People who work professionally in Architectural visualisation generally make post-process of the images (once they are been rendered with their 3D modeller program) using Photoshop. I mean, the only render never is enough for a realistic or final high quality image. So, if you want to improve the final render with photoshop, to get a more realistic result, just go on. It is a standard practice in arch-vis.

The main goal was to find out how accurate and realistic you can get with SH3D. This rules out any post processing of the rendered images.
I find the result quite satisfying. I also like the fact that I can reproduce the exact same result anytime. I can also change perspective, zoom in and out and render new images at the same quality quite easily.
There's a lot more work to do if you add in post processing in a third party program.

Best, slimpy