Slashdot Mirror
The State of OpenGL
CowboyRobot writes "No longer vapor, but a true 3D-embedded engine, OpenGL is on the move. Pixar and others would love to be able to render their movies in realtime, and that desire has prompted the intended release of OpenGL 2.0, due in a few months. Khronos is now in charge of further extending OpenGL to cellphones and handheld gaming devices."
Pixlet is for playback (a codec), not for rendering films. Unless I'm missing something in your post.
The guitars sound good, now give me about 10db more on the cow bell.
I thought we told you not to come back here after that Monkey Boy Dance incident, Steve.
When will they figure it out OpenGL is not necessarily desirable in a cellular phone?
I want business class reliability, not a the ability to rent subpar games on my cell phone for $5/month.
When I'm on the phone all day because of my work I want it to be there for important calls, not fizzle out after an hour because it's got a 640x480 pixel screen with 24-bit color.
Although right now OpenGL is all that's out there for low-cost portable embedded 3D software, no one is going to develop with it until hardware support emerges. Who wants a handheld 3D mapping device that takes 10 seconds to redraw a frame using an ARM9 software renderer?
I don't know about availability, OpenGL is cross-platform (works on OS X, Linux, Windows, etc.) while DirectX is Windows only. OpenGL is also included with many if not all graphic cards. So it's just as widely available, if not more, than DirectX.
What about the ability to stick to-do notes on the BACK of your cellphone? Wouldn't that make mobile 3D worthwhile?!
Don't call me a cowboy, and don't tell me to slow down!
For so long, DirectX had to struggle and claw to keep up with OpenGL - they did just that, while OpenGL sat mainly idle (well, John Carmack was a big help to it)... Now it seems the shoe is on the other foot, and OpenGL is going to have to move deftly to surpass DX9, and soon enough 10...
I sincerely hope it happens. I wish developers felt more inclinded to make their 3D engined GL based rather than DX based, so the day where I can play any game in linux may actually arrive. Of course, we have to give massive amounts of respect to those who do make OpenGL platforms for their games (ID, Epic), but what about those who feel DX is easier and more practical for what they do(Valve).
Maybe if we're lucky, the Carmack will drop in to this discussion and tell us exactly what he thinks needs to happen to really make GL a reality for most gamaes again.
"Touch one hair on her head and I'll render you limb from limb!"
Hopefully, this will prompt more developers to join efforts to create a feature rich gaming framework for *nix. SDL is a great start, but lags behind DirectX in a number of ways. I look forward to seeing this 2.0 release breathe new life/blood into this area of development.
Thank you for your time,
BBH
OpenGL is used in the Torque engine alongside Direct3D (D3D on Windows, OpenGL on Mac and Linux). It would be great if OpenGL could eclipse Direct3D, and become the premiere 3D platform once again. Perhaps we will see this with the release of OpenGL 2.0, but for a few years Direct3D has been slowly but surely catching up and then surpassing the aging OpenGL standard.
A lot of our customers demand Linux in their solutions (networked gaming terminals) to avoid the cost of licensing Windows XP Embedded for each machine, and the option so far has been to go the Mesa/OpenGL/SDL route (WineX is still too slow for what we do), which, while it has worked, is technically slightly inferior to our Windows equivalents. Hopefully OpenGL 2.0 will change this.
I'm going to learn opengl in a few months during the holidays before I start my PhD. I work with simulations of the Sun and use IDL to visualise the results. but I think it would be cool to have more "realistic" pictures, plus having the hardware acceleration has benefits when dealing with a lot of data (IDL gets real slow for 2D simulations at resolutions above a few hundred squared)
these simulations are done on beowulf clusters (imagine that!) so I think opengl is the best (the only other API I know of being directx)
Say what?
You don't get much higher-level than a scenegraph API like Java3D.
I think the author may have been confused, although he did get the overall point right. OpenGL ES on J2ME will probably be the way this goes.
FUNK!
OpenGL 2.0 is not as exciting as the new major version number might indicate. Probably the most important new feature of OpenGL 2.0 was going to be the GLSL high level shader language. However, in order to speed up its support by hardware companies, this was instead put into OpenGL 1.5 spec when it was announced last year; GLSL already has implementations by 3DLabs, ATi and nVidia. OpenGL 2.0 will still add some useful new features, but it won't be the world-shattering event that 3DLabs promised in their original proposals.
When the article talks about rendering in real-time it isn't talking about the compressed/flattened video playing a full frame rate, it's talking about OpenGL being able to calculate/shade/render a model in realtime verses waiting X mins/hours for a frame to render. It's talking about the process of converting vector data to raster data.
The guitars sound good, now give me about 10db more on the cow bell.
...Microsoft rules. They really are the most innovative and resilient software companies in the world.
So you acknowledge that they should be broken up into multiple companies? Groovy!
Jan 16th 2002: SGI transfers 3D graphics patents to MS
Jul 09th 2002: Microsoft Claims IP Rights on Portions of OpenGL
Jul 11th 2002: 3D graphics world shaken by patent claims
Jul 13th 2002: Microsoft patent claims may affect OpenGL
Mar 3rd 2003: Microsoft quits OpenGL board
My cousin's husband works for Sun and he said that the next version (1.5?) of Java will have Swing ported to OpenGL underpinnings... that way, even 2D apps will be MUCH faster.
He said they're realizing 4X speed increases on plain old 2D apps.
They're also working on making 3D game demos (some with 3rd parties) to demo that Java can actually now compete in the desktop game market...
---- It puts the lotion on its skin or else it gets the hose again. It does this whenever it's told.
No longer vapor, but a true 3D-embedded engine...
Since when has OpenGL been vapor?
Which could be your target as a glowing orb, or a character in of a video game super-imposed on the actual landscape, or the trail your friend took through the same city two years ago, or just some construct representing an interesting thing about your environment, or ...
I think that would be a real killer app.
For one, direct3d is integrated into the direct api which handles a multitude of things, multimedia and game input devices among others, that game developers are almost naturally drawn to by the appeal that so much work has already been done for them
OpenGL can't and really shouldn't have to address all these requirements, but it's just part of why there's been this ongoing struggle. SDL is a reasonable answer to portability while still accomplishing the integration that MS has achieved, but SDL isn't really as mainstream as OpenGL is.
I've seen soap opera plots that were less convoluted than this mess.
File under 'M' for 'Manic ranting'
The original article post seems to confuse different forms of OpenGL. OpenGL|ES is the embeded stripped down OpenGL for mobile & embeded systems. OpenGL 2.0 is just a proposal from 3DLabs and may never get off the drawing board. Most of the significant changes that OpenGL 2.0 introduced have been implemented and released either as extensions or as part of OpenGL 1.5, so it's just not clear if or when OpenGL 2.0 will actually arrive, there's a lot of resistance because 2.0 intended to throw some stuff out and many developing, selling & using OpenGL implementations think that it's a REALLY bad idea to do that. With OpenGL|ES there is already a version 1.0 and you can actually get this in several forms from implementations that run on phones to wrappers around OpenGL that you can use on the desktop to emulate OpenGL|ES. OpenGL|ES is in the process of developing version 1.1 right now.
The notion that Pixar would use OpenGL for final rendering if only it were fast enough comes up every time a new video card or GL enhancement comes along just indicates how little people understand how Pixar actually makes their films. Oddly, Pixar really doesn't make this information much of a secret, and they'll even sell you the same software they use.
Most frames in Pixar movies are rendered using some form of ray-tracing.
Technically, no. Renderman (the Pixar renderer) does not perform ray tracing. It uses a scanline renderer that is much faster than any ray tracer I've ever seen. They've been at this for literally decades, and are very good at it. Still, the most complex images in their movies can take many hours -- sometimes more than a day -- to render. The time-to-render doesn't seem to improve much from picture to picture because as computers get faster, they just add complexity to the scene.
As TD who works in the computer graphics field, let me state that the technology required to render a Pixar film in 'Real Time' is far off and ridiculous. Just because OpenGL looks better does not mean that it can support the shader functions that Renderman utilizes, not to mention the Fur and cloth APIs. Also, the majority of shots in movies aren't even single comp shots they involve many rendered elements, which you still have to comp together. I'd be all for the guy talking about how OpenGL 2.0 will benefit the artists by allowing them to get more feedback about the quality of the shot they're working on without preview renders, but thinking that OGL could replace final renders any time soon is wrong. Perhaps we are geting to a place where we could render the original Toy Story realtime and a general viewing audience might not know the difference. Perhaps. But I remember some really great PRman shaders from the film that wouldn't be posible in the real time version.
No, you just need to buy two phones, one for the left eye, and one for the right eye, and a special new 'hands free' kit that holds the phones in front of your face.
1) Its Apple's implementation of GL that's less than perfectly optimized. On Windows and Linux, OpenGL is as fast as D3D.
2) OpenGL has numerous releases in the last few years. 1.3, 1.4, and 1.5 were all released in quick succession. What rock have you been hiding under?
A deep unwavering belief is a sure sign you're missing something...
"Over the next year or two, I think you're going to see a whole range of applications that use your graphics board as a supercomputer," Trevett says enthusiastically.
was the most interesting part of the article?
SETI@home, Finite Element Analysis, video recoding are all areas which could benefit from vector processing , matrix calculation and/or huge register sizes provided by GPUs.
Peercy and Olano (Click on "PDF" in the upper right)
Presentation
ASHLI
GPGPU
More than Moore's Law
Moore's law : still for wimps
Using programmable graphics hardware (possibly through OpenGL) for final rendering is not that far off. (Definitely not in real-time, but as a more cost-effective way to do it, anyway.) Especially with the massive parallelism of rendering, and the fact that GPUs are far outpacing CPUs in terms of their speed and transistor counts.
OpenGL is much more similar to micropolygon rendering (REYES) than it is to raytracing in the first place. The shaders are where you spend all of your time, anyway.
Heck, do you think nVIDIA bought ExLuna (Larry Gritz, author of BMRT, and former Pixar employee) just for the fun of it?
Software for translating from RenderMan Shading Language to Cg?
And what about RenderMonkey supporting RenderMan?
Do you even remember PixelFlow from Pixar? Do you see the name Marc Olano on that paper? The same Marc Olano who talks about rendering on consumer-level graphics hardware? These things have far more in common than you seem to realize.
Education is the silver bullet.