Back again. I wanted to call your attention to some screen shots of the 2 Tridents:
http://devluchadore.wordpress.com/2011/04/04/issues-converting-from-away3d-3-5-to-3-6/ They show how things are getting twisted in my view by showing the 3.5 vs the 3.6 Tridents. Sure would like some help understanding how to fix this. In fact, I'm pretty much stuck not moving forward with 3.6 unless we can figure it out. In the meanwhile, some brief notes in the article (the link is a blog post, pardon my Luchadore persona) may help others who want to be able to switch back and forth between 3.5 and 3.6 easily. There wasn't much to it for my app. Thanks in advance for the assistance with these issues. Cheers, BW On Apr 3, 12:06 pm, Michael Iv <[email protected]> wrote: > As being one that had hard times with getIntersect would like to add ,you > have to transform the faces of your object into the view space before you > perform the test. :) > > > > > > On Sun, Apr 3, 2011 at 9:55 PM, Fabrice3D <[email protected]> wrote: > > Just uploaded a 3.6 and 4.0 update for Drag3D class > > > the getIntersect method works as usual, returning mouse coordinates/hit > > plane by default > > but if you pass a custom, x and y getIntersect(myX, myY); will return the > > hit based on these. > > > cheers > > > Fabrice > > > On Apr 3, 2011, at 8:42 PM, Bob W wrote: > > > Another problem with 3.6 versus 3.5--the "upAxis" parameter for the > > camera's lookAt is acting funny. > > > I am just passing it unit vectors: {1, 0, 0}, {0, 1, 0}, or {0, 0, -1} (Z > > is inverted in my model's space). This worked great in 3.5. In 3.6 I get > > all sorts of odd effects: > > > - My "perspective" view, which is the camera hovering above and somewhat > > left of the object has a lookAt of {0, 0, -1}. You would expect Z to be up, > > and it is, but mysteriously, the X and Y are rotated in the process. The > > old lookAt didn't do this--the other axes stayed put. > > > - I also have "Top", "Right", and "Front" views. Similar bizarre things > > happen to them. > > > Must be something quite different about the matrix math used in 3.6 vs > > 3.5. > > > How do I fix it? > > > Best, > > > BW > > > On Sat, Apr 2, 2011 at 5:02 PM, BobWarfield <[email protected]> wrote: > > >> I'm trying to switch my fairly large app from Away3D 3.5 to 3.6 and > >> I’m seeing some problems: > > >> Scene3D > > >> No more updateLight property. I don’t use lights. Is there an > >> equivalent I should use to make things run a little faster? > > >> There is an autoUpdate, but no update. If autoUpdate == false, what > >> am I supposed to call to update? Seems like you should either have > >> both autoUpdate and update or neither. > > >> Drag3D > > >> Fabrice, we used to be able to pass in coordinates to getIntersect. I > >> found this really handy and use it for all sorts of things. Now > >> apparently it only accepts the mouse coordinates and no args? > > >> Comparing the two versions (3.5 and 3.6), the tiny bit of performance > >> gained sure doesn’t seem worth the loss of flexibility. It can’t have > >> been a very performance intensive op anyway. I bet you could keep the > >> flexibility and get back most of the performance diff using Jackson > >> Dunstan’s improved “isNaN” code. > > >> Matrix3D > > >> Is there code somewhere to convert a vertex or Vector3D to screen > >> coordinates? > >> I had been using some code ripped off from ZoomFocusLens.project: > > >> v.x = (v.x * view.camera.viewMatrix.sxx + v.y * > >> view.camera.viewMatrix.sxy + v.z * view.camera.viewMatrix.sxz + > >> view.camera.viewMatrix.tx); > > >> etc. > > >> Of course it is no longer viable since Matrix3D came along and the > >> code in 3.6’s ZoomFocusLens.project seems a lot more daunting to > >> repurpose. > > >> You will notice the similarity of purpose to what I use > >> Drag3D.getIntersect for. Seems like utils to help people go from > >> Away3D coordinates to ordinary Flash graphics coordinates would be > >> very popular. > > >> Best, > > >> BW > > -- > Michael Ivanov ,Programmer > Neurotech Solutions Ltd. > Flex|Air |3D|Unity|www.neurotechresearch.comhttp://blog.alladvanced.net > Tel:054-4962254 > [email protected] > [email protected] Hide quoted text - > > - Show quoted text -
