WebGL around the net, 7 April 2011

Oof, still quite a backlog of links to get through — a new project at work called PythonAnywhere is making “free time” a distant memory — but there are still some great links this week:

You can leave a response, or trackback from your own site.

23 Responses to “WebGL around the net, 7 April 2011”

  1. I’ve been wasting many hours I could be putting on my app playing cyclebob :p Now you just did the same to a bunch of other people I bet.

    That list of framework was nice. I have to second his comments about O3D, my application is based on it and while the library is fairly complete, Google doesn’t seem to want to invest any new efforts into it. It works fine but the lack of support by google prevents other from trying it which is sad.

  2. Florian says:

    I didn’t like O3D. I was hunting around for frameworks before rolling my own, and O3D was just… convoluted.

    Anyway, I just went trough the list of frameworks, and they pretty much all suffer one or the other of these issues:

    * They allocate objects at rendertime which kicks in the GC which makes animation extremely unsmooth.
    * Lacking (or completely missing) examples
    * Pages not responding
    * Lacking documentation
    * Awful API styles
    * Many common things left unsolved and as an exercise for the user
    * Other common things over-abstracted for no good reason

    I feel like most of these frameworks aren’t target/application driven, and those that are, are incredibly cought up in their own idiosyncracies.

    :)

  3. brainjam says:

    @Florian, your critique of the frameworks is fair enough, but they are still valuable as high level examples for all sorts of techniques (both in terms of graphics operations and Javascript software engineering).

    Also, I’ve found that in at least one case (GLGE), the author has responded graciously to problem reports and has made efforts to improve things, and I’d expect that to be the case in general. It’s early days yet.

    By the way, I’d echo some of your comments to some degree when it comes to math/matrix/vector frameworks. They are useful, but at the moment aren’t quite giving me the functionality and reliability I need. At present I use a combo of three frameworks, each to cover the shortcomings of the others.

  4. cooliobject says:

    Any one have problems with shadows demo on XP/Firefox? I see one frame of correctly rendered scene, after that it’s as if completely in shadow.

  5. cooliobject says:

    Works on XP/Chrome though, when experimental-webgl context is replaced by webkit-3d. I’ve had problems with firefox previously, too, so it’s probably safe to blame them and not the demo.

  6. Florian says:

    @coolieobject

    Webgl is a standard, and as such, a webgl user can’t do any more then code against that standard. If that standard compliant code fails to work for some users, there’s jack the webgl user can do about it, because, by definition, it’s not his fault, ever.

    It’s not like html or css, and you can do some hacks to make the problem go away. For the most part, if there’s a bug in some driver, operating system and browser combination with webgl, there’s no way for the webgl user to “fix” it by some hack.

  7. Benoit Jacob says:

    @coolieobject, @Florian, despite WebGL being a standard and having good conformance tests, there still are various factors that can explain that a demo would work in one browser and not in another; chief among them is the fact that different implementations hit different driver bugs. If you see bad rendering in Firefox, the first thing to do is to go to the JS error console (Ctrl+Shift+J) and see if there is a WebGL warning (webgl.verbose in about:config).

  8. Benoit Jacob says:

    That said, I can reproduce this bug here on NVIDIA/Linux so it is indeed looking like a Firefox bug.

  9. brainjam says:

    (cooliobject said:) “Works on XP/Chrome though, when experimental-webgl context is replaced by webkit-3d”

    Anybody know the situation on what allowable context names are? “webgl” doesn’t appear to be supported, “”moz-webgl” isn’t supported by Firefox, “webkit-3d” works with Chrome. Why is “experimental-webgl” still around? And is “webkit-3d” different from “experimental-webgl”? How should we write our WebGL apps so they get the right context in future browser versions? Is “experimental-webgl” going away anytime?

  10. Benoit Jacob says:

    @brainjam: “experimental-webgl” is the only thing that you need to use, for the time being. As browsers start passing 100% of the conformance tests, they will start supporting “webgl”, but “experimental-webgl” will obviously stay supported for compatibility.

  11. Jon says:

    @giles and @other blogers:

    I need to do a 3D app to simulate a process in real time. I have read about WebGL and other things and was wondering what you all think about X3D (X3DOM)?? What should I use for my app?? are there other things (better ones) I could use?

    I’m not an expert web-programmer (only know C and assembler) but I have time to learn.

    Thanks xD

  12. DK_wu says:

    I’m a Chinese boy.I like 3D development.Also in reference to the moderator document.But there are many places don’t understand.My English is not good.Please take help!thank you!

  13. Erik says:

    Incidentally, if anyone is curious as to how close their browser is to full conformance go here: https://cvs.khronos.org/svn/repos/registry/trunk/public/webgl/sdk/tests/webgl-conformance-tests.html

    Mine (chrome 10, windows vista) passes 5782 out of 6002 tests so it’s pretty close.

  14. Lindsay says:

    Check this out – Stephen Bannasch has provided some graphical benchmarks of matrix libraries intended for use with WebGL:

    http://stepheneb.github.com/webgl-matrix-benchmarks/matrix_benchmark.html

  15. Lindsay says:

    We submitted the BioDigital Human to Chrome Experiments a couple of weeks back .. no response ;)

  16. DK_wu says:

    Hello,I am a beginner,English is not good,May I ask how the transformation model.Use .dae into json format it?

  17. mariuz says:

    Thanks for the update
    Also the google-chrome –ignore-gpu-blacklist really rocks

    I use nouveau and radeon open source drivers and they are really capable of rendering simple levels from quake3 in webgl

    So i keep webgl enabled on all opensource drivers … even on intel ones

    What to add in the next update ? Hellknight demo works again
    http://blog.tojicode.com/2011/04/hellknight-demo-works-again.html

    WebGL Starter Package
    http://blog.tojicode.com/2011/04/webgl-starter-package.html

    Quake3 , Doom3 demos updated to use requestAnimationFrame
    http://blog.tojicode.com/2011/04/requestanimationframe.html

  18. mariuz says:

    What to add in the next update ?

    Advanced WebGL – Part 1 & 2 & 3
    http://codeflow.org/entries/2011/apr/11/advanced-webgl-part-1/
    http://codeflow.org/entries/2011/apr/13/advanced-webgl-part-2-sky-rendering/
    http://codeflow.org/entries/2011/apr/18/advanced-webgl-part-3-irradiance-environment-map/

    WebGL Capabilities Analyzed
    http://codeflow.org/entries/2011/apr/16/webgl-capabilities-analyzed/

    ps:For me the above demos didn’t worked on radeon open source driver
    lshw -C video
    description: VGA compatible controller
    product: RV710 [Radeon HD 4350]

  19. wheelie says:

    Hi there!

    Just a little comment on ‘A simple physics demo from Stephen: a rolling ball.’

    The Player object sets the global variable shaderProgram as it’s own, however the shader parameter is for that purpose:

    function Player(shader, level) {
    // variables
    this.shaderProgram = shader; //<- CORRECT
    //this.shaderProgram = shaderProgram; <- TYPO

  20. Stephen says:

    @wheelie Thanks for that catch. Did it cause an error for you or were you just digging through the source?

  21. wheelie says:

    @Stephen No errors, just noticed it in the source code :)

  22. giles says:

    @Nicolas — yup, I was feeling cruel ;-)

    @Jon — X3DOM looks pretty good, though it is a very different approach to the other frameworks, being markup-based instead of code-based.

Leave a Reply

Subscribe to RSS Feed Follow Learning WebGL on Twitter