Thanks, but I don't have a SB-910, or experience with it and its filters, so nothing there about this action. And so I'm just talking, without that experience, but my suspicion is the question might just be a problem in comparing results.
The D3100 manual says it does Nikons Flash Color Information Communication, which used to be a slightly different feature than the filter, but still reporting actual WB values into Auto WB, so I assume it is supported, and should work in Auto WB.
If we take a picture indoors with incandescent lighting alone, that picture is orange, so we must use Incandescent white balance to correct it.
I am assuming light bulbs of incandescent WB. Some bulbs are Daylight type, and today, many are CFL. So we may not actually know anything about the expected color, but Auto WB should try to sort it out. Better than nothing, but such automation is not always reliable however.
If we add flash to that picture, flash is white, and so it shows differently. So we put the CTO filter on the flash to make it be orange also, to match the incandescent, so the same Incandescent WB still works, and the lighting is not mixed then.
I would have liked to have heard a little about the procedure used for the comparison? The report was "not much difference". No difference in what? I think that means the picture is still fully Orange? And I would have liked to have seen some comparison pictures of it posted, showing before and after...
I wondered if Auto ISO was used. I think (not positive, but do believe) that the D3100 is of the Nikon era where Auto ISO goes high with the indoor ambient, and ISO remains high even though flash is being used (earlier and later models don't do it that way, ISO stays low with flash). So wondering about the actual ISO value, and if the orange ambient was well exposed (by high ISO), or far underexposed by low ISO? And therefore, how much the flash actually filled it.
Saying...
Incandescent ambient alone should be orange, and Auto WB should help it not be orange.
Flash alone (meaning with low ISO so with insignificant ambient level) should be orange with the filter, and Auto WB should help it not be orange.
But I suspect this was NOT the case, I think D3100 Auto ISO will be rather high indoors. And the orange filter idea is for high ISO, i.e., so that the orange ambient would show up to be orange, requiring Incandescent WB, and thus the filter on the flash to match it. (if the ISO were low, and the ambient very underexposed and dark, we would just use flash and no filter and Flash WB).
Both together, meaning with high ISO to let the ambient show up, should be orange, and high ISO keeps the flash from contributing much, and Auto WB should help it not be orange.
So not sure there should be much difference. We need some other plan to see a difference.
Was the report of "not different" that the picture was orange anyway?
I would suggest (as a test to learn more about the situation) repeating the test using Incandescent WB, which is the proper assumed value, including for the flash filter. Should work as well or better (if incandescent bulbs), but then if the light bulbs are not orange, the ambient (farther than flash range) should show up blue.
And try the flash without the filter, so it should stay white. These expected differences should show and emphasize what should be happening.
The point is, we'd know something else was going on.
Possible test variables... Auto ISO or low ISO. Auto WB vs Incandescent WB. Flash with and without filter.
Seeing these things can only enhance our experience and knowledge of the situation.
I'm not really the one to comment, because I have two incompatible notions. One, if I know it is incandescent, I see no reason not to use Incandescent WB. But two, I really don't even do that, I depend on correcting it better later in Raw, after I can see what is happening.
