Sure, it makes no sense to make your pixels smaller than what you need to have the shortest wavelength of interest covered. Or smaller than what your tip can image.
That's long enough for me :)
I was indeed assuming all the pixels are independent, which is sort of true if the pixels are considerably larger than the tip radius but for smaller scan sizes that's not true. As you thoroughly show in your paper, one can indeed introduce more correlation by background subtraction and other processing algorithms. One has to be careful though. If I have an ideal scanning system and an ideal surface, such that the image is so flat that all the heights I measure are in exactly the same digitized value,...
I completely agree with David here. I've had this discussion many times with people who want to measure roughness. Anytime you try to describe a three dimensional surface by a single number, you are throwing away a lot of information. The roughness number contains about 6 orders of magnitude less information than the data it was derived from. When throwing away that much information, it is extremely important to know what you are doing and why you are doing it. There are some video courses on this...
For the work I usually do, that would be fantastic. Is the selection remembered or does it need to be unchecked every time?
I finally found out what was going on! I often run multiple iterations of subtract background-adjust color range, with the color range set to 'explicit'. If you don't select another item from the 'tools' part of the main window, the color range window remains active (even if you click 'hide' or close it with the cross) . By 'active' I don't mean it remains on the foreground but the color range is adjusted in the same way as when the window is in the foreground. And from what I can tell, single clicking...
'remove polynomial background' and 'remove scars' also affect color mapping, by the way.
'align rows' changes the color scale, and it has done so for a very long time. I don't remember if it was ever different.