This is my latest effort with a Kemp slide:
![Image](https://i.imgur.com/i6hWwyg.jpg)
Seems more common to measure the apparent distance it takes for the intensity to go from 10% to 90% relative to the values far from the edge, which is also an effect of diffraction as crb5 mentioned earlier. Measuring the distance from the center of the transition to that dark band is conceptually similar but more sensitive to the exact shape of the response, so not as robust and more difficult to interpret without prior knowledge of what the edge response should look like for the specific optical system in question. Figure 25-4 in that dspguide.com link is illustrating how 10% to 90% transition width is not strongly dependent on the exact shape of the PSF/MTF.Stonius wrote: ↑Wed Jul 21, 2021 8:58 amThe insert at 100% shows a line separated from the edge of the black division marker by about 50 pixels, or 0.0005mm. I take it this line is an effect of diffraction and represents the point beyond which no further detail can be resolved, even if all else were perfect?
If you will infer resolution from this dark band diffraction artifact then why not from the apparent transition width, which is also a diffraction artifact? (Assuming the physical transition on the slide is sharper than the objective can resolve.) Transition width 10% to 90% (as opposed to other details/features of the edge response) appears to be widely used and has some practical benefits discussed on the two pages I linked earlier.Hobbyst46 wrote: ↑Wed Jul 21, 2021 10:03 amThe thin line around the scale mark, which is probably an artifact (besides diffraction, there is chromatic aberrations), is nevertheless a distinguishable object in the image, So I treat it as if it were a real object.
Its distance from the scale mark visible border on my monitor is ~2mm. Hence, its true distance in the image created by the microscope is 5.6*2/32 = 0.35 um.
Yeah as the air gap approaches zero, it is simpler, rays corresponding to NA >= 1 are just lost due internal reflection going glass-to-air but there are no other effects on imaging. With significant air gap it gets more confusing to think about because of spherical aberration. In the extreme case with only air like Markus's first photo (guessing no cover glass either in that one) I believe the paraxial focus is shifted closer to the front element by a factor ~1/1.5, similar to the focus shift caused by plane parallel plate elements. But marginal focus approaches the surface of the front element as NA approaches 1 indicating extreme spherical aberration. So I guess the real question is, at what air gap thickness does resolution become limited by spherical aberration, so that NA is no longer a useful metric?
hans wrote:...
Thank you both for the link and comments. Admittedly my previous claims are only relevant within the two popular (at least in microscopy texts) definitions of resolution. I agree that the width of the 90% to 10% decay is a reliable indicator of resolution for the image obtained under oil immersion. It is far less reliable, IMHO, for the image of the scale bars under 100x objective, dry.crb5 wrote:...
There is a great deal of skill in light microscopy, and as such it can be considered a craft as well as a science.Hobbyst46 wrote: ↑Wed Jul 21, 2021 3:35 pmHowever resolution is not everything. Contrast is as important. Illumination is important. I would assume that there is nothing wrong with the resolution of your microscope and start looking at interesting specimens instead. After some specimen-related adventures, one forgets about theoretical vs practical resolutions.