Eforen> on a bad day my'n is up at 5-6GB[21:32:24] <Eforen> then agean thats what i get for useing windows[21:32:25] <Eforen> lol[21:32:46] *** ViRUS has joined ##opengl[21:34:37] <Eforen> are there any tutorials for biginers in opengl 2?[21:34:44] <Eforen> wait has 3 come out?[21:36:47] *** TheLorax has quit IRC[21:37:09] <Eforen> also any ideas of how i can do a face morph type thing so users can make the face how ever they whant[21:37:31] <RTFM_FTW> use the programmable pipeline[21:37:41] <Eforen> hu?[21:37:55] <RTFM_FTW> i.e. stuff like ARB_fragment_program, ARB_vertex_program or The OpenGL Shading Language[21:38:28] <RTFM_FTW> using a texture containing the user's face which you manipulate via the GPU[21:38:30] *** rsaltini has joined ##OpenGL[21:38:31] <Eforen> for changeing the shape of a face model?[21:38:34] <RTFM_FTW> using the above extensions[21:38:36] <rsaltini> hello[21:38:52] <rsaltini> if GL_VERSION gives me "1.3 Mesa 7.0.1"[21:38:55] <Eforen> no no not the uses real face i mean a model head[21:39:08] <rsaltini> but mesa 7.x should implement opengl2.1[21:39:09] <RTFM_FTW> if the fact is a vertex model then you can simply mutate the vertex data directly[21:39:26] <RTFM_FTW> then you can update it via VAs (a optimal approach being VBO)[21:40:04] <RTFM_FTW> you can even re-tesselate the structure for additional detail on the GPU via geometry shaders[21:40:27] *** TheLorax has joined ##opengl[21:40:32] <RTFM_FTW> you can also warp and mutate the model directly on the GPU via the programmable vertex pipeline[21:40:43] <RTFM_FTW> which saves you the cost of updating the data[21:41:17] <Eforen> ok...[21:41:40] * Eforen is a little lost but thats ok[21:43:21] *** fsvend_ has quit IRC[21:46:24] *** rsaltini has quit IRC[21:48:24] *** nplus_ is now known as nplus[21:54:48] *** mm765 is now known as mm765^away[21:55:48] *** TheChuckster has quit IRC[22:04:30] *** korff_ has joined ##OpenGL[22:09:57] <Eforen> i just tryed to comple the example at the bottom of http://www.glprogramming.com/red/chapter01.html and it is throwing all kindsa errors at me from gl.h[22:10:22] <Eforen> over 143 errors[22:11:25] <HuntsMan> pastebin them[22:12:14] *** predaeus has joined ##opengl[22:12:22] *** korff has quit IRC[22:16:18] <Eforen> http://pastebin.com/d2a00b7e4[22:18:39] <Eforen> any ideas HuntsMan?[22:20:11] <HuntsMan> sounds like bad headers[22:21:47] <MatthiasM> show the code where you include gl.h[22:22:47] <Eforen> MatthiasM: http://www.glprogramming.com/red/chapter01 the last code example[22:23:29] <Eforen> mabey i have old headers and libs where can i get the curent ones?[22:23:34] <MatthiasM> try to include windows.h first[22:25:11] *** fargiolas has joined ##OpenGL[22:25:16] <Eforen> now i get a linker error[22:25:21] <Eforen> Error 1 error LNK2019: unresolved external symbol _WinMain@16 referenced in function ___tmainCRTStartup MSVCRTD.lib[22:25:41] <MatthiasM> compile you app as console application[22:25:57] <Eforen> i thoght you could not do that?[22:26:12] <MatthiasM> you can /subsystem:console[22:26:26] <MatthiasM> it's a linker option[22:26:48] <Eforen> sweet got it[22:27:11] <Eforen> but i was told as some point that opengl aps had to be win32 not console?[22:27:46] <Eforen> thank you tho that fixed it much thanks[22:28:29] <MatthiasM> it doesn't matter - you can open windows from a console app - and you can open a console from a windows app[22:29:42] *** thesquib has joined ##OpenGL[22:32:23] *** void256 has joined ##opengl[22:37:47] *** wersf has quit IRC[22:40:47] *** lolage0 has joined ##OpenGL[22:43:22] *** braahyan has quit IRC[22:48:34] *** Ademan has quit IRC[22:48:51] *** braahyan has joined ##OpenGL[22:55:46] *** void256 has quit IRC[22:59:46] *** predaeus has quit IRC[22:59:59] *** Walt has quit IRC[23:03:17] *** Ademan has joined ##OpenGL[23:04:31] *** rsaltini has joined ##OpenGL[23:07:24] *** fargiolas has quit IRC[23:08:50] *** Ademan has quit IRC[23:11:32] <rsaltini> can't i have opengl 1.5 over ati rv250?[23:16:53] * thesquib is away - bbl - [AcidJazz - Log ON][23:18:45] *** Arc_ has joined ##OpenGL[23:19:44] *** jparishy has joined ##OpenGL[23:19:52] <jparishy> Hey,[23:20:41] <jparishy> is there a way to use ortho where (0, 0) is teh top left and everything is ...upside up?[23:21:05] <jparishy> I am doing glOrtho(0, width, height, 0, -1, 1); and it does that, but everything is drawn upside down[23:21:43] *** Ademan has joined ##OpenGL[23:24:13] <MatthiasM> lol[23:24:18] *** servus_ has joined ##opengl[23:24:44] <MatthiasM> set (0, -height) as bottom left then :D[23:25:27] <jparishy> yeah, that's what i was doing[23:25:33] <jparishy> it just kinda bugs me :([23:27:37] <MatthiasM> if you swap the coordinates from top and bottom - every thing is ofcourse bottom up if you don't fix the coordinates of the objects[23:30:56] <rsaltini> i've asked this also yesterday.....but it doesn't convince me[23:31:28] <rsaltini> i have mesa 7.0.1 library[23:31:44] <rsaltini> anyway GL_VERSION returns me "1.3 Mesa 7.0.1"[23:31:57] <rsaltini> mesa 7.x should implemente opengl 2.1[23:32:22] <rsaltini> and shouldn't opengl be indipendent from the hardware?[23:32:35] <MatthiasM> no[23:32:41] *** TheChuckster has joined ##OpenGL[23:32:51] <rsaltini> shouldn't they implement by software what lacks in hardware?[23:33:05] <bobbens> software fallbacks are slow[23:33:11] <bobbens> and will totally kill performance[23:33:47] <HuntsMan> rsaltini: besides, free drivers don't always support all features of the HW[23:33:57] <HuntsMan> as documentation of the hardware internals is lacking[23:33:58] *** Arc has quit IRC[23:34:20] <MatthiasM> check if you can force full software rendering :)[23:34:23] <bobbens> or manpower is lacking when they have the documentation :P[23:35:26] <rsaltini> Are you saying that mesa 7.0.1 is capable of opengl2.1 but only with specific video cards?[23:36:01] <HuntsMan> yes[23:36:11] <bobbens> or software implementation i believe too[23:36:12] <HuntsMan> not all cards support GL 2.1 BTW[23:36:18] <HuntsMan> yeah too[23:37:24] <rsaltini> my professor today said me that opengl version don't depend on hardware[23:37:49] <rsaltini> my problem is that i'd like use shadow map[23:37:56] <bobbens> opengl is a standard, independent of hardware, but you won't have hardware accelerated functions on all cards[23:37:59] <rsaltini> but with opengl 1.3 i can't[23:38:09] <bobbens> and software fallbacks kill the entire point of hardware accelerated opengl[23:38:35] <rsaltini> ok, but at least software fallbacks should exist[23:38:51] <bobbens> you can force full software fallback[23:38:56] <rsaltini> how?[23:39:05] <bobbens> it can actually get decent speed depending on what you're doing and with what CPU[23:39:16] <bobbens> option NoAccel "on" <-- or something like that[23:41:12] *** servus has quit IRC[23:44:13] <HuntsMan> that's to disable XRENDER[23:45:02] <rsaltini> i'd like opengl uses direct rendering up to video card capabilities ang then uses software fallback[23:45:05] <rsaltini> is it possible?[23:45:33] <HuntsMan> that's driver dependant[23:45:39] <HuntsMan> what capabilities you need?[23:46:20] <rsaltini> shadow mapping[23:46:34] <HuntsMan> that's a technique, not a HW capability[23:46:50] *** PHilton has joined ##OpenGL[23:46:57] <rsaltini> in opengl 1.3 is not possible to copy depth buffer to texture[23:47:30] <HuntsMan> use FBO's[23:47:36] <PHilton> hi, how do i know which diagonal of a quad will be chosen?[23:47:50] <MatthiasM> PHilton: you don't - use triangles[23:48:20] <rsaltini> so you are saying me that i can implemente shadow map using opengl 1.3[23:48:21] <rsaltini> ?[23:48:25] <PHilton> MatthiasM I see, thanks a lot.[23:48:55] <HuntsMan> yes, using GL_EXT_framebuffer_object[23:51:06] <rsaltini> is that the only way? can i do without using that extension?[23:51:20] *** LordMetroid has joined ##OpenGL[23:51:44] *** LtJax has joined ##opengl[23:52:09] <HuntsMan> sure, using pbuffers, but it will be a pain in the ass[23:52:21] <MatthiasM> rsaltini: check if you have GL_ARB_Shadow extension[23:52:39] *** PHilton has left ##OpenGL[23:53:11] <RTFM_FTW> you aren't going to get FBO support on a GL 1.3 level renderer so mentioning "use FBOs" is somewhat of a moot point here[23:53:27] *** Xmas__ has quit IRC[23:53:50] <RTFM_FTW> in any case as long as you have support for depth texturing (i.e. ARB_depth_texture) you can implementation the R compare through the FF path[23:54:05] <RTFM_FTW> via ARB_shadow, EXT_shadow_funcs, ARB_shadow_ambient and friends[23:54:26] *** jann has quit IRC[23:55:04] <RTFM_FTW> if you have shader support as well you can use either ARB_fragment_program or GLSL (via the shadow*(...) methods) using ARB_fragment_program_shadow[23:55:43] *** TheChuckster has quit IRC[23:57:22] *** BadChoice has quit IRC[23:57:25] <rsaltini> glxinfo | grep depth say nothing[23:58:52] <rsaltini> i think i can give up[23:59:33] *** TheChuckster has joined ##OpenGL[23:59:52] <MatthiasM> rsaltini: try stencil shadows