Need help from compositing gurus for lens effects

[sv]

can anyone through me a bone here - client wants some to add some lens effects onto some spot lights in the ceiling, and also for them to be reflected on the limestone floor. Using vray i can't seem to get the lens effects rendered, particuarly when sending the file through backburner to the farm, but i have found a work around: by setting up your render stats before hand and the lens effects, then clicking the "interactive" button under max's effects menu, max will start to render the image, using vray and the settings you've put in. Once the image has rendered, you can then save it from the frame buffer with the lens effects intact.

However, i want to farm out the image, as its an internal and it taking a fair amount of time to render. It would be good, therefore, to render out the lens effects as a separate pass, and composite them afterwards. Being a new-be to compositing, i can't work out how to render a channel which only holds the lens effects - or even if this can be done. (looking through the g-buffer, things like atmosphere are pretty self explanatory, but no post lens effects channel).

What would be the best way to go about doing this? What i've come up with so far, is to render the lens effects on there own with a black background, and use "lighten" in photoshop to composite them, but i am loosing quite a bit of detail, and obviously can't affect the layers below.

The other point - is it possible to reflect lens effects on the floor using vray, or is this an inherant problem with using a 3rd party renderer because essentially the lens effects are added by max once vray has finished its rendering process?

Thanks to anyone who can add some light, and for taking the time to read this post as know sometimes it can take a while!

[Gijs]

One thing, if you are adding a layer like this on top of your rendering in Photoshop, you are probably better of with Linear Dodge, which is a not so well chosen name for add. (I think this is present in Photoshop 7+) This means it just adds pixel values of the linear dodge layer to the underlying layer. Example: if your underlying pixel is RGB 125, 180, 240 and your linear dodge pixel is RGB 0, 70, 55 the resulting pixel is RGB 125, 250, 255.

[sv]

Is there a difference in functionality concerning blending names between composite programs? For instance, testing this last night on combustion, add seemed to do it fine, but this morning in photoshop, lighten seemed to be the better option?

[instinct]

Photoshop has kinda different Blending interpretations. There is a shake macro just to mimic the looks of the PS Blending modes.

Thorsten

[joconnell]

Yep - most people just render the lens flare against a black background and use add or screen mode to layer the flare over its background plate.