Hey bakedbeing for us, the Lite problem for the buggy player was located in the Projection code of the camera (now found in the lens classes).
the toMatrix3D() method of the PerspectiveProjection class produced incorrect results, causing the entire view to scale to 0.05 of its correct size. tbh, the 10,0,12,36 was a beta release of the player most likely installed by developers who have since upgraded, i wouldn't expect a significant number of users out there to be knowingly running the old player. do you have any stats on the percentage of users to your banners who still use that version? Rob On Mon, Jan 25, 2010 at 8:10 AM, bakedbeing <[email protected]>wrote: > PS: I saw in the previous thread that there was a fix for Matrix3D, > but I couldn't figure out where in the A3DLite engine this patch would > be best applied, as there are so many uses of the Matrix3D class. Am I > using it on the viewport, or just for scaling/rotation on object 3d? > > My game only contains rotation and translation of planes and cubes and > a camera being positioned once. > > Ta! > > On Jan 25, 7:08 pm, bakedbeing <[email protected]> wrote: > > Bugger, just typed out a long reply-to-author and it disappeared into > > the ether, I have to get used to the Google groups setup :) > > > > Thanks for the reply Ken, that definitely looks like it. > > > > I wonder if there's any chance of seeing a fix for this - I know it's > > not a bug in the engine but rather the Flash Player, but when you're > > pushing out a banner (which a3dLite seems perfect for) publishers and > > clients really want maximum user eyeballs, and it'd be great if there > > was a workaround that let us apply the Matrix3D fix at runtime, so we > > could deliver our game to all the Flash 10 users out there. > > > > At the moment I'm having to divert all the Flash Player users with the > > buggy version of FP10 to a simple 2D frame rather than our mini-game, > > but with no Adobe stats for dot-releases of the player I have no way > > to predict (and tell our guys) how many users we're losing :( > > > > Once again thanks to everyone who's working on the engine, and if > > everyone is Katopz: thanks Katopz! > > > > On Jan 15, 2:21 am, Ken Railey <[email protected]> wrote: > > > > > > > > > Sounds like the flash player scaling bug to me. > > > > > Previous discussion on this list: > > > > >http://groups.google.com/group/away3d-dev/browse_thread/thread/83f3ba. > .. > > > > > HTH > > > -Ken > > > > > On Wed, Jan 13, 2010 at 8:13 PM, bakedbeing <[email protected] > >wrote: > > > > > > Heya guys, I think I may have uncovered the problem. > > > > > > From the release notes to Flash Player 10,0,22,87: > > > > (Fixes) > > > > * Matrix3D AS3 class implementation is difficult to use for general > > > > math. (2216936) > > > > > > Does anyone know what those changes were? Is there any chance that > the > > > > engine could be modified to support the old form of Matrix3D > > > > implementation for those players it detects to be < 10,0,22,87? > > > > > > I ask this because when we're pushing it through in banners we quote > > > > the statistics from Adobe's penetration studies, but these don't > > > > provide stats for dot releases. As a result it gets tricky convincing > > > > the client and publisher (and yourself!) that it's a wise move to > push > > > > out an Away3DLite banner. We essentially aren't sure who we're losing > > > > till we track the numbers from the first banners. I'm so bloody keen > > > > on getting them rolling tho, building mini games in banners is one of > > > > our most fun and effective practices! > > > > > > P.S. Performance is great! With two copies of the swf embedded in a > > > > heavy news page (including other flash content), and with that page > > > > open in two different browsers, all four banners are running at their > > > > full 21fps. > > > > Nice work guys! > > > > > > On Jan 14, 12:40 pm, bakedbeing <[email protected]> wrote: > > > > > Hiya guys, I've been doing some user testing on some Away3DLite > work > > > > > we're doing, and have found that serious problems arose on this one > > > > > version of Flash Player 10. > > > > > > > Functioning normally(10,0,32,18 10,0,42,34): A plane scrolling > towards > > > > > the user in the Z. The plane fills the views, as the camera is > > > > > slightly down-tilted. Basically looks like a mode 7 game, nice. > > > > > > > Broken (FP 10,0,12,36): The plane is much smaller and surrounded by > > > > > empty black space. It still scrolls toward you, but the entire > plane > > > > > is now tiny and flying at you. > > > > > > > It seems like maybe an issue with the projection in the final > > > > > rendering stage? Was something fundamentally borked in that > particular > > > > > Flash Player? > > > > > > > Thanks everyone! > > > > > Danny. > -- Rob Bateman Flash Development & Consultancy [email protected] www.infiniteturtles.co.uk www.away3d.com
