MTF, PSF of Painting Photograph , What to do for painting

[Mustafa Umut Sarac]

As it is possible to measure MTF of a photograph to evaluate the lens performance.

I read it is even possible to measure PSF of a photograph to measure lens performance.

What to do for measuring psf , otf , mtf of a painting from photograph.

Umut

 

[Mike Crawford]

You are going to have to give a slight clue to us earthlings
as to what on earth you are talking about with your MTFs and PSFs. Then perhaps, and more importantly, why? Or did I miss something in my photographic studies that was mathematical instead of visual?
Regards
Mike
PS. To think, I almost wrote 'what the fuck are you talking about?'

 

[Nodda Duma]

Mustafa,

The transfer of a photograph to a painting by an artist is a non-linear system, whereas the MTF applies only to linear systems. Therefore, discussion of the MTF of a painting is non-sensical. The MTF is, of course, the modulus of the OTF, the linear transformation of a scene to an image.

The only linear system when viewing a painting (even a reproduction of a photograph) is your eyeball, which has a well-established optical model with a known MTF.

But I like your thinking. Look up the research into the specifications of the camera obscura used by Johannes Vermeer.

 

[Dr Croubie]

Technically possible, I would think, if you can hand-paint a test chart?
Measuring MTF or a painting of someone's face would be just like measuring MTF of a photo of someone's face, ie not very easy or useful because there are no standardised LP/mm with standardised contrast-ratios on a face (at least, on most faces).

 

[bernard_L]

I use this to optimize focus and USM on my s*****r.
http://rsb.info.nih.gov/ij/plugins/se-mtf/index.html
need first to install:
http://imagej.nih.gov/ij/
For s*****r evaluation I use a razor blade as a sharp edge. For a regular photo camera a painted/printed black/white boundary will do. You need to understand the frequency units in the plots of Slanted Edge MTF.

Fine points:

measure MTF of a photograph to evaluate the lens performance

Scene = input; lens+film = system; photo = output. The MTF is a property of the system, to evaluate it you need to know/measure both input and output.

Will you keep us informed of your results?

 

[Nodda Duma]

The other thing you can do to analyze a painting is to record an image of the painting at high-resolution, then perform a 2-D Fourier transform of the image to analyze the spatial frequency content, but this will not reveal an MTF of the painting itself. Spatial frequency content should provide useful information about brush stroke sizes, effects of aging, etc, especially if combined with spectral content analysis.

A painting is not a linear system, a fundamental requirement for an optical transfer function (OTF) and it's corresponding modulus (MTF). Beware of GIGO: Garbage in / garbage out.

If you tried to use the slant-edge method to measure the MTF of an image of a hand-painted edge, then the resultant MTF will simply be that of the camera you imaged the painting with, degraded by the effect of the painted edge not being ideal.

 

[Mustafa Umut Sarac]

Lets make a mind experiment.

There are thousands of painting scans at google earth. If I zoom a face , it would come with cracks. If I fourier transform it , I think these hundreds of cracks effect the analysis.

If someone wanted to design a lens which renders like an specific painting - may be low high pass filtes and lots of pgmented filters - , is there a chance for science.

I am going high and want to design a eyeglass which renders like a painting ! If I could design it , 3d printers are today capable to print very specific phase plates on to polymer eyeglasses for 20/20 vision with all high order aberrations are inside.

Umut

 

[Old-N-Feeble]

No purely analog/optical process will render life images as paintings. At any rate, what types/styles of paintings? There are many dozens of styles and countless individualized variations of those styles. I assume you aren't considering rendering like a a Picasso but more life-like similar to da Vinci or Michelangelo. Even so, the closest you'll come to automating that process is digitally.

 

[Mustafa Umut Sarac]

I want something wearable , not require computers or cameras. I think careful - like Leitz does- adjusting low pass high pass - I dont know whether aberrations and their physical represantations phase plates acts like low high pass filters but I have some ideas- filters are something makes a scene rounder , smoother or sharper.

I found canon digital cameras are not product of lenses and sensors but smooth degrades , shadows and highlights in shadows are the product of sensor low pass filter.

I learned that low pass filters are two glass pieces , first render the image vertical and one render the image horizontal and two low pass the details and moires production.

I bought a Leitz Dygon and it was rendering a point just like as vertical coma 3.

I point the lens to a bright point and blurred halo appeared around the bright point and two copies of point appeared vertically. I looked to the zernike renders and this was exactlly Coma 3 vertical.

As you see , an aberration can do what canon low pass filter can do.

I think my lens was 50 years old and Canon analyzed many similar lenses and reached to a conclusion.

I posted Elcan paper on 35mm Summicron and there were lots of intense aberrations before and after the stop.

As myth told , Leica did this to carefully adjust the especially relief effect and the low high pass trick.

I will try to simulate aberrations on 35mm summicron lens by effecting a reference image but computers does not process 3d images as a lens can do today. I think I will produce and try phase plates by 3d printer.

Tonight I will look at zernike images and try to find a horizontal symmetric aberration and the place on summicron 35.