Here is my take on image based GI with shadow catching in P 11 and Superfly

I think this exact way hasnt been described yet and it seems the most straight forward and useful way to me

1. add a background sphere that is slightly larger than the GROUND object, put your HDRI on it, ambient channel only, switch visibility for raytracing and shadows off this sphere is for preview only

2. add an IBL light, put your HDRI on it, this light is for preview only and will not be rendered in Superfly later delete all other lights (unless you really need some)

3. put your image on the Background part of the GROUND Object, ambient channel only, increase ambient amount factor to taste ( I seem to use 3 most of the time), shader as follows:

TextureCoordinate Camera --> Vector transform Camera to World --> Environment Texture Vector --> Ambient Color

Set other channels to 0/black

4. put your image on the Ground part of the ground Object, as follows:

TextureCoordinate Camerea --> Vector transform Camerea to World --> Environment Texture Vector --> Diffuse Color

TextureCoordinate Camerea --> Vector transform Camerea to World --> Environment Texture Vector --> Ambient Color

Now use the 2D Wave shader to fade out the Ambient amount on the Ground towards the center of the object so it does not emit on that part, but is only illuminated passively for most parts (use upos: 1.5, vpos 0.5, frequency 1, phase offset -0.25 for instance for a large black center area)

**5.: **now switch GROUND prop on for rendering, off for preview.

this works for me better than all other solutions, and you can modify it easyly for uneven terrain or ojects in the background image which catch shadows via dummy objects

Do you have any renders available set up with this technique? It would be nice to see some comparisons between regular and your technique renders.

Hi

the result is not too different to other methods it's just easy to use

in the image you can see the horizontal line just above the white ball, since the mask didnt match well, its too large.

a trick to deal with the color shift is to sample the color of the sky and rotate it by 180° (use th opposite color) to tint the diffuse passive part.

Sounds and looks good. Reading through it, it seems all the setup could be automated as a PoserPython script?

I guess so, but I assume it's not worth it.

You can set this up once and save it all, then you just have to reload the HDRI 3 times.

But sometimes you want to make adjustments, so automatizing this is not much of an advantge.

For instance you may want to adjust the size and fading of the 2DWave shader mask, or you may want to switch the shadows and diffuse part off (you can do this by just using the same material on the ground as on the background of the GROUND object.)

Generally, you want the mask and "passive diffuse" part which catches the shadow to be as small as possible, to reduce noise in the rendering, but at the same time you want the transition between both materials to be as smooth as possible to reduce any ever so slight color and shading differences that night occur.

another thing worth mentioning that people often seem to miss is that your figure should not be at 0,0 but offset. 0,0 is actually where the image was taken from, so you actually want to place your camera around there.

At least thats what I think, I havnt played around with this too much.