Anybody else having discrepancies between Viewport Preview and Render when using Shadow Catcher?
only in the Shadows… with AO on, the difference between Viewport and final Render is about half strength shadows for the Shadow Catcher shadows… what to do?
Blender 2.79 official
Tracked .jpg sequence as camera background
Basic World settings at strength 1.0
One Sun lamp set at .9 strength
Ambient Occlusion enabled at strength of .2
Well… i guess my only solution for now is to drop the AO setting down to half of what it is… not ideal… but it kinda works out.
Then add AO as a separate pass.
Here, the render is a bit darker.
Guess it has something to do with lighting used. Since HDRs for IBL really have no standard (the eye of the beholder).
Following test is done using slightly modified IBL/HDRI texture.
yeah that looks pretty good on yours… Im actually not using an hdri at all though… just a basic world setting with a medium grey color and a sun lamp… The backgrund image is just in the camera and doesn’t add to the scene… but for some reason, the shadows render about 50% lighter than they look in the viewport when using Ambient Occlusion… All my settings for preview and render are matched, so i wonder if its just a little bug in this kind of scene setup… Regardless, dropping the AO to half of what looks correct in the Viewport results in a pretty accurate Render result for me…
aaah… had missed the scene description / maybe you have Simplify checked 
it influences AO, might cause difference
also
Hi, I think you’ll find it’s just the long standing difference between the way the viewport handles alpha, and the way the compositor handles it. Try adding a square in your scene over a transparent background area, and add a material that’s half diffuse and half transparent. It comes out darker on the viewport with a background image than when it’s composited with an alphaover node. Pure emission is even more different, with it being shown over the background image in the viewport, but having no alpha in the compositor.
The viewport is just a quick approximation. It’s been reported as a bug a number of times, but apparently the behaviour is deeply burried in the viewport. Likely fixed in 2.8 though.
I’d suggest you render the shadow catcher in a different render layer, and composite this between the background image and the foreground image. It only need a small number of samples as you can generally blur it. That way you can quickly adjust the shadow intensity and colour balance to match other shadows in the image. This is very had to do in the viewport because any changes affect both the model and the shadow.
@Burnin, @Nerk… thanks for your replies friends… i have tested it all sorts of different ways now, and you are both correct… it does seem to mostly be an unavoidable discrepancy between viewport render and final render process at this time… but its not really too big a deal and is managable… thanks again for your replies…
