alanr0

It was wrong in the context you originally presented it. Wrong in several ways. This was the first, and fundamentally wrong assumption in your argument. Half a degree is not small and point-like. W ...

Not a good assumption in the present case. Diameter of the solar disk is 0.533 degrees or 0.0093 radians. Image size with a 400 mm lens is 3.7 mm. Even reducing the focal length to 40 mm, a 0.37 mm ...

Correct. Energy per unit area is proportional to square of numerical aperture, or inversely proportional to T-stop squared. F-number dominates, but there is a second order effect associated with ...

Very true.  You have made that point repeatedly, and I agree with it.  However, that is not what I suggested. I presented a simplistic exposition to help GB, and anyone else who might be ...

<snip> a is just a scale factor which determines how large the aberration blur is, just as d is a scale factor for the diffraction. I could (and perhaps should) have defined Ra = Ro (r/a0)^n, so ...

Suppose blur due to diffraction and blur from aberrations can both be treated as convolution with a Gaussian of appropriate radius. Diffraction blur radius is inversely proportional to aperture ...

OK that's clear at last.

Speaking rather loosely, if the effective blur size due to aberrations varies as a higher power of the aperture than the diffraction blur size, then the peak does not occur exactly where the two ...

So you are saying that if the theoretical diffraction-limited MTF50 of an ideal lens at f/4 is 180 cycles/mm, but aberrations reduce this to 179 cycles/mm, then the lens is limited by aberrations, ...

<snip> So when you said and You didn't mean that for aperture radius a : Wavefront distortion = waves x ( r/a )^4 You instead meant a positive and negative-going distortion whose peak-peak ...

As far as I can tell, Airy is saying that both primary spherical aberration and axial colour reduce as the aperture stops down. I believe he is saying that axial colour decreases more slowly than ...

So your refutation is a spurious call a unspecified authorities. RMS value of r^n over the unit circle follows trivially from basic integral calculus. Mean square < r^2n >= { int[0 to 1] r^2n r dr ...

<snip> Good point. Dodgy numbers. RMS wavefront distortion is measured with respect to the mean value over the full aperture, not the central value at r=0. A constant offset corresponds to a ...

We can share the blame.  I was responding to your post with the Edmund link, and got distracted trying to remember the analytic formula for MTF vs F-number and wavelength. I neglected to refer back ...

Hi DM. Did you intend to link to the same page as my "diffraction limited" reference, or did you have something more comprehensive in mind? I linked to the top of the page, as I thought the table ...

So what is your opinion on applying RSS to non-orthogonal "W" aberration coefficients . If its OK to get a bit sloppy at 3:00 am, and not even offer an estimated tolerance band, then cut me some ...

Never? I was trying to figure out why there was a factor of 10 difference between Ted's 450 lp/mm, and the 46 lp/mm Lenstip resolution quoted by GB. I have done enough numerical simulations to know ...

FWIW, the Edmund optics site quotes 481 lp/mm at F/4 for the diffraction-limited extinction resolution (0% Contrast Limit) @0.520μm. Formula used is 1 /( λ F# ) https://spie.org/publications/tt52_1 ...

These look very much like transmission curves for the filter material only. Presumably after UV cure, and for the three thicknesses indicated - 0.7 μm, 0.9 μm and 1.1 μm. Note how the fractional ...

First difficult question: what do you mean by " ideal " in this context? Let's assume you mean " capable of accurately reproducing colours ", as opposed to " subjectively pleasing landscapes " or " ...