Thank you in advance and I don't know if that is possible since Maxwell and Fry seems that can't do it and i don't know if its impossible or not implemented yet.
(This also could be to speed up bake textures
)Could indigo reuse the calcs for a flytrhough render?
-
VictorJapi
- Posts: 58
- Joined: Sun Jan 21, 2007 1:27 am
- Location: Zamora - Algeciras || España
Could indigo reuse the calcs for a flytrhough render?
Since the MTL based renderers calc the correct light equations, and later draw the pixels in screen based in that equations... my answer is: could that equations be stored to speed up other renderer where only the camera has moved?
Thank you in advance and I don't know if that is possible since Maxwell and Fry seems that can't do it and i don't know if its impossible or not implemented yet.
(This also could be to speed up bake textures )
Thank you in advance and I don't know if that is possible since Maxwell and Fry seems that can't do it and i don't know if its impossible or not implemented yet.
(This also could be to speed up bake textures
)winmain(){
japi victorJapi;
victorJapi::Signature();
}
japi victorJapi;
victorJapi::Signature();
}
-
VictorJapi
- Posts: 58
- Joined: Sun Jan 21, 2007 1:27 am
- Location: Zamora - Algeciras || España
I have a big hard diskzsouthboy wrote:Yes it can be done, no, indigo can't do it.
Basically the renderer requires enough ram to hold the light data for every single face in the scene.
Additionally, techniques like bi-directional path tracing would not work with this, as there is no "eye" to trace back to, from a lightsource.
winmain(){
japi victorJapi;
victorJapi::Signature();
}
japi victorJapi;
victorJapi::Signature();
}
Again, bi-dir MLT is dependent on a camera.
Right now, indigo, as an example, simple bounces rays around, to the camera viewpoint. The data that MLT creates along the way are not "cached" at each triangle/model (or "baked"). The results from the data are only stored at the viewport buffer level.
Think of it as indigo taking a polaroid of a scene that was described to it. The scene doesn't exist, but the polaroid does.
And RE: your hard drive space, I know you were joking, but it's not feasible.
Right now, indigo, as an example, simple bounces rays around, to the camera viewpoint. The data that MLT creates along the way are not "cached" at each triangle/model (or "baked"). The results from the data are only stored at the viewport buffer level.
Think of it as indigo taking a polaroid of a scene that was described to it. The scene doesn't exist, but the polaroid does.
And RE: your hard drive space, I know you were joking, but it's not feasible.
i have completed a couple of short animations, as a demo. on is here:ZomB wrote:Perhaps its possible to cache some of the calculated data,
that could be reused for animated content,
so it would still speed things up, but cost some temp (hdd) storage...
btw:
Does anybody had ever shown some animated Indigo work???
http://ftlfm.com/doug/video/bounce.avi (430k, xvid codec i think.)
Who is online
Users browsing this forum: No registered users and 6 guests