|
From: Frank C. <dev...@ch...> - 2004-09-28 18:18:12
|
On 28-Sep-04, at 1:52 PM, James W. Walker wrote: > Peter Michelsen <Pet...@mi...> wrote: > >> I tried this but it had no effect as the objects are >> transparent..... and I guess ( from testing ) use the last style >> illumination style set in the submit loop. ( does this sound right >> or am I way of base? ). > > That seems right. Quesa saves the illumination style with each > transparent primitive, but only uses it to decide whether to use > specular highlights. That is, it distinguishes Phong and Lambert > illumination from each other, but does not consider NULL illumination. > This should probably be filed as a bug, if it hasn't been already. I pointed this bug report out earlier in the thread: [ 967773 ] Lighting state isn't respected on transparent objects <http://sourceforge.net/tracker/index.php? func=detail&aid=967773&group_id=45158&atid=442052> Simple fix really, but some other states can be merged into a state cache as well, like fog, orientation, and fill - setting those for each polygon is a waste. I can put together a fix but was kinda waiting to see an update to IRTransparent.c first, since the sorting code has proven to been unusable for my purposes. James: Has there been any progress with your new sorting algorithms? I would very much like to see your fast centroid method checked in. There are also some related fixes that would help bring the transparency renderer up to snuff: <http://sourceforge.net/tracker/index.php? func=detail&aid=973022&group_id=45158&atid=442052> <http://sourceforge.net/tracker/index.php? func=detail&aid=967958&group_id=45158&atid=442052> Frank. |