# Canon’s new 600mm and 800mm lenses

Started 3 months ago | Discussions
 Forum
Plane of Best 'Focus'

bclaff wrote:

Jack Hogan wrote:
... I see that Alan has provided a list of available alternatives for criteria to be used.

I had a friend who worked for DARPA who used to say "the great thing about standards is you can pick whichever one you want".

Maybe the difference is minimal with well corrected lenses, though it seems that something of the order of one p-v wavelength OPD for rho^4 aberrations is somewhat normal for such at lowish f-numbers. The question is not trivial because it is a determinant of the plane of best 'focus' in the center. Which of course affects DOF, as Alan suggested (not to mention a lot of other stuff).

Where is best 'focus' from a photographer's perspective? W is p-v OPD in microns, N is f-number. The minimal geometric CLC occurs further than suggested by minimum RMS.

The plane of best 'focus' in the center of the field of view in practice is often taken to occur when the p-v Optical Path Difference caused by, say, SA3 is offset by defocus' ( W020 = -W040), based on RMS units*. So we can estimate the location of the plane of best 'focus' in either case.

I went through this thought process here (see around Equation 4):

https://www.strollswithmydog.com/dof-diffraction-image/

The choice W020 = -W040 in the center comes from reasoning in RMS units but it is also evident when looking at aggregate MTF, hence energy concentration, as shown there in Figure 1, and below in Figure 7

https://www.strollswithmydog.com/simple-model-for-sharpness-in-digital-cameras-spherical-aberration/

Jack

* Geometrically the plane would instead be further away, at W020 = -1.5*W040, as can be easily gleaned from the optical bench image above. Wyant is our friend for stuff like this.

Complain
Re: Plane of Best 'Focus'
2

Reminds me of something a divorce attorney told me once; he said:

"The best divorce is one where the parties are equally unhappy."

-- hide signature --

Bill ( Your trusted source for independent sensor data at PhotonsToPhotos )

Complain
Re: Plane of Best 'Focus'

Jack Hogan wrote:

Where is best 'focus' from a photographer's perspective? W is p-v OPD in microns, N is f-number. The minimal geometric CLC occurs further than suggested by minimum RMS.

The plane of best 'focus' in the center of the field of view in practice

FWIW, the 'plane of best focus' using only diffraction effects (not geometrical) can also be defined in a number of ways. Here is one way as in the link below - note that for usual photography it doesn't matter much, though, because of the range of typical F#'s. So it is more instructive. But, outside that range it has a marked effect. And, it is interesting to note it goes against the thinking that smaller cone angles are covered better with paraxial optics:

https://www.dpreview.com/forums/post/60963794

-- hide signature --
Complain
Re: Plane of Best 'Focus'

Joofa wrote:

Jack Hogan wrote:

Where is best 'focus' from a photographer's perspective? W is p-v OPD in microns, N is f-number. The minimal geometric CLC occurs further than suggested by minimum RMS.

The plane of best 'focus' in the center of the field of view in practice

FWIW, the 'plane of best focus' using only diffraction effects (not geometrical) can also be defined in a number of ways. Here is one way as in the link below - note that for usual photography it doesn't matter much, though, because of the range of typical F#'s. So it is more instructive. But, outside that range it has a marked effect. And, it is interesting to note it goes against the thinking that smaller cone angles are covered better with paraxial optics:

https://www.dpreview.com/forums/post/60963794

That's an interesting post and thread Joofa.

Complain
Re: DoF of Canon’s new 600mm and 800mm lenses

AiryDiscus wrote:

The original author of LightTools has released an open source raytracer. Odds are quite high there isn't a better open source one out there, since he is probably among the most knowledgeable people out there on the topic.

Hi, do you have a link for above?

Thanks and Regards

Dibyendu

Complain
Re: DoF of Canon’s new 600mm and 800mm lenses

Dibyendu Majumdar wrote:

AiryDiscus wrote:

The original author of LightTools has released an open source raytracer. Odds are quite high there isn't a better open source one out there, since he is probably among the most knowledgeable people out there on the topic.

Hi, do you have a link for above?

I think you meant this:

https://github.com/mjhoptics

Complain
Re: DoF of Canon’s new 600mm and 800mm lenses

yes

Complain
Re: DoF of Canon’s new 600mm and 800mm lenses

I tried using it a while back - it failed to run at the time:

https://github.com/mjhoptics/ray-optics/issues/4

I haven't tried since.

Like all ray tracing opensource products the documentation is poor, but that's understandable in this case as it is still under development.

A more complete and working opensource package is https://github.com/quartiq/rayopt.

Unfortunately the author is not developing it any further it seems.

Complain
Re: DoF of Canon’s new 600mm and 800mm lenses

Dibyendu Majumdar wrote:

I tried using it a while back - it failed to run at the time:

https://github.com/mjhoptics/ray-optics/issues/4

I haven't tried since.

Like all ray tracing opensource products the documentation is poor, but that's understandable in this case as it is still under development.

A more complete and working opensource package is https://github.com/quartiq/rayopt.

Unfortunately the author is not developing it any further it seems.

My comment is simply to say that Mike all but certainly knows more about making a raytracer than anyone else that's released an open source one.  LightTools is a Goliath among commercial raytracers with a long heritage and revered stability.  It's listed first in the design competition submission formats for a reason.

Complain
Re: Canon’s new 600mm and 800mm lenses
1

Old Greenlander wrote:

Chris in UK

tell me how you get a shallow DOF with those lenses...

Shallow DOF depends on how larger the pupil looks as an angle, from the subject's perspective. It has no direct relationship to f-ratios.

F-ratios do not appear visibly in images. Pupils do.

Like so many other similar situations in this big world, people have a tendency to confuse casuation and correlation, ignoring confounding variables.

Yes, if you vary the f-ratio, all else being equal, the higher f-ratios give the deepest DOF. That does NOT mean that DOF is a direct characteristic of f-ratio.

Perhaps I used to be confused myself at one time, but it is still hard for me to fathom why people ascribe so much absolute meaning to things like f-ratios, ISOs, and exposure, when in fact, they are removed from direct relevance by a number of other factors. This may not affect the landscape photographer much in a way that causes miscalculations, but for people using long lenses, or who always crave for longer lenses, the pupil and distance paradigm tends to be more direct, as you do more object- or subject-relevant calculations, and less sensor-frame-relevant ones.

What is real in a photograph (as far as the underlying analog image is concerned), is exposure time, perspective, and effective pupil size. What we are left with is whether or not the size of the sensor is too narrow to host our composition, and whether there are enough pixels to resolve our composition to our satisfaction.

f-ratios are for completing exposure triangles, if that floats your boat, or for AF systems, where f-ratios of open lenses can affect AF ability, because of low light flux, but also because of optical geometry, despite sufficient light quantity. Low (not high!) f-ratios can run into problems with photosite/microlens optics, losing more of the light from the outer regions of the exit pupil, but this generally only occurs at f-ratios less than 2.8.

From the same distance, 400/6.3 or 200/3.2 have a little bit more noise, a little bit more DOF, and a little bit more diffraction than 800/11, in the underlying analog image. 800/11 is photographically equivalent to 400/5.6 or 200/2.8, and one would only get more subject noise, DOF and diffraction with the 800/11 if the narrow angle of view, when combined with the sensor size, forces one to back away from the subject to a longer distance.

-- hide signature --
Complain
Re: Canon’s new 600mm and 800mm lenses
1

This is a rather dogmatic way of looking at DOF. With the same framing, the DOF depends on the f-stop only when it is shallow enough. The size of the upupil makes no visible difference. From the same distance, things change.

We have many variables, some of them dependent on the others, so absolute statements like this make no much sense.

With long lenses we often get strong background blur which is well outside the DOF.

Complain
Re: Canon’s new 600mm and 800mm lenses
2

John Sheehy wrote:

Old Greenlander wrote:

Chris in UK

tell me how you get a shallow DOF with those lenses...

Shallow DOF depends on how larger the pupil looks as an angle, from the subject's perspective. It has no direct relationship to f-ratios.

Depth of field and depth of focus are two sides of the same coin; object and image space.  They are related by the longitudinal magnification.

F-ratios do not appear visibly in images. Pupils do.

(2)

pupils do not appear in focused images. The focused image is the convolution of some truth with the magnitude squared of the fourier transform of a pupil.

Your way of thinking is irreconcilable with lenses like these, which have both pupils at infinity and of undefined size:

https://www.thorlabs.com/newgrouppage9.cfm?objectgroup_id=7381

On the other hand, these lenses have finite NA/F#, so nothing is lost the way that has "no direct relationship."

Like so many other similar situations in this big world, people have a tendency to confuse casuation and correlation, ignoring confounding variables.

Yes, if you vary the f-ratio, all else being equal, the higher f-ratios give the deepest DOF. That does NOT mean that DOF is a direct characteristic of f-ratio.

Perhaps I used to be confused myself at one time, but it is still hard for me to fathom why people ascribe so much absolute meaning to things like f-ratios, ISOs, and exposure, when in fact, they are removed from direct relevance by a number of other factors. This may not affect the landscape photographer much in a way that causes miscalculations, but for people using long lenses, or who always crave for longer lenses, the pupil and distance paradigm tends to be more direct, as you do more object- or subject-relevant calculations, and less sensor-frame-relevant ones.

What is real in a photograph (as far as the underlying analog image is concerned), is exposure time, perspective, and effective pupil size. What we are left with is whether or not the size of the sensor is too narrow to host our composition, and whether there are enough pixels to resolve our composition to our satisfaction.

f-ratios are for completing exposure triangles, if that floats your boat, or for AF systems, where f-ratios of open lenses can affect AF ability, because of low light flux, but also because of optical geometry,

that "geometry" argument, without commenting on its legitimacy, would lead you to support an NA/F# based formalism, no?

despite sufficient light quantity. Low (not high!) f-ratios can run into problems with photosite/microlens optics, losing more of the light from the outer regions of the exit pupil, but this generally only occurs at f-ratios less than 2.8.

The microlenses on most sensors these days are on the order of F/1.5.  Canon puts NA 0.8 lightguides on their cine cameras.

Complain
Re: Canon’s new 600mm and 800mm lenses

I think these lenses will not sell well and have limited productivity. Have you shot at f/9-11 much? Prepare to clean your mirrorless sensor more and frankly, forget shooting in low light.