Hmm.. to be honest this information is a bit of a surprise to me, simply because I've been able to get good lighting on all parts of the sphere. Using an inverseSceneTransform to move the light around, I found a few cases where it would move very quickly, and lap the sphere as it rotates -- lighting the relevant portion of the sphere correctly. Maybe I'm misinterpreting how the normal map is used in this case; I'm not sure what workarounds would be effective in this situation, but it definitely has to do with the normal map itself?
On Oct 1, 5:20 pm, Peter Kapelyan <[email protected]> wrote: > When you create a normal map, you are basically baking in the light and > shadows from a certain direction. So when you light up the dark side of a > normal map, it will still be dark. > > -Pete > > > > > > On Fri, Oct 1, 2010 at 4:52 PM, mikerz <[email protected]> wrote: > > Hi All, > > > I have a AWD model loaded (a specially UV-mapped sphere), which I > > apply a Dot3F10BitmapMaterial to. Looks great, the lighting works > > correctly. When I rotate the model, the light moves with it. This > > doesn't make sense, given that it's a directionalLight3D which is > > added to the scene and not the model. > > > Still, I tried a few workarounds in trying to map the light to a new > > position based on the rotating object's transform matrix ... also > > tried using sceneTransform, inverseSceneTransform and a wide range of > > similar techniques. I tried adding an empty Object3D into the rotating > > object at the light's desired, normalized direction as an automatic > > way of calculating the new direction. > > > Nothing has worked in terms of workarounds so far; is there something > > going on "working as intended" which I'm not realizing? > > > -Michael > > > Any help appreciated, thank you > > -- > ___________________ > > Actionscript 3.0 Flash 3D Graphics Engine > > HTTP://AWAY3D.COM
