irradiance map with animation

[Ryan_Chiong]

hi guys,

i cant seem to find a thread related to this.

this thing is really is confusing,

which irradiance mode am i goin to use for animation

1. multiframe incremental

2 add to current map

3 incremental add to current map.


regards
Ryan

[aphex79]

multiframe

[Ryan_Chiong]

furthermore... this is just camera animation. no moving objects at all .

[DanielBrew]

Ryan have you read this?
http://www.spot3d.com/vray/help/150R...ials_imap2.htm

Dan

[Ryan_Chiong]

Hi Dan

i have read the tutorials but what bothers me is that what advantage do we get using multi frame when it disregard the computed solution of the previous frame on the otherhand, incremental add refines the solution of the previous one, thus lower render times.

[DanielBrew]

Hi Ryan,

I think you've misunderstood what multiframe incremental does.

Yes, it clears the memory of previously calculated IR and LC data but then it calculates the light solution for the first frame of your animation and then it adds more samples as it renders subsequent frames. It doesn't clear the memory before rendering each frame, only before the first frame.

Dan

[Clifton Santiago]

Exactly what DanielBrew said...

If Multi-frame incremental cleared the memory of every previous Nth frame of IR data it would be called Single frame

Multi-frame incremental lowers render times because it doesn't recalculate the IR solution for all the geometry.. only for new geometry it sees from the last Nth frame.

Additionally, you need to calculate it in advance (on a single machine) in order to prevent IR map flickering in your animation.

[Ryan_Chiong]

those were really comprehensive

thanks for taking time to explain. it wqas greatly appreciated!

Ryan

[DanielBrew]

No problem Ryan, hope that's a bit clearer now.

[Ryan_Chiong]

yes its very much clearer now,

One more thing, is there a way to calculate IR in reverse sequence? last frame to the first.

[Clifton Santiago]

Well, for calcualting it forward, you just click Active time segment, and specify every 10th frame (depending on camera speed), so to run it in reverse, you could specify individual frames in reverse order, i.e. 100,90,80... etc. Though I don't know why calculating it in reverse would make a difference.

[Ryan_Chiong]

well im trying to test using "Incremental add to current map" on a walkthrough. just wanna test and see what happens. my idea is calculating the IR of the farthest object from the last frame and have a good IR map of it, eventually building up the IR to the first frame where i do not have to recompute IR every nth frame.

Ryan

[Clifton Santiago]

Oh, you're not using your camera path to calculate your IR. Well in that case, you would use incremental add. But chances are using the IR map Nth frame option, would be faster because you will only be calculating the things the camera sees. However, if you are using lightcache, which is time dependent, you will need to calculate using the camera path with the same time frame, like Vlado describes in his tutorial.

[Ryan_Chiong]

yes i am using camera path. though the path is linear and doesnt have much turning.

[Ryan_Chiong]

can i compute say.. frame 0-100 on 1 node and 101-200 on another, then merge the IR map?