Optics science and simulation of a lens on image questions

I am researching to simulate very expensive , impossible to buy lenses on my photographs.I first wrote my thoughts on f295.org and than i found a paper .
white.stanford.edu/~brian/papers/ise/Maeda-Optics-SPIE-EI-5678-5.pdf
You need imageval.com / iset2.0 , matlab 7.0 and a lens analysis program like KDP2.It takes 2 hours to change a very very small picture. I did not do it but paper claims.
Astronomers do such a things for decades .
please fill the form :
even write yes or no.

A - First of I need to learn is ray tracing methods are universal ?Or do results vary from leitz factory to zeiss factory ?

AA - What do ray tracing find ?

B - What is the logic between aberrations , psf and mtf ?

AAA- Do aberrations change with the color of ray ?

C - Do psf change on the lens , photograph locations ? If yes , why astronomers change the whole picture with single psf ?

D - May be astronomy way is faster .Which astronomy photography software accepts above listed inputs for to change the photograph ?

E - Do Zeiss or Leica publish their methods , algorithms ? Where and by which name ? How can I follow their research in time ?

Regards ,

Mustafa Umut Sarac

...I am researching to simulate very expensive , impossible to buy lenses on my photographs.

Click to expand...

I don't quite understand, here. Do you mean "simulating the EFFECT that another lens would have if used to take the photograph in the first place?"

If so, that would be an extremely difficult thing to do. Not impossible (nothing is) but EXTREMELY difficult.

please fill the form :
even write yes or no.

A - First of I need to learn is ray tracing methods are universal ?

Click to expand...

Ray tracing principles are universal. How it is done? I wouldn't expect a great deal of difference among lens manufacturers.

Or do results vary from leitz factory to zeiss factory ?

Click to expand...

No.

AA - What do ray tracing find ?

Click to expand...

Everything about the lens; focal length; where the individual rays converge after entering and leaving the lens ... the "efficiency" and uniformity of coverage ...

- What is the logic between aberrations , psf and mtf ?

Click to expand...

First, I don't understand what is meant by "psf".
The answer to the relations between aberrations and "MTF" (Modulation Transfer Function is far beyond the scope of the APUG Forum. It would take a great deal of study to even begin to understand the information that the answer would provide.

AAA- Do aberrations change with the color of ray ?

Click to expand...

Aberrations are caused by errors in "where the rays go" ... that IS affected by their "color".

C - Do psf change on the lens , photograph locations ? If yes , why astronomers change the whole picture with single psf ?

Click to expand...

I've already asked that ... What is "psf"?

D - May be astronomy way is faster .Which astronomy photography software accepts above listed inputs for to change the photograph ?

Click to expand...

I have NO idea. Possibly an "Astronomy" site would the place to search for an answer.

E - Do Zeiss or Leica publish their methods , algorithms ? Where and by which name ? How can I follow their research in time ?

Click to expand...

That would be best answered by the Lens Manufacturers themselves. I would try a Google search to try to contact them.

As far as I am aware, no-one uses ray tracing any more: modern design programs use wave-front analysis.

Cheers,

R.

Here are my comments after a brief skim through the paper. It was a brief skim, so please forgive any misunderstandings.

The paper you refer to is about simulating the image a camera would produce, not about converting an image taken with one lens into that produced by another.

The paper concerns itself with the representation of two-dimensional subjects. That is difficult enough. What it doesn't address, because it isn't within the purpose of the work, is how the lens performs in three dimensions.

A lot of the character of a lens is how it handles the out-of focus areas. As the two-dimensional image produced by one lens lacks information about the third dimension it is going to be extremely difficult to remodel the blur created by one lens into that created by another. You might want to begin with a pair of in-focus images if you want to simulate the image of a three-dimensional scene. The process might then be:
1) Create a three-dimensional digital model of the scene.
2) Use lens simulation software to step through the scene one slice at a time to build up the resulting two-dimensional image.

That is not a trivial task.

I wonder whether it would be worth the effort. Most people who use digital techniques to simulate analogue photographic properties seem to be happy with rather poor imitations. What might be more effective would be a piece of software that did a very rough approximation of a dimly remembered and poorly observed lens characteristic, but was accompanied by a load of pseudo-technical nonsense about how accurate the software is.

This appears to be a 100% digital subject. APUG may not be the best place for it.

Best,
Helen

A - First of I need to learn is ray tracing methods are universal ?Or do results vary from leitz factory to zeiss factory ?

There are different flavours of raytracing. Point-to-point raytracing fixes the start and end points of the ray and tries to find an optical path between them, honouring Fermat's least-travel-time principle. Wavefront construction methods take small time steps from a starting position and build up a picture of the propagating wave. They are both high-frequency approximations to the exact solutions of the wave equation, which means they work when the variations in material properties (ie the refractive index) occur over a much larger scale than the wavelength. This means that lens coating effects, for example, may not be well-described by raytracing, since the coating thickness is small (comparable to light wavelengths). If you really want to model that, it is necessary to use numerical solutions of the full wave equation. Of course, that also involves some approximations.

B - What is the logic between aberrations , psf and mtf ?

'psf' is point spread function. If you have a point source of light imaged by an optical system, because of the imperfections in the system, it will not image as a point but be slightly blurred. Its actual image is the psf, which can vary across the image because the energy has followed different paths through the system. Different aberrations cause different shapes of pdf. mtf is related to the width of the psf in different directions and at different spatial frequencies.

AAA- Do aberrations change with the color of ray ?
Yes

C - Do psf change on the lens , photograph locations ? If yes , why astronomers change the whole picture with single psf ?

Astronomers can try to identify the psf by choosing images of very distant light sources if they are far enough away to be effective point sources. This means that in theory they could get a different psf for each point in the image. However, correcting the image would be very expensive and is mathematically unstable (the problem is almost singular). It is more stable, and MUCH cheaper, to assume that there is only one psf for the whole system. This is still unstable, but can be controlled by suitable regularisation techniques. You end up with an approximately corrected image where the residual psf varies across the image, which is probably better than the original as long as the psf doesn't change too much.

Pete H said:

A - First of I need to learn is ray tracing methods are universal ?Or do results vary from leitz factory to zeiss factory ?

There are different flavours of raytracing. Point-to-point raytracing fixes the start and end points of the ray and tries to find an optical path between them, honouring Fermat's least-travel-time principle....

Click to expand...

There you go ... getting all "accurate" on me - and reminding me how long it has been since I've been in the "business".

I've had experience testing lenses on an Optical BENCH!!! Remember those .. dating from "When the Earth cooled"?

Ed Sukach said:

I've had experience testing lenses on an Optical BENCH!!! Remember those .. dating from "When the Earth cooled"?

Click to expand...

My only point was that you might get slightly different answers depending on the algorithm used.
I'm a seismologist, not an optical scientist, so I use ray tracing to show that the earth hasn't finished cooling yet. You can continue to use your optical bench if you want to !

cheers
Pete

Pete H said:

My only point was that you might get slightly different answers depending on the algorithm used.

Click to expand...

Agreed.

I'm a seismologist, not an optical scientist, so I use ray tracing to show that the earth hasn't finished cooling yet.

Click to expand...

Oh ... @#$$%@ - All right ...:

"From when the Earth cooled to where the atmospheric and core temperatures are now." Wait, wait "... to where they were thirty, or so, years ago ...."

I hope that is sufficiently accurate.

You can continue to use your optical bench if you want to !

Click to expand...

Thanks, but I did not own them. They belonged to the "Specialty Optics" Companies I worked for as a Quality Assurance Specialist.

Speaking about "accuracy" ... Would you be interested in joining the f/63 Group?

Ed Sukach said:

"From when the Earth cooled to where the atmospheric and core temperatures are now." Wait, wait "... to where they were thirty, or so, years ago ...."

I hope that is sufficiently accurate.

Speaking about "accuracy" ... Would you be interested in joining the f/63 Group?

Click to expand...

Well, as a seismologist we measure various quantities to high precision, but the accuracy of our deductions is probably no better than 100%. So I guess I can shoot at f/63.00127 for precisely 1/8.31578 sec and easily be 100% out on the exposure, but I won't sweat it. Hey, I should fit right in

Pete H said:

Well, as a seismologist we measure various quantities to high precision, but the accuracy of our deductions is probably no better than 100%. So I guess I can shoot at f/63.00127 for precisely 1/8.31578 sec and easily be 100% out on the exposure, but I won't sweat it. Hey, I should fit right in

Click to expand...

Welcome!

I'm wondering about this "sloppiness"; Just how accurate ARE those jagged lines on that paper?

We could only infer our levels of accuracy - and performance - in statistical terms: The area under the curve, determined from empiracal statistical analyses .... Oh, by the way ... my "venue" was later in Metrology and Calibration - read, "WAY anal".

It is good to be in f/63 - and not give a damn, even remotely, about the "small stuff".