I made a sort of discovery after a few hours of frustration today. I don't have pics to show this currently, because I kept going until I resolved the issue. (And at the moment, I don't feel like waiting on more render time.) I do know it affects Carrara 6.2 and older, so YMMV if you have the newer versions. Hopefully DAZ recognized this issue and has either fixed it or put it on a bug tracker.

Basically I had a region of a model that was UV mapped in other software. It was just done only on a part where silkscreened decals would be on the real thing. So I expected that a simple layers list or multi-channel shader to work. It would seem like a simple job of putting in a reference shader in one slot and dropping in a wizard preset for the decal paint in the other and then putting my texture image in to drive the channel separation or layer opacity. But it didn't work. No matter what I did trying either of those, the area of the model with the UV material zone always appeared a shade darker than the materal zone without UVs. (And I tested other aspects. Cloning the material in a new shader would work seamlessly, however a reference based material always ended up darker.)

Now I think this applies to a specific case. This is because I was using a shader with multiple lighting models in it as the base material. (Has a phong for a nice metal-flake paint.) So somewhere along the way, Carrara wasn't properly digesting that particular bit of info from the reference - and it kept changing some aspect so it didn't match. When lighting models aren't used in channels, this didn't seem a problem. Labels elsewhere on the model where the background didn't have that quality could be done the "easy" way.

The fix?
Instead of applying the decal in what would seem the easy way, I had to duplicate the base shader containing a multichannel mixer with the different lighting channels. Then in the respective channels of those - I had to modify the colors etc. with the mixer driven by the texturemap. It was a bit more work than I expected. At least there's some degree of copypasta funcionality with shader channels between different shaders and mixers would retain an old value in one channel when set, so it wasn't too difficult to get a match. But definitely not a direct approach.

I may do an example later, unless somebody else cares to try. At least I hope this little bit of info will prove useful.

 good to know