Focus shift

[jaapv]

Advertisement (gone after registration)

I think this is hard to 'get' if you come from a background of SLR's where the lens is actually being focussed via the mirror onto the focussing screen (which has its own potential for error) which is where I guess the OP comes from, so can well understand his confusion...

Err... isn't it the lens that's prone to 'focus shift'..?

Nothing to do with whether its a rangefinder or an SLR.

Thus...

[adan]

It is interesting that I never encountered any mention of focus shift, nor thought of it nor noticed it, in my 30 years shooting SLRs and/or film. I never knew it existed. Although as Jaap says - obviously it has been in existence all along, and can affect SLR lenses focused at f/1.4 but shot at f/2.8 just as much as RF lenses.

 

One presumes that the confluence of being able to view images at 100% pixels (whether digital originals or film scans) - plus the pickiness of digital's infinitely-thin focal plane (compared to film's relatively "fat" gelatin coating) - plus improved lenses and higher expectations - plus the internet's ability to disseminate information - have all made this a "modern problem."

 

I mean, an old MF/SLR Nikkor-N 35 f/1.4 (and to some extent the f/2 as well) was so soft at any aperture above f/4, that who would notice "shift" as opposed to simply "dreamy?" We assumed that if you wanted really good sharpness, you stopped down 3 stops (where DoF would cover any shift), and if you used the wider apertures, don't expect too much. A "big print" was 11x14/A3, and who would notice shift at that size, without a microscope? Except maybe in macro shots. Negs were usually view at 8x (effectively 8x10/A4).

 

Of course, I was also creating "bokeh" for 30 years without knowing it.

Edited September 3, 2016 by adan

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[edwardkaraa]

With an SLR, focus shift causes front focus at wide open, even though the lens is wide open when you focus, the focusing screen is unable to the see the light rays from the f/1.4-2 apertures. That is why we were front focusing our 1.4 lenses during the old days without even knowing it.

[jaapv]

Sorry, you've lost me here. I struggle with a "screen unable to see light rays".

Do you mean the split image? That is indeed unusable in low light or with slow lenses as it blocks up.  The microprism ring was a bit better, but still not optimal. However, we knew to use a full matte screen or the matte portion of the universal screen in those circumstances. I fail to see what the lens aperture has to do with this in good light, though.

[Reeray]

Just extending this, what prompted my enquiry is an intention to acquire a 50mm Summilux Asph, which seemingly is subject to variable quality resulting in some lenses being very prone to focus shift.

 

I'd value the members thoughts on this.

 

My concern is that I'm shortly to visit the UK where the lens will be purchased and, should I encounter problems with shift, returning the item from Thailand would be problematic.

 

How would I check in the shop that I have a good copy?

 

Thank you

[jaapv]

Ask the shop (if it is really a good one) to assist you. Really testing a lens is a bit of a hassle. You need a fixed setup: tripod and target and time to evaluate the result. Just snapping a few random "test" shots in the shop is not going to work.

[edwardkaraa]

Advertisement (gone after registration)

Sorry, you've lost me here. I struggle with a "screen unable to see light rays".

Do you mean the split image? That is indeed unusable in low light or with slow lenses as it blocks up. The microprism ring was a bit better, but still not optimal. However, we knew to use a full matte screen or the matte portion of the universal screen in those circumstances. I fail to see what the lens aperture has to do with this in good light, though.

Focusing screens by design are blind to light rays from wide apertures so even when the lens is set at 1.4 what you effectively see is the image at 2.8-4. There are screens that are designed for wide apertures but these will turn dark when you use a lens of 2.8 widest aperture or smaller. I'm on mobile right now so I can't find the pertinent links, but you can scroll down to the focus shift paragraph in this link where it is briefly mentioned:

 

http://www.zeiss.com/camera-lenses/en_de/website/photography/what_makes_the_difference/manual_focusing.html

 

"Due to their surface structure, focusing screens in fast lenses are not capable of factoring in all the incoming rays. This can easily be seen from the fact that – depending on the design – the brightness of the focusing screen no longer changes with lenses such as those that are faster than 1:2.8. This means that the change in the optimum focus position can no longer be evaluated on the focusing screen at wide apertures when using very fast lenses."

[edwardkaraa]

Just extending this, what prompted my enquiry is an intention to acquire a 50mm Summilux Asph, which seemingly is subject to variable quality resulting in some lenses being very prone to focus shift.

 

I'd value the members thoughts on this.

 

My concern is that I'm shortly to visit the UK where the lens will be purchased and, should I encounter problems with shift, returning the item from Thailand would be problematic.

 

How would I check in the shop that I have a good copy?

 

Thank you

Just shoot anything with small print like a newspaper or magazine at 45 degrees wide open and see where the sharp image is in relation to your point of focus. Repeat at 2.8. This will give you a clear idea. Edited September 4, 2016 by edwardkaraa

[Exodies]

I'm intrigued by film's fat gelatin focal plane preventing focus shift from being visible. Does this mean that everything is as out of focus as it is when focus shift has taken place in film photographs?

[edwardkaraa]

I'm intrigued by film's fat gelatin focal plane preventing focus shift from being visible. Does this mean that everything is as out of focus as it is when focus shift has taken place in film photographs?

It isn't as dramatic as it sounds. Film has a physical thickness, with different layers, so it's more forgiving for slight focus errors or shift. But not all films are created equal. Black and white film is extremely thin, followed by negative color film, thickest being color reversal film. That is why slide film has actually the lowest acuity of all three. Digital sensors have virtually no thickness, as the image records superficially on the top of the microlenses, giving much less depth of field. Also film users are less critical about sharpness as they rarely enlarge to significant sizes, and film looses a lot of sharpness when enlarged anyway. But focus shift is present on both mediums.

[lct]

It's just that we did not pixel peep in the film days i suspect but focus shift was visible out of my 35/2 v4 on trannies. I though it was some sort of field curvature then and/or an effect of the Colorplan CF 90/2.5.

[jaapv]

Focusing screens by design are blind to light rays from wide apertures so even when the lens is set at 1.4 what you effectively see is the image at 2.8-4. There are screens that are designed for wide apertures but these will turn dark when you use a lens of 2.8 widest aperture or smaller. I'm on mobile right now so I can't find the pertinent links, but you can scroll down to the focus shift paragraph in this link where it is briefly mentioned:

 

http://www.zeiss.com/camera-lenses/en_de/website/photography/what_makes_the_difference/manual_focusing.html

 

"Due to their surface structure, focusing screens in fast lenses are not capable of factoring in all the incoming rays. This can easily be seen from the fact that – depending on the design – the brightness of the focusing screen no longer changes with lenses such as those that are faster than 1:2.8. This means that the change in the optimum focus position can no longer be evaluated on the focusing screen at wide apertures when using very fast lenses."

Ah- I see what you mean. Yes, but that depends on the fineness and structure of the screen. The R  screens, for instance, do not show this effect, as they used a kind of Fresnel structure to avoid this. I always used full matte screens on my R cameras - it also avoids the inherent problems with long slow lenses.

Come to think of it, lenses without spherical aberrations, i.e. focus shift, will not easily show this, as the edge rays, which cannot be captured due to the shape of the focusing screen grain, will not differ from the center rays.

[adan]

Focusing screens by design are blind to light rays from wide apertures so even when the lens is set at 1.4 what you effectively see is the image at 2.8-4.

 

Well, there are focusing screens, and then there are focusing screens.

 

Modern "bright-screen" AF focus screens are manufactured with a surface of microlenses (not unlike the microlenses on some image sensors). And as we know, microlenses do have a limited "acceptance angle" for incoming light rays, and may very well not pick up rays from the edges of wide-aperture lenses. They are really bright, since the microlenses "redirect" and focus all the light they do accept towards the eyepiece.

 

Old-school SLR screens, especially prior to the AF era (~1990) and in a lower-tech manufacturing era, simply had a random, "sand-blasted" surface. Literally, a "ground glass" or ground plastic. Which was far less efficient at redirecting light to the eye - but had a much wider acceptance angle for incoming light.

 

Get a hold of a pre-AF SLR, a Canon F-1 or Nikon F/F2, put on an f/1.4 lens, and use the stop-down lever to flip from f/1.4 to f/2 or 2.8 viewing. I'll guarantee you'll see the image darken dramatically and the DoF change - they use all the light from f/1.4, and it is noticeable if you remove those "f/1.4" rays by stopping down even a little.

 

The downside is - they don't direct light to the eye as effectively (they are darker/dimmer at all apertures) - and they get really dark getting down to f/3.5 or below. Condenser lenses on the back improved the brightness somewhat, especially in the corners, and, as Jaap says, fresnel plates did the same thing in a thinner package. The pro Nikon/Canon screens used both - and were really thick as a result. http://www.dryheatphoto.com/images/D7K6242xcrop.jpg

 

Great for focusing - not as pretty to look at, especially at small apertures. And the top-end interchangeable screens required a removable prism to swap out - too thick to feed in through the front.

 

So in an era of consumer f/3.5-5.6 kit zooms and AF, they had to go. Different era, different priorities.

 

For me, manual focusing took priority over "pretty" - so I tossed my otherwise brilliant F100s and went back to Nikon Fs, and eventually to RFs.

 

Canon offers optional "Super Precision Matte" screens, which are more like the old-school SLR screens. Great for improved manual focusing with fast lenses, using all the light rays at f/1.4 (as Canon says - "areas that are out of focus will appear MORE out of focus") - but "will appear dark and grainy with apertures below f/2.8".

 

https://www.learn.usa.canon.com/app/pdfs/quickguides/CDLC_FocusingScreens_QuickGuide.pdf

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Edited September 4, 2016 by adan

[Jeff S]

None of your lenses should have focus shift but it's easy to verify. Put the camera on tripod and shoot a ruler at 45 degrees at apertures from wide open to f/8, focusing of course with the RF. Not all lenses are optimized for the same aperture.

 

The focus shift in some samples of the 35 Summicron ASPH has been discussed frequently here (even though most folks think it only applied to the Summilux ASPH).  After careful tripod testing and screen peeping, I realized that mine exhibited the same mild shift at f4-f5.6 as the sample Sean Reid tested years ago.  In practice, though, it can be easily accommodated and presents no problems in my actual prints.

 

Jeff

[edwardkaraa]

Well, there are focusing screens, and then there are focusing screens.

 

Modern "bright-screen" AF focus screens are manufactured with a surface of microlenses (not unlike the microlenses on some image sensors). And as we know, microlenses do have a limited "acceptance angle" for incoming light rays, and may very well not pick up rays from the edges of wide-aperture lenses. They are really bright, since the microlenses "redirect" and focus all the light they do accept towards the eyepiece.

 

Old-school SLR screens, especially prior to the AF era (~1990) and in a lower-tech manufacturing era, simply had a random, "sand-blasted" surface. Literally, a "ground glass" or ground plastic. Which was far less efficient at redirecting light to the eye - but had a much wider acceptance angle for incoming light.

 

Get a hold of a pre-AF SLR, a Canon F-1 or Nikon F/F2, put on an f/1.4 lens, and use the stop-down lever to flip from f/1.4 to f/2 or 2.8 viewing. I'll guarantee you'll see the image darken dramatically and the DoF change - they use all the light from f/1.4, and it is noticeable if you remove those "f/1.4" rays by stopping down even a little.

 

The downside is - they don't direct light to the eye as effectively (they are darker/dimmer at all apertures) - and they get really dark getting down to f/3.5 or below. Condenser lenses on the back improved the brightness somewhat, especially in the corners, and, as Jaap says, fresnel plates did the same thing in a thinner package. The pro Nikon/Canon screens used both - and were really thick as a result. http://www.dryheatphoto.com/images/D7K6242xcrop.jpg

 

Great for focusing - not as pretty to look at, especially at small apertures. And the top-end interchangeable screens required a removable prism to swap out - too thick to feed in through the front.

 

So in an era of consumer f/3.5-5.6 kit zooms and AF, they had to go. Different era, different priorities.

 

For me, manual focusing took priority over "pretty" - so I tossed my otherwise brilliant F100s and went back to Nikon Fs, and eventually to RFs.

 

Canon offers optional "Super Precision Matte" screens, which are more like the old-school SLR screens. Great for improved manual focusing with fast lenses, using all the light rays at f/1.4 (as Canon says - "areas that are out of focus will appear MORE out of focus") - but "will appear dark and grainy with apertures below f/2.8".

 

https://www.learn.usa.canon.com/app/pdfs/quickguides/CDLC_FocusingScreens_QuickGuide.pdf

Thank you Andy for the detailed info. In fact, I was aware of different screen designs, as also mentioned by Jaap, but not as knowledgeable as you My understanding for simple one layer screens is that the ones designed for wide aperture lenses became too dark with lenses with smaller maximum apertures. That is why most standard screens are optimized for around f/2.8, as they won't get dark until f/5.6 but won't show the blur and depth of field of luminous lenses, just a compromise to make them work with a wide variety of lenses.

[edwardkaraa]

The focus shift in some samples of the 35 Summicron ASPH has been discussed frequently here (even though most folks think it only applied to the Summilux ASPH).  After careful tripod testing and screen peeping, I realized that mine exhibited the same mild shift at f4-f5.6 as the sample Sean Reid tested years ago.  In practice, though, it can be easily accommodated and presents no problems in my actual prints.

 

Jeff

Thanks for the info, Jeff. I have the feeling many rangefinder lenses do have some focus shift, even if the contrary is advertised by the manufacturer, probably because they consider it as being negligible. The ZM 35/2 is one of those as well. It shifts about 2 cm at MFD from f/2 to f/2.8 but behaves well beyond that.