« AUG Finland meeting tomorrow: Flex | Main | One month 'till CS3 »

Category: Flash

February 22, 2007 - artman - Flash

My 2 cents on 3D in Flash

3d

There's been a lot of buzz around 3D in Flash since Carlos Ulloa decided to decided to go Open Source with Papervision 3D (as a side-note, if you plan on going Open Source, why wait to release the source code and enable anyone to contribute?). And as promising as it looks and as much as the guys earn respect for creating something as complex as PV3D I must say that everybody is a little too excited.

I'm not saying that PV3D wouldn't find it's way into many projects down the line. It'll definitely find it's place in a number of cool (yet simple) UI's.

Aral Balkan even went as far as to say that we finally have PlayStation quality 3D in Flash. Now quality in my mind means speed, when you talk about 3D. Although the original PlayStation was not equipped with a GPU for 3D drawing to what we're used today, it did have a Gemoetry Transformation Engine, which increased the speed of 3D calculations required to draw flat, shaded or textured polygons and was able to draw 180.000 texture-mapped and light-shaded polygons per second. Thats still 6000 polygons if you want everything to run 30fps. PV3D is not going to be able to come even close to that (with the average PC that's out there).

Now jump back quite a few years to almost the beginning of the century and look at one of the most popular graphics cards: The GeForce 4 MX series. Even in its lowest configuration its able to paint 31 million triangles per second. And all with anisotropic filtering, pixel shading, multitexture rendering and lot of a heck more that wouldn't be possible (and is never going to be possible) to achieve with CPU rendering.

What's my point? Well, my point is that we are Flash developers (at least the ones still reading...). We're amazed by what guys like Carlos are doing, since we know what we're up against. We know we don't have hardware acceleration. We know that AS3 is still a scripting language. That's why we tend to go whoaaa whenever a new 3D engine or C64 emulator sees the daylight. But our end-users, the ones that use the end-result of our projects and in the end truly matter, don't know what constraints we have. They use hardware accelerated 3D all the time. And they will pitifully laugh at any 3D game created in Flash we can throw at them. The engine that turns us on will look embarrassing in the eyes of someone who has played a few 3D games.

What we need is a hardware accelerated Flash Palyer, because with the introduction of AS3 the bottleneck in the player is again graphics rendering. Not only will hw-acceleration give us awesome 3D, but accelerating vector & bitmap rendering should also give the Player a performance increase that'll blow your mind (You know, bitmap rendering with a decent gfx-card should increase performance some 10 000%).

I'm sure the the Flash Player team is looking closely what Microsoft is doing on that front, and am quite certain that we'll see a fully accelerated Flash Player 10.


Comments

I agree with what you're saying on the game aspect. But for simple demos in 3d and interfaces papervision opens up a world of hurt. You can now potentially create full 3d interfaces that look halfway decent on the web with VERY minimal download time. This is something we've wanted in flash for a long time. Even if it isn't hardware accelerated. It opens up an entire avenue of interfaces. That's what is great about it.

Posted by: Austin | 22 Feb 2007 20:52:08


Hardware-accelerated 3D has been on the wishlist for every version of the Flash Player that I've been involved in. It's my view that if we can get enough people involved in creating 3D for Flash, it might pave the way for this feature request to become a priority. In light of this, I see PV3D as a wonderful advance that will result in a lot more 3D Flash content being created and, hopefully, demonstrate to Adobe that it should invest in building hardware-accelerated 3D into the Flash Player. You're right, that's the ultimate goal; PV3D in my mind is both a means to an end and will, in its own right, result in some very cool 3D Flash work.

Posted by: Aral Balkan | 22 Feb 2007 21:03:44


A good reference post here is "The limits of software rendering" from Tinic Uro, Flash Player Engineer: http://www.kaourantin.net/2007/02/limits-of-software-rendering.html

Posted by: darron | 22 Feb 2007 21:04:54


Papervision3d is getting us started with 3D in Flash. Whenever Adobe decides to implement hardware acceleration in the Flash Player, a framework will already exist.

The way I see it, Flash developers have good reasons to be excited about Papervision3d: these early works are making the case for a long neglected need.

Posted by: Oscar Trelles | 22 Feb 2007 21:43:40


I built my career making seemingly useless bits of things that made people go ooh and ahh. :) It's this kind of playing and experimenting that drives the technology forward.

Posted by: Keith Peters | 22 Feb 2007 22:08:23


Oscar, if Adobe implements hardware acceleration, they are going to implement a 3D framework of their own, rendering PV3D completelty useless. That's why deeply respect Carlos and the rest of the group working on PV3D. They know that their years of work can be rendered useless with one single announcenement from Adobe.

But again, I completelty agree that until them, PV3D rocks and is going to help a lot of people do really cool stuff.

Posted by: Tuomas Artman | 23 Feb 2007 08:10:49


One might wonder why a Flash Player engineer is looking into the issue of how to speed up PV3D's software rendering first, and then question when the hardware accelerated bomb will drop.

Posted by: UnitZeroOne | 23 Feb 2007 10:20:37


Papervision3D might not be the 3D engine of the future, but it aims to be the seed of a Flash 3D community that has an awful lot to learn, and even more fun to have in the process.

This is not about polys per second, it's about the Flash people discovering 3D.

Posted by: C4RL05 | 2 Mar 2007 12:52:14


jeezy standing young

Posted by: underground jeezy young | 2 Mar 2008 13:59:40


jeezy standing young

Posted by: underground jeezy young | 2 Mar 2008 13:59:51


I think the best and risky strategy is for carlos to make a company and then Adobe buys them out heh. Just a thought tho :D peace y'all. :)

Posted by: Wilmon | 27 Mar 2008 11:22:16


 

Post a comment