@Digitalis
Quote: Also if it was a spot on the filter or sensor it would be dark, not bright.
That's a simplification which is not helpful.
Of course the physical pixel stays "dark" if it is not met by a light beam (or that cell/amplifier does not work anymore). But the user will never see that directly. Even in the raw file the values of each pixel are added ones of what the four physical pixels surrounding it got.
So, what happens with that pixel if the light path to one of the surrounding physical ones is blocked, depends.
If that spot in the picture was dark anyway, nothing changes.
If it would have got some light, the colour balance of the resulting values will change. In most (but not all) cases, this pixel will not show the correct colour, and it will be darker. But there are many combinations possible. For example, if the physical pixel was covered by a blue filter, and the light blocked did not contain much blue, changes will be very small.
What I wanted to say in my last post (and you rejected it) is:
If the light path to one physial pixel is blocked, or the cell is not working anymore ("cold pixel"), you will see in most of you pictures at that position a point with a wrong colour, which also may be slightly darker than it should.
As far as I know, the test facilities of the manufacturers do map these faults, but I have no information whether the built-in pixel mapping of the bodies can do it as well.
I think I stumbled on this when I remembered the discussions in a German forum about the faulty ROM routines of the K20D for making hot/cold pixels unvisible. It probably was cured with an update, I never owned that model. The result was that, if such a pixel was met by a contrasty line of the scenery, coloured points could emerge (even if the scenery shot was just b&w in this area). I think this was dependend on whether the line was horizontal or vertical. Pixel mapping did not help, as in some respects the fault was a result of pixel mapping.