Missing things TOP3: Graphiccard Acceleration

Calls for graphiccard acceleration for the flashplayer can be dated to over 7 years ago ? Anyway, the fast broadband internet is reality and it’s not necessary to reduce the player size under 1MB or whatever the internal preferences are, anymore. It’s not only about 3D. As a developer, I want to create new effects besides the ‘presets’ in the flashplayer. The new bitmap effects are nice, but too slow to combine them for new effects. With graphiccard acceleration we could source out a lot of computations to have more power on the application site.

It’s not just a feature, its the possebility to extend the player heavily with new effects in actionscript.

8 thoughts on “Missing things TOP3: Graphiccard Acceleration”

  1. Many years ago (2001?) I saw a Flash player (perhaps Flash V3.0, unofficial, beta) using OpenGL to draw any (non AS) flash movie.
    It was incredibly fast and you could “split” the various layers and “turn around” the running movie … but no Action Script, so it was almost useless, not considering that was a stand-alone application.
    Now that Mac and PC have both a good GFX card support, Macr… Adobe could think about it!

  2. I really would like to see this too.
    Every time you try something new ( video with
    alpha layers, blurring, … ) the computer
    starts to choke.

    I actually want to use these features at a proper
    speed. Graphiccard Acceleration would make this possible. Good luck Andre with your TOP3 mission ;)

  3. the real question is: how will the acceleration be accomplished?

    When the flashplayer runs on different systems it will have to use different APIs.
    DirectX, OpenGL, whatever…

    So who can garanty that the results will look exactly the same right down to the very pixel??

  4. nay,it should use open gl on all OS as that´s most widly available.
    I strongly support this feature request,let´s go for it :)
    Years ago i made an april first joke on FK where i posted fake images of a new flash version which had 3d capabilities (i took some hours to modify the IDE in photoshop to feature several cam views etc..)
    It was half an april 1st joke and half the whole heartedly wish of a long time flash user.
    I heard some positive feedback from some Macromedia fellas but also heard that some were mad about that thread because i “would have made people get too high expectations for next flash”.
    Come on people,that was years ago,give it to us finally ;)

  5. Well allright, I admit that OpenGL as a platform independent API has the power to do this.
    But the OpenGL-drivers that will be written to translate the OpenGL-Commands for a specific graphicscard may differ, and produce slightly different results on different Hardware.

    Today we get exactly the same results with flash on every system because flash lets the CPU create the screenery, and no external hardware that might influence the pixels in a way the user had not intended on his system… maybe I am wrong, but right now that’s what i think.

    Dont get me wrong, I would so much love to see graphics accelaration.
    But I get the impression that’s not what macromedia wants. Since flash gets more and more into the mobile device market I question that there will be a dependency of flash to use graphics hardware…
    Most palm and mobile devices do not have OpenGL features.

    We must look at the lowest common denominator of all systems that run flash…

  6. I´m not informed enough about in which way (technically speaking) adobe would make full use of the openGL api to be fully able to say that there won´t be different results on different hardware for sure; but having openGL as a standard which leads to good (and pretty much the same) display results (on same gen graphics cards) on different hardware for years makes me not worry about the feasability of trying to achieve good results on all hardware.
    Talking about what adobe intends to do in the future without having first hand coments on that particular topic is always speculation of course, but regarding the evolution of the flash plattform in the last few years i have the feeling macromedia learnt a lot from the ups and downs of the shockwave plugin and is going an own special route with the flash player.
    Shockwave director seemed to be aimed towards beeing the one for everything super milk cheese flesh cow thing. It should have 3d card acceleration when that wasn´t a standard feature of the most low end systems and one should be able to extend it even more with lots of extensions. Macromedia seemed to be ok with a way bigger filesize for the plugin in return (although fast internet connections still were way less common than they are now).
    The result was that although shockwave was and is used for many higher spec requiring apps/games,the download rates quickly were passed by the ones of the (technically) much weaker flash plugin and even today on my very new box i get stability issues when running shockwave content which makes use of certain extensions.
    With flash it seems like macromedia went for the least common denominator regarding hardware specs and download bandwidth requiring plugin which is proably one of the biggest if not the biggest reason for the success of the flash plugin.
    Because of that i was ok with them not integrating 3d capabilities while 3d cards weren´t that common in the standard household/workplace pc for a long while.
    But yeah,the situation changed a lot over the years,as far as i know ther´s no normal pc sold anymore (for a good while) which has no 3d capable card ,so the least common denominator plugin could be an open gl acceleration offering one.
    For me the expansion of the flash plattform with flash lite shows that macromedia was trying to go for the least common denominator on the mobile plattform,too,so offering the fitting least common denominator plugin for each completely different plattform now.
    I strongly hope meanwhile adobe continues that path with F9 and at least includes features which can be accessed by pretty much all standard machines on each plattform.

  7. Yes OpenGL rendering quality and appearance will vary from hardware to hardware. BUT the point is you can design for this. Game developers around the world know this. We are primarily talking about, aside a major overall performance boost, effects-type programming enhancement… things like additive layers upon layers and many many particles and so on which don’t necessarily always have to look exactly the dead same to be effective (for one).

    OpenGL support is way past due, for certain. Its really sad to watch an application choke on a few thousand particles or a handful of composite objects when you consider the stuff going on elsewhere.

    Hardware accelerating rendering could also be one of the only ways Flash would be really viable on ‘device’ platforms. Take PSP for example. Flash6 support was recently announced for the PSP browser. I’m sure it ‘works’ but is slow as all, virtually useless to deliver ‘Flashy’ things, physics, games, ‘rich interfaces’. But the PSP has a perfectly suitable OpenGL capability out of the box. And Nvidia/ATI are coming out with scaled down portable and mobile chipsets as we speak…

    So I think the key to Flash really staying ‘rich’ in the face of products like Micro$oft’s Expression framework, or emerging competition, is going to be adopting hardware…

    On another note, a game developer called Secret Level (www.secretlevel.com) DID write an OpenGL port of Flash Player (5) years ago, including ActionScript support! There was a public OpenGL Flash Player another company had put up, but it didn’t really support AS or Flash5. But Secret Level’s Flash Player was massive…

    They basically had it tessilate all the spline stuff in realtime, passing the resultant geometry to OpenGL and using some trickery to deal with precalculated ‘end caps’ and other special case display bits. They had also enhanced it with other functionality by where you could define a canvas for the stage within 3space and interact with other GL functions via ActionScript (so the story goes).

    They had built this thinking that most game developers would be interested in using Flash in games (for menus, guis, ad delivery, etc.) — with the sidenote being it would also be an easy content vehicle for people to get content onto gameplatforms (digital magazines, etc.). They had tons of interest from major developers and had set up a big promo thing at an old E3 show… I think this was around 1999 if memory serves me correct (6 years ago +).

    From what was relayed to me by friends working on the project, it could massively outperform normal flash — thousands of objects running at 120 fps in realtime, etc. Tons of bitmap compositing and being able to do things like mess with blend mode on MovieClips, etc.

    What follows next defies logic. Macromedia decided to ask them not to release it, stating something like that they do not have their ‘strategy for game platforms defined’ (or something like this was relayed to me).

    From what I got, Secret Level ended up shelving the project under Macromedia’s lack of official support for their development, which sounds like it had something to do with some bizarre politics. Macromedia did offer something like that they (Secret Level) could license their CPU-only (software only) port of Flash Player for PS2/XBOX — which they did and was used in a few games. But sadly that version underperformed because (of course) it was just a port of Flash to what were actually slower CPU’s than a typical PC of the time.

    For a while Secret Level had a white paper up describing exactly how they managed to do the Flash-to-OpenGL trickery (the conversion), but it was taken down probably under legal demand from MM at the time…

    Now, 6 years later, my shorts still get in a twist when I think about what could have been, while I spend 8 hours writing some code to create some basic blitter feedback like people did in WinAmp back in 1997…


Comments are closed.