Canon’s new 600mm and 800mm lenses

Started 3 months ago | Discussions thread
alanr0 Senior Member • Posts: 2,326
Re: DoF of Canon’s new 600mm and 800mm lenses

bclaff wrote:

alanr0 wrote:

bclaff wrote:

alanr0 wrote:

bclaff wrote:

Not quite. If you toggle between the two scenarios while zoomed in close to the focal point you'll see the angle changes slightly.


While it's true that nothing changes behind the diaphragm the changes in front of it change the angle that the marginal rays pass the diaphragm and ultimately the angle at which they arrive at the image plane.

But doesn't a change in angle at the edge of the diaphragm mean mean that the marginal ray intercepts the axis at a different location?

No necessarily. A bundle of rays can leave one point and arrive at another; right?
I think it simply means that it arrives at the same point but at a different angle.

This is the case when the first point is imaged to the second point by the final lens group, which does not apply here.

If the aperture, final lens group and image plane are indeed fixed, my guess is that this is due to a small focussing error, possibly combined with a change in spherical aberration in the preceding optics.

If you zoom into the image plane, and switch scenarios, there is a small shift in focus. The minimum in beam width is slightly short of the image plane.

there is a small dependence of near and far focus distances on pupil magnification, due to the image-space asymmetry.

I don't think it's small.

However, the total depth of field (far-near distance) is independent of pupil magnification.

I don't think so.

See my equations 18, 19 here, or in more detail here.

I'm curious, what do you get for near and far focus for a circle of confusion of 0.29mm at closest focus?
(I have numbers from the ray trace and I'm curious to see if there is agreement.)

From your Optical Bench for the Nikon 800 mm f/5.6, at f/6.35 (re-setting exact aperture is tricky) and closest focus:

The Optical Bench Aperture Stopdown slider works in 1/3 stop steps so 1 click is f/6.35 at Infinity and 6 clicks is f/11.32 at Infinity.

OK - got it. I need to click within the circle of the slider, but not at the centre to get a single 1/3 stop step. A click in the centre does nothing, and a click outside the circle jumps multiple steps.

FWIW, once the slider has focus you can use the up/down arrows rather than the mouse.

Good to know.

And yes, anticipating your question below. I carelessly omitted a zero and did intend 0.029mm COC.

Focal length f = 540.91 mm

Magnification 6.50:1 m = 0.154

At working f-number Nw = 6.35; CoC = 0.29 mm, I calculate:

  • Exit pupil diameter p A = 29.3 mm
  • Ufar = 4132 mm; Unear = 3977 mm; 2 Δu = 155 mm.

Closer to the Canon aperture, using working aperture f/11, smaller circle of confusion diameter CoC = 0.029 mm, but the same magnification as above m = 0.154

  • Exit pupil diameter p A = 16.9 mm
  • Ufar = 4066.8 mm; Unear = 4053.3 mm; Total DoF = 2 Δu = 26.9 mm.
  • Ufar = 4066.8 mm; Unear = 4039.9 mm; Total DoF = 26.9 mm. ** EDIT **

Why 2Δu ? Isn't Total DOF = uFar - uNear = 13.5mm ?

In the equations I referenced, Δu is difference between geometric focus and either near or far point, so total DoF is twice this.

Still baffles me. I expect total DOF is uFar - uNear

Sorry - my transcription error.

Ufar = 4066.8, Unear = 4039.9, U_geometric_focus = 4053.3 mm.

Δu= +13.5 mm, -13.4 mm, Total DoF = 2 Δu = 26.9 mm

In any case, for f/11.32 in scenario 2 the Optical Bench gets ufar = 4056.9mm and uNear = 4029.4mm (and u at focus of 4031.1mm)
I have complete confidence in these ray traces.

27.5 mm total DoF - which is broadly consistent with my 26.9 mm (2.2% difference)

The Optical Bench values are shifted lower but this is almost certainly OK and due to the Optical Bench having full precision values such as for magnification.

But the high total DOF is due to pupil magnification less than unity.
This is what I suspected we might find.
My equations account for this. I suspect the ones you're using do not. (Even though I thought they had been reconciled.)

I still don't see a significant effect of pupil magnification. How are you estimating its impact?

Calculated using my equation 18, with similar total DoF predictions from my equation 19. Object distances (Ufar, Unear) are measured from first principal surface, so add 1236 mm for working distance from first element, and 1746 mm for distance to image plane.

You are quite careful in your work and usually correct so I wonder why we're not lined up on this. Perhaps there's a misunderstanding.

Even with your large CoC of 0.29 mm (did you mean 0.029 mm?) and an aperture of f/11, the image-space depth of field is (± CoC Nw) = ±3.2 mm. Exit pupil distance to image plane is 187 mm, so there is only ±1.7% asymmetry in the effective aperture at near and far image points. The first order contribution to object-space DoF cancel, so the impact on the total depth of field is negligible.

-- hide signature --

Alan Robinson

Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow