Any time I’ve worked with an image background I approach it in a similar way as I did when I was doing image manipulation in Photoshop. Mainly, keeping the light/shadow angles, the light color, and the “quality” of the light as closely matched as possible across my source images. If there are no glaring problems with the colors and the shadow directions there is actually a good amount of wiggle room.
It’s even easier in Blender since one is able to rotate an HDRI all over the place in order to get the best angle. It’s also possible to drop adjustment nodes (HSV, RGB Curve, Brightness/Constrast, etc) after the HDRI in the world shader to really dial things in.
So yep, it’s pretty much about picking something close and tinkering until it looks believable. And/or color grade the heck out of it later. 😀
