Nikon exit pupil size, effect on potential 17mm PC-E lens capabilities

[free1000]

I was showing my D800E to someone yesterday, and he mentioned that one reason why Nikon will have difficulty creating PC lenses as good as the Canon TS-E 24 and 17 is due to the smaller aperture in the camera body,  I believe he was saying that this limits the size of exit pupil of the lenses.

So am I right to reframe his conjecture that the smaller exit pupil possible with lenses on a Nikon camera means that lens designs with movements (lets stick to shifts which interest me most) are inevitably going to be compromised. I'm expecting that I've made a variety of technical errors in the way I've framed this, but hopefully the sense of the conjecture comes through.

I've decided to keep both my Canon with its clutch of stellar TS lenses for my commercial architectural work, while adding the D800E, so for me its not a massive deal breaker if Nikon are less able to match this capability of the Canon. It would be great if my friend were wrong, but there are some other benefits to me keeping both kits as well.

Is the exit pupil size always going to place a theoretical and/or practical limit on how great a Nikon PC lens could be?  It seems to me it might be, as it would have an effect on the size of the image circle, as well as lots of other effects on distortion, the shape of the field of focus etc.

Some people have mentioned Nikon have a patent for a 17mm lens. My feeling is I wouldn't buy a D800E solely on the expectation that Nikon can at some future date produce a 17mm as good as the Canon one. I think there are a bunch of other reasons for using the D800/E. I'm interested in what our optical experts think about this.

[Petrus]

I am just in the process of migrating from Canon system to Nikon, before a lengthy sick- & summer leave which started in May when I left all my Canon gear at work. I made a quick visit to the offices yesterday and was given a D4 with new 14-24, 24-70 and 70-200 zooms to play with and learn before returning to work in October. It immediately struck me how strangely thin the Nikon glasses are at the flange, coming from 10 years with Canon. It surely must affect TS lens design in a bad way. It would be easy to measure and make some drawings to find out how much shift you can put on a Nikon TS lens before serious vignetting, or even total cut-off will happen. Certainly much sooner than in Canon. That was not taken into account when the lens system was designed in the fifties and Nikon has stuck to their design since then, which is not a bad thing in itself. I have used the Canon 17 TC a bit and it truly is an amazing tool. That said, D4 is a fantastic press photographer's camera, maybe the best, and as I will get a D800E as a backup later and still have also Canon gear at my disposal when needed including all TS lenses (and X-Pro1 as a hobby camera), I am in a hog heaven right now!

[Ray]

I'm reminded here of one of the reasons I switched from Minolta to Canon in the film days before DSLRs became affordable.

I'd bought a Sigma 400mm lens which had had a good review in its Canon version. The review praised the lens for its relatively low vignetting at full aperture.

I bought the Minolta version, assuming that the same lens qualities would apply, but was very disappointed in the vignetting at wide apertures. What was the reason for this discrepancy between the Canon model and the Minolta model? Was it simply due to QC variation?

I then discovered on further investigation that the camera body aperture on the Minolta, where one attaches the lens, was smaller than the Canon body aperture. I wondered at the time if this could be the cause of the increased vignetting of this lens when used on a Minolta body.

I never found out for sure, but this factor was in the back of my mind when I switched to Canon.

Just out of curiosity, I've just measured the openings in my 5D and D800E bodies. They are 2&1/8th inches for the Canon and only 1 & 7/8th inches for the Nikon, approximately. That's a 1/4th inch difference, or a 6.3mm difference.

I would surmise that this might make it more difficult for Nikon to produce a PC lens of the same quality as the Canon TS-E, but I'm not sure.

[free1000]

Just out of curiosity, I've just measured the openings in my 5D and D800E bodies. They are 2&1/8th inches for the Canon and only 1 & 7/8th inches for the Nikon, approximately. That's a 1/4th inch difference, or a 6.3mm difference.

Doesn't seem much,  but my 'O' level maths tells me thats a difference in area of about 60%. Maybe thats more significant overall. Suprising result. 2.25" diameter gives 3.96 square inches and 1.175" diameter gives 2.4 square inches. So Canon is roughly 40% larger light gathering capacity.

I was getting hopeful when I read your measurements... now I've disappointed myself ;-)

[Petrus]

So Canon is roughly 40% larger light gathering capacity.

You are jumping into conclusions a bit too fast. Certainly Nikon can stuff just as much light trough their smaller opening, but it makes lens design more restricted, as the exit pupil of the lens has to be small, even if the maximum aperture was the same. For T/S lenses the problem is mechanical, with even a small shift the edge of flange starts getting on the way.

[ACH DIGITAL]

You are jumping into conclusions a bit too fast. Certainly Nikon can stuff just as much light trough their smaller opening, but it makes lens design more restricted, as the exit pupil of the lens has to be small, even if the maximum aperture was the same. For T/S lenses the problem is mechanical, with even a small shift the edge of flange starts getting on the way.

Already corroborated. You're right.

[free1000]

You are jumping into conclusions a bit too fast. Certainly Nikon can stuff just as much light trough their smaller opening, but it makes lens design more restricted, as the exit pupil of the lens has to be small, even if the maximum aperture was the same. For T/S lenses the problem is mechanical, with even a small shift the edge of flange starts getting on the way.

This more restricted lens design could be the kind of problem I was referring to.  Not saying that a design cannot be found to equal the Canon's, only with more constraints getting a better result is harder to achieve.  It would be nice if the total weight/size of a Nikon alternative was a little smaller, but with shift lenses I'd trade additional size and weight for the best results in terms of distortion, resolution, colour etc.

[Ray]

Doesn't seem much,  but my 'O' level maths tells me thats a difference in area of about 60%. Maybe thats more significant overall. Suprising result. 2.25" diameter gives 3.96 square inches and 1.175" diameter gives 2.4 square inches. So Canon is roughly 40% larger light gathering capacity.

I was getting hopeful when I read your measurements... now I've disappointed myself ;-)

Sorry! I've put you into the difficulty of converting from these imperial measures to decimal. 1 & 7/8ths of an inch for the Nikon is 1.875 inches. 2 & 1/8th inches for the Canon is 2.125 inches. There's a difference of 1/4" or 0.25 inches.

The area of a circle is given by: pi x the radius squared. The radius of the Nikon opening is 1.875/2 =  0.9375. Squared = 0.879. pi x 0.879 = 2.76 square inches for the Nikon opening.

Therefore, the opening for the Canon body, by the same procedure, is 3.54 square inches, the difference being 0.78 square inches, which represents a 28% increase in area.

How significant this is for lens design, I don't know. However, if my total wealth were to increase by 28% today, I'd be very pleased. 

[free1000]

Ha,  of course, Imperial measurements, I should know better, its the naughty step for me for the next hour.  Well I guess if NASA can do the same thing (and miss Mars) I needn't feel quite so bad about it.

28% is still a fairly noticeable difference.

[rethmeier]

Nikon will have to pull out all stops here.
It has to be as good as the Canon one.
I guess we will have to wait and see,
Cheers,
Willem.

[BernardLanguillier]

Weren't people saying before the D3/14-24 f2.8 that the F mount was probably not suitable for Full Frame sensors or wide angle lenses?

The current generation Nikon 24mm T/S seems superior to the previous Canon generation, so it is probably just another example of ping pong play.

I would be very surprised if Nikon bothered releasing a 17mm T/S that is not best in class, but measurments will tell. For now that lens doesn't exist yet.

Cheers,
Bernard

[rethmeier]

Well said Bernard and we want a D4x as well!
And a 24 PC-e MkII

[indusphoto]

Every 35mm lens has to focus the light to the same area. Both mounts provide openings that are sufficiently wide to illuminate the entire 35mm sensor (or film) without vignetting. The difference that probably matter more is the flange distance (the distance between sensor/film plane and mount). Canon has the shortest of this distance, and Nikon is probably the longest. What this means is that the rear-most element on Nikon can not be as close to the sensor plane as in Canon.

Given the geometry, the Canon can afford to make light fall at a more oblique angle through its opening (wider and closer) then Nikon. Which simply means that Nikon has to make its lenses a little longer (so that the light rays can be straightened before entering the body). If this is true, then it will explain why Nikon lenses are physically longer then their Canon counterparts for the same focal length.