light cache in fly trough

[serge]

am i wrong or light cache is saved with irmap?

i make some test with light cache for secondary bounce

settings:
precalculation of irmap:

-irmap+light cache
-irmap incremental add to current map

-light cache mode single frame
-on render end don't delete

rendering:

-irmap+none
-irmap from file

result is the same if i use
-irmap+light cache
-irmap from file
-light cache from file

[tom schuelke]

Hi

lightcach is saved in an own file.

usualy you render lightcach as a flightthrough and save it .

this can be taken from disk in the next step (also in one step) and used for i-map precalculation, incremental. this works fine for architectural flight throughs. When you move many things of course you have to choose singleframe calculation or rendering with layers.

if i am not wrong, after i-map calculation you dont need lightcache no more, because you can take i-map instead.

Tom

[flipside]

Yes you are correct, you can turn off light map for the final renderings.

[serge]

if i understand light map calculation is used for irmap and store with it

what is the best way for irmap calculation over network

actually if i have 1000 frames i send frames 0-200 to node 1 201-400 to node 2 and so on...

i use imap viewer for merging all the irmap and send the final rendering
over network

with light cache how can i networking the irmap calculation and use imap viewer for merging all the irmap

do i use single frame or fligh trough

[Da_elf]

light cache has nothing to do with imaps. they are two totally seperate GI solutions and dont save in an imap file

[flipside]

Yes they do!

Lightmap info is used during imap calculations. If you calculate LM and IRmap, then save the IR map to disk, you can rerender the scene with this loaded IR map while lightmap for second bounce is turned off!!

[natty]

Yeah this is true ...

[serge]

strange ??

i've just make a test
precalculate ir map with settings i said before
render the whole sequence with saved irmap and without sec bouce
all was fine that's why i said light cache is used for irmap calculation and store with it

what i don't understand is the diff between single frame and fly trough in the light cache mode

is there some one who can explain in detail the right way to use light cache in fly through with network rendering

[serge]

Yes they do!

Lightmap info is used during imap calculations. If you calculate LM and IRmap, then save the IR map to disk, you can rerender the scene with this loaded IR map while lightmap for second bounce is turned off!!

lightmap is dead
long life to light cache
this the new way of our rendering

[flipside]

strange ??

i've just make a test
precalculate ir map with settings i said before
render the whole sequence with saved irmap and without sec bouce
all was fine that's why i said light cache is used for irmap calculation and store with it

what i don't understand is the diff between single frame and fly trough in the light cache mode

is there some one who can explain in detail the right way to use light cache in fly through with network rendering

Serge, we agree that lightcache is stored inside the IR map, don't get confused

Fly through is for fly throughs It calculates 1 lightcache for you entire camera path. Please refer to the help file for some more info, it's explained in there.

Don't have experience with network rendering but I think it's best you calculate the lightcache first with flytrough mode turned on. Save it. Then load it and calculate IR map for each n-frames, IR map set to single frame (because you render on different PC's!). The lightcache should be computed on 1 PC. Then merge all your IR maps with the IRmap merger or whatever that tool is called. This final merged IR map can be used to render your final frames, with second bounce set to nothing so it doesn't compute lightmap over and over again.

wouter

[Da_elf]

wow. ya learn something new every day. so basically its a waste of ram to calculate imap and lm together and save them both then use them both in the load function? and if it saves with irradiance map in the primary bounce then wouldnt re-loading a saved lm in the secondary bounce make things too bright since its already saved in the primary?

[flipside]

I was thinking that too, not sure how vray handles this, but I thought the result is the same as without second bounce and loaded ir map

[baumgore]

Saved IR plus saved LightCache is indeed a waste of memory, and the scene is not getting brighter with both on. The making of both is also a waste, but of working time, the flythrough for the Light Cache is done in a single frame, the IR maps must be collected over several frames. A simple optimized "Make Light" algorithm would make work a lot faster, distributed, multiframe, incremental IR calculations WOULD be a starting point.

[RErender]

I don't know what all the fuss is, works great to render every nth with single frame lightcache and single frame imap. Saving only the imaps, which includes the info from the lightcache. Then just merge them together...

[Da_elf]

it doesnt work with Incremental add to current IM and single frame LM?