Medium Format sharpness

[peter_hughes]

I don't really understand everyone's preoccupation with mammoth tripods. I use a Bogen 3021 with a 3055 ball head for my P6x7 and Hassy w/135mm & 150mm lenses respectively. Many would say that this 'pod is inadequate, but I've experienced no problems. I usually lock the mirror up on the P6x7 and trip the shutter with a long mechanical release, but I sometimes shoot handheld (at speeds of 1/125 and above). I'm a little more daring with the Hassy, often shooting on a "loose" ball head, no MLU, tripping the shutter directly. Of course, with my portraiture, the absolute n'th degree of sharpness is not as important to me as capturing a "moment of truth", so to speak. But if you'll check out my site, you'll see that my technique is more than adequate.

 

http://www.ravenvision.com/rvapeter.htm

[scott_jorgenson1]

Just wanted to say, with respect to the reply I made yesterday to Dan Brown:

 

Oops, I sloppily misread and thought you were using a Pentax 67 or 67II, but you clearly state you're using a Mamiya 7. Certainly, then, a major reason you're not experiencing any shutter vibration is due to the comparatively-tiny and quiet leaf shutters in that camera system.

 

Sorry to have misread your posting, Dan.

 

Scott

[russ_arcuri]

Someone above asked if 35mm lenses were sharper than medium format lenses. In general, this is true for comparable lenses. For example, a Canon 50/1.4 can resolve more lines/mm than almost any 80 or 100mm lens for medium format out there. Does this mean the prints will be sharper? Almost never. The negatives are much bigger in medium format, and therefore require much less enlargement for the same size print. So, unless the medium format lens is a real dog, medium format will produce better prints.

[john clark]

Ross, I've read somewhere that the Mamiya 7 lenses can comfortably exceed the lp/mm values of the Canon L glass. Can anyone confirm or deny this?

[michael_schuch]

I'm new to medium format but it sounds like people are overlooking the obvious. Is the focus screen the same distance from the lens as the film? I am a complete novice but my understanding is that you check this by opening the camera and putting a piece of ground glass on the film plane. Focus through the finder. Use a loupe to see if the image is sharp on the ground glass. If the image isn't sharp change the focus on the lens until it is. If you can get a sharp image on the ground glass but it is different than the focus screen you can either get the camera fixed or compensate?

 

Just a thought.

[dan_brown1]

..."Ross, I've read somewhere that the Mamiya 7 lenses can comfortably exceed the lp/mm values of the Canon L glass. Can anyone confirm or deny this?"...

 

Per photodo.com, which tests MTF at various resolutions, generally indicating lens sharpness:

 

EF20 -- 3.4

 

EF24 -- 3.9

 

EF50 -- 4.4

 

M43 -- 3.7

 

M80 -- 3.9

 

Seems to me they are pretty much comparable.

[john_hicks___]

Lenses for 35mm and medium format cameras will easily resolve several hundred lp/mm in aerial image.

 

A Perkin-Elmer engineer once told that me he checked out his 100mm Kinoptik lens and could clearly see 600 lp/mm, the limit of his test equipment.

 

The now-defunct Photopro mag hired an optics lab to test three Zeiss Hasselblad lenses several years ago; if I remember right, the lowest resolution figure was higher than 400 lp/mm.

 

The problem is that we're trying to project a 3-D world on a flat surface, and that surface often isn't flat.

 

Field flatness, or lack thereof, is one limiting factor. The more a lens has to be stopped down to gain adequate depth of focus, the lower the overall performance due to diffraction.

 

In larger formats, the biggest limiting factor is film flatness. The bigger the film the more it tends to buckle, particularly if it's wound around reverse-curve film-magazine inserts.

 

It's not too difficult to achieve 100 lp/mm on film with 35mm, but it takes quite a bit of care to achieve 80 lp/mm on medium format. So one could say that although both lenses perform equally or perhaps the lens for 35mm performs better because it can be designed for so much smaller coverage, 35mm is generally sharper than medium format.

 

But that's not the point.

 

Assuming perfect hardware and essentially ignoring the printing paper, in an 8x enlargement that 35mm neg isn't going to project more than 12.5 lp/mm on the paper no matter what, while the medium format neg is good for 20 lp/mm for the same size print.

 

Where many photographers new to medium or large format make a mistake is using their same loupes to examine the film without realizing that they're visualizing a significantly bigger print size.

 

If 8x is used for 35mm, then the loupe that should be used for medium format _assuming the same repro size_ is 4x.

 

Of course probably none of us actually do that. I use 8x, and if a medium format neg looks ok then I'm certain that it'll be sharp printed at 4x.

[struan_gray]

John, if a lens really resolves 600 lp/mm at visible wavelengths it

must have an aperture around f0.7 to avoid diffraction blurring. Your

Perkin Elmer engineer might well have seen a non-zero value of the

lens MTF at 600 lp/mm, but that's a very different thing from

resolving 600 lp/mm in any photographic sense.

 

Most modern MF and 35 mm lenses are limited by the film's resolution,

which for non-specialised emulsions maxes out somewhere around 100

lp/mm for high-contrast subjects and can be as little as half that for

normal pictorial contrast of 6:1. A better lens does help (the errors

add up), but beyond the film resolution limits the real benefit of a

good lens is the lack of contrast-reducing flare.

[mike_long]

To Michael's Point ---

 

If there is a point of sharp focus before or after the subject then,

yes, the screen could be out of position (or the back). This would be

more likely on the old Pentax 67 where a different screen had been

installed. Spacers can (and have for me) shifted and the point of

sharpest focus was behind the subject. However, as I recall the

specifications, the PX67II has a user replacable screen which would

indicate a manufacturing defect rather than a spacer shift if this

were the problem. While not impossible, it is less likely to be the

problem in this case.

 

A suitable test for this is placing a yardstick on the wall and

photographing it at a 45 degree angle. The mark you focus on should be

the sharpest point of the negative.

 

Mike

[tim_brown]

Struan sez:<br><i>"if a lens really resolves 600 lp/mm at visible wavelengths it must have an aperture around f0.7 to avoid diffraction blurring"</i><br><br>

Actually, if the diffraction limit is conservatively rated at 100 lp/mm at f11 then 600 lp/mm would correspond to about f2, quite possible.

[ken_heflinger]

600 lines/mm! Wow .. that is some lens! Can it achive this resolution corner to corner? How about for all colors? I have seen 300 lines/mm in the center of the field, using a microscope without ground glass to examin the image. But I think that test is unfair because it does not reflect what really can be achieved using film. Examining the image with a microscope and no ground glass permits my eyes to do a lot of correction to the image, in particular some degree of cromatic aperations can be corrected by the eye.

 

The best I have accieved on my negitives is about 80 lpm (TechPan film), with Tri-x only about 40 lpm. 600 lpm sounds a bit optimistic. Even Kodak only claims 400 lpm with TechPan. Most of the lenses I have seen start to loose their ability to focus accuratly when opened up all the way. What I have seen is that 1 or two stops down generally gives the best result, beyond that difraction takes over. The total resolution of the image is degraded by each limiting factor, diffraction, film resolution, and lens resolution. Even printing the image loses some resolution (printing involves lenses, and difraction too).

 

One item I have found is that the focus on the ground glass does not always match the "split image" part. The reason is that the split image view uses the edges of the lens .. the part that is not used when the photograph is taken if the lens is stopped down. A lens that has some aberations where the outter part does not focus to the same part as the center of the lens is now a problem. Focusing using the split image gives me slightly unfocused results. Focusing on the ground glass, or stopping down one stop and focusing on the ground glass gives me better results.

[john_hicks___]

> 600 lines/mm! Wow .. that is some lens!

 

It was a 100 f2 Kinoptik. I'd expect other modern 100 f2 lenses to be capable of similar performance.

 

But that has hardly anything to do with photography; that was eyeballing the aerial image through a scope.

 

My point is that what limits on-film resolution (for a given contrast) or if you look at it the other way, MTF, are mechanical things such as accurate focusing, camera alignment, accurate placement of the groundglass at the film plane or, in a reflex system, at the same distance, film positioning within the channel, film flatness and various vibrations including floor or ground vibration.

 

So it seems that no matter what we do we hit the wall of perhaps 80 lp/mm for medium format and maybe 125 lp/mm for 35mm....and that other things being equal, the more square inches of film the better.

[struan_gray]

<blockquote><i>

"Actually, if the diffraction limit is conservatively rated at 100

lp/mm at f11 then 600 lp/mm would correspond to about f2, quite

possible."

</i></blockquote></p>

 

I should probably have put more explanation into my comment - sorry if

I came across as dogmatic. The usual definition of 'resolution'

(which gives the f2-ish figure) assumes a pattern of white and black

bars and that that no more than 80% of the light from the white bars

ends up on the black ones. Roughly speaking, you end up with a

reduction of contrast from 1000:1 to 1.2:1 - which might satisfy a

surveillance team, but is a pretty lousy picture by most photographic

standards. A more useful definition for conventional imagery is that

the diameter of the Airy disc is equal to the width of one bar in the

pattern - this gives the f0.7 I quoted. Intermediate definitions are

also possible, and relevant to different types of imagery. You pays

yer money....</p>

 

It is perhaps worth pointing out that looking at the aerial image

through a microscope can produce all sorts of artifacts and

photographers trying to evaluate a lens are probably better off with

film-based testing. Unless the microscope objective is carefully

chosen, the effective f-number of the combined optical system can be

higher than that of the lens, which leads to spuriously high values

for resolution at wide apertures (ie, you think you're measuring a

wide aperture when you're not). Also, if your target consists of sets

of bars of various sizes it is hard to measure the phase error in the

projected pattern - at its crudest this appears as a reversal of

contrast, but more subtle phase shifts will affect images on film more

than is obvious from inspection of the aerial image.</p>

 

Personally, if I have a tripod-mounted shot that is unsharp I blame my

focussing. I find MF is where I make the most focus errors: LF is

less likely to be used on the run, and 35mm has more intrinsic DOF

(OK, so that's another can of worms, but you know what I mean).</p>

[glen_johnson]

Maybe this has already been hashed out above, but I wonder if its the difference between autofocus and no autofocus? I have tried the Pentax 67 on a few occasions (not the P67II), and found the finder to be very difficult to assess focus in.

 

When I was younger, this was never an issue. Now it is critical, and it is the major reason why I have been so happy to see autofocus come to 645. I wonder if you're in the same boat.

[bobatkins]

For comments on why aerial image testing is full of pitfalls and can lead to blind leaps of faith and totally incorrect conclusions if you don't know what you are doing, take a look at my web page <a href="http://bobatkins.photo.net/info/resolution.html">How NOT to get fooled by the results of aerial image

resolution testing of lenses</a>

[john_owlett]

When I were a lad, I learnt a rule of thumb for amateur photography that -- with prime lenses, slow film and a tripod -- you can safely enlarge a negative 8 times. This rule seems to be confirmed by an article "Is bigger better" in this month's issue of the UK magazine "Practical Photography".

 

It sent a photographer out to take landscape shots of a landscape (a stone barn with a lane in front and rolling downland behind) using

 

o (35mm) a Minolta Dynax 800si (sorry, don't know the US name)

 

o (645) a Bronica ETRSi

 

o (6x6) a Hasselblad 501CM

 

o (6x7) a Pentax 67 II

 

o (6x7) a Mamiya RZ67 II

 

The photographs were taken on 100 ISO E6 film (the article doesn't say which) and enlarged to 20 inches wide.

 

The writer commented -- as people here have -- about how the larger transparencies stand out on the lightbox when submitted for a competition or for publication; but he concluded

 

"If you never have a picture enlarged beyond 10x8in, don't enter competitions and have no serious intention of selling your work ... stick with 35mm."

 

(A 10x8 print is an 8 times enlargement of a 35mm transparency.)

 

"The quality available from the 6x7cm format is remarkable and the best here by some way. The benefits, even compared with 6x4.5cm and 6x6cm, are patently obvious."

 

(A 20x16 print is over an 8 times enlargement for 6x6, but less than 8 times for a 6x7.)

 

Owl

[john_hicks___]

FWIW, the Pentax 67 has a slightly different apparent focusing-screen difference than any other MF camera I've ever used.

 

I had the devil of a time getting the focus to snap in; this was finally solved by putting a +1 diopter in it. I don't need the correction with any other camera.

 

Of course Pentax diopters can be hard to find; a Nikon HP diopter will fit fine. Nikon's numbering system is odd, though; a Nikon #0 is in real life a +1.

[jim_chow]

If you enlarge full frame (why crop if you don't have to?), then 6x6 and 6x7 both enlarge to 7.26x on 16x20 paper. Also, most 6x7 cameras also have a frame length less than 70mm long..more like 68mm. Now figure in 1/4 inch of margin along the sides to allow for matting. The frame height is 56mm, so 7.03x in 56x56mm gives 15.5x15.5 while for 56x68mm, you get 15.5x18.8 inches. You pay a lot in weight for those extra 3 inches. You could be shooting a 4x5 field camera, and it would be lighter (and cheaper). :-)