First of all, thanks for the replies!

@hborre: At this stage I was not trying to achieve anything other then matching the initial look that came with the set (without GC). In particular I was missing the vibrance and contrast. Also I know that it is good advice in general, and I've read it before, but I also always find it very confusing when someone has a problem with the image being too dark, and people suggest that they should lower the light intensities...

 

@BB: Thanks! I've saved that to a text file :) I am aware of the whole "linear workflow". I've read quite a few of your posts here and on RDNA on the subject and I'm all on board :) I also remember that the subject of color swatches was discussed somewhere, but somehow I thought that when you turn on GC poser will do the "anti-GC" phase both on the images and the color swatches, but perhaps I remember it wrong...

Also a related question: You are suppoed to set the gamma of "value" type images to 1. But sometimes the same image is plugged to a bunch of color type of channels as well as value type channels and you can only set one gamma per image...

What's the best way to solve this? I guess, rather then creating and loading a separate copy of the image just so that you can set the gamma independently, one could just set the image gamma to 2.2 and add an adjustement node before connecting it to the value channels?

Or (and this just occured to me) if the image was created with the diffuse channel in mind then it may require gamma correction for the value inputs as well?

Which actually leads me to a question that I always wanted to ask: I understand in theory that you do not want to (anti) GC value maps like bump, transparency etc. However what I always wondered is that if the image is/was created in Photoshop or similar, then Photoshop would not know that the image was meant to represent values and would (I assume) perform the usual gamma encoding when saving the image, wouldn't it? And if yes, wouldn't we need to compensate for this to get back the intended gray levels?

Back to the image at hand:

I've adjusted the light levels as instructed and it helped a lot:

Note however that the image is still not as "crisp" as the original. The texture clarity (?) and contrast on the sofa is still not as good as the original. The column at the stairs and the table is overexposed/blown out. The lights now are probably too much. (Note: the lighs intensity varies between 35%-60%-ish...) The reflections on the lapshade look weird and for some reason the left side of the bug vase below the stairs lost its green color... I wonder if this last thing is related to me unchecking the lite multiplication or to the fact that the marble texture is added on top a greenish base diffuse color (0 68 68)...

I originally did not touch the supplied lights as originally I did not think there was anything fundamentally worng with them (Apart from turning off AO, which should be unnecessary with IDL, right?) However I also did a test render where I've thrown out all the lights and added just a single infinite light at 0.9 intensity. Here's the result:

This is now much better. Though some things could be improved. Especially with the small objects below the stairs.

Some specific questions:

• One of the original lights is a diffuse IBL with just a color. In fact it had a "simple color" node attached with a gray color to the "color" input. The light also has a diffuse_color channel as well which was also a gray color. What's the difference between the two channels? In the first image with the adjusted original lights I've removed the simple color node from the color channel, set it's color value to white and adjusted the intensity based on the formula from BB. I did not know what to do with the diffuse color channel gray value, so I've just set it to white as well...

• A lot of the surfaces have a non zero ambient value, often combined with a non-white ambient color... How does one deal with that? (I've seen a lighting training video suggesting that adding an ambient component to all sufraces is an acceptable way of providing general ambient lighting, but I've always found it a bit strange...)

• Some of the maps are added onto non-white (and often non-gray) base diffuse, specular and other channels. How does one adjust thise (if at all)?

• Often the same image map is linked to amost all the base input channels... This seems to make it rather difficult to deal with them properly...

 

Again, many thanks for your help in advance!!!