realistic car compositing

[thundering1]

Actually, you can use the HDRI image IN Maya to render "all in one".

"Is it important to have the exact same background image that I used for the lighting?"

It's whatever you want it to be for the final image. Granted, I'd make a floor for it ot sit on for proper shadows, but the entire background would be the hdri image used.

No compositing needed, unless you're going for something else as far as the final image - like it's not actually in the warehouse, but in a garage - let's say you just wanted the properties of the lighting from the hdri image used.

And yes, I'd use PS for compositing (I'm a PS junkie) - you can render out an alpha of the car, that way you get a perfect cutout in PS and you can put it anywhere you want it, in front of any background you feel is appropriate.

Did this help in any way?

-Lew

[ryanstruk]

thanks, that makes sense. I have done a quick render, is that the general idea? I suppose once I work out how to cast a shadow on a render invisible plane it should "anchor" it to the ground. Strangely if I dolly the camera further out it doesn't appear to change the position of the background in the render despite apparently being changed in the view port. how does one get better resolution out of the background considering the hdri used was around a 50 MB file?, I expected it to look a little cleaner

http://i210.photobucket.com/albums/bb176/Mingzhus/xfire.jpg

[thundering1]

I suppose once I work out how to cast a shadow on a render invisible plane it should "anchor" it to the ground.

Yeah - that's what I was talking about - I'd make a ground plane of concrete and put it in place so the car is actually sitting on a piece of geometry which will react to the lighting.

And as far as the resolution of the HDRI map - they're usually stretched all over the place, having been taken from a silver ball which is why they don't look "crisp". In this render you're only seeing a tiny portion of it blown up - there's another 300-probably degrees in that image we don't see from this render.

The only way to beat this is an image map that is from something like the Spheron camera - it's freakin' enormous!! It's a 50MP camera that spins and takes images of 10 degree strips - bizarrely detailed!

As your image currently is, what you can do is output with an alpha map of the car itself, and make adjustments both to the background image, and the car to match them as far as lighting. This would allow you to also make your own shadows if you're comfortable doing that. You can go all over the place with this render in Photoshop - just how comfortable are you at messing with re-lighting a scene will determine how far you go with it.

Did this make sense? Hope this helps-
-Lew ;-)

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.