Leigh Youdale
Well-known
To quote Miss Piggy, "What? Moi?"
learning about focus shift
- Thread starter meandihagee
- Start date
- Latest activity Latest activity:
- Replies 39
- Views 7K
To quote Miss Piggy, "What? Moi?"
Spyderman
Well-known
Don't forget that the focus shift is a direct result of spherical aberation.
Modern more lenses (e.g. planar design) have better corrected spherical aberation and thus less focus shift.
Theoretically aspherical lenses should have the least if any focus shift.
On the other hand lenses like Sonnar are very special because the spherical aberation is inherent in the design and it's a part of its signature and the reason why people buy these lenses. In this case the focus shift is more pronounced.
Modern more lenses (e.g. planar design) have better corrected spherical aberation and thus less focus shift.
Theoretically aspherical lenses should have the least if any focus shift.
On the other hand lenses like Sonnar are very special because the spherical aberation is inherent in the design and it's a part of its signature and the reason why people buy these lenses. In this case the focus shift is more pronounced.
Most of the Pop Photo and Modern Photo lens tests of the day noted focus shift of the lenses tested. It certainly did not just appear in the last few years. I certainly was aware of it decades ago. By F4, it is lost in the DOF with most lenses except Fast Sonnars. And with the latter, the lens is optimized for a specific aperture. By F5.6, the shift on a fast Sonnar is also lost in the DOF. If anyone is interested, I have a circa 1991 Pop Photo test of the 5cm f1.4 Nikkor-SC in LTM. It clearly shows a performance drop at F2 and F2.8 as compared with F1.4. This was due to focus shift.
With an SLR, the DOF preview was quite handy. But most of us just set the lens to F5.6 or F8 on an SLR and forget about shift. using wide-open, no need to worry. You focus and get what you see.
Shimming a Sonnar, using the M8- it is easy to fine tune the best performance to where you want it to be, or set a compromise between wide-open and stopped down. My own J-3, I optimized it for F2. The focus shift was 1" at 36", stopping down from F1.5 to F4. So optimize for F2, the F1.5 was within the DOF as was F4.
With an SLR, the DOF preview was quite handy. But most of us just set the lens to F5.6 or F8 on an SLR and forget about shift. using wide-open, no need to worry. You focus and get what you see.
Shimming a Sonnar, using the M8- it is easy to fine tune the best performance to where you want it to be, or set a compromise between wide-open and stopped down. My own J-3, I optimized it for F2. The focus shift was 1" at 36", stopping down from F1.5 to F4. So optimize for F2, the F1.5 was within the DOF as was F4.
Last edited:
This is the 35/1.7 Aspheric on the EP, 100% crops with the camera on a tripod.
Point of focus is the hole in the ruler.
F1.7:
and F4:
I could readily see the shift in the 5cm F1.4 Nikkor doing the same test.
Focus Shift is related to spherical aberration. The Ultron is an Ashperical lens. Focus shift from F1.7 to F4 is pretty much not an issue, at least when photographing rulers...
Point of focus is the hole in the ruler.
F1.7:
and F4:
I could readily see the shift in the 5cm F1.4 Nikkor doing the same test.
Focus Shift is related to spherical aberration. The Ultron is an Ashperical lens. Focus shift from F1.7 to F4 is pretty much not an issue, at least when photographing rulers...
Attachments
Sparrow
Veteran
There is much talk on the interweb about the Yeti, but it is seldom seen and seldom, if ever, photographed ... but then if it were I'm sure the prints would be abominable ...
Ranchu
Veteran
The old school gauss/biotar symmetrical lenses have less shift than the (non aspherical) Ultron types or the more assymetrical 'gauss' types too. Not so fast, though
Last edited:
Ranchu
Veteran
ferider
Veteran
A couple of related facts 🙂
1) lenses that shift noticeably, mostly have quite pronounced focal plane curvature.
2) most interweb focus shift measurements are done with dead center focus, irrelevant in practice.
3) many classic Leica lenses shift (Summilux 50, Noctilux, etc). They rarely get the same credit, as, say CV lenses 🙂
4) When you read about focus shift in the Leica context, more often than not the root cause of the problem was a mis-adjusted rangefinder.
Cheers,
Roland.
1) lenses that shift noticeably, mostly have quite pronounced focal plane curvature.
2) most interweb focus shift measurements are done with dead center focus, irrelevant in practice.
3) many classic Leica lenses shift (Summilux 50, Noctilux, etc). They rarely get the same credit, as, say CV lenses 🙂
4) When you read about focus shift in the Leica context, more often than not the root cause of the problem was a mis-adjusted rangefinder.
Cheers,
Roland.
I saw a Who pop out of a spec on a dandelion, used to test the Summicron, while pixel-peeping the images. "Who are you?", I cried. And he answered "Yes", and blew away in the wind.
Freakscene
Obscure member
I know that this is meant sarcastically, but taking this seriously, what you overlook is that until the early 1980s the thinnest film emulsions were ~ 25 microns thick. If the image is in focus anywhere in that 25 microns the photo will be sharp. Later, with advances in film coating technology that meant that thin emulsions were ~4-15 microns thick and then digital sensors, which effectively have depth approaching zero, these issues became much more important.
It's also worth noting that many Nikkors were slightly over-corrected for spherical aberrations, part of what gave then their reputation for sharpness and also, coincidentally, limits focus shift. it also meant that some of them show distinctive double-line out-of-focus rendition.
Marty
Last edited:
rxmd
May contain traces of nut
What makes you think so? If you look at the projection of a point into a space 200µ thick you get a cone (or two), which has a non-zero area. The photo won't be sharper for the thickness of the emulsion.
While there are some differences (mainly with regards to flatness of sensors, probably not thickness - ever wonder about Foveon sensors?) the main difference with digital sensors is that people are much more easily tempted to blow up their pictures 100x and look at them pixel-by-pixel.
If you do look at 100x60 prints of Cartier-Bresson's negatives with a loupe, I'm sure you'll find traces of focus shift. Then again, I think we all agree that only a complete dork would do that. The only reason why people feel differently about it with digital pictures is that it's so easy that one don't notice how silly it is. Next time you look at a 1:1 picture onscreen, just imagine yourself standing in an exhibition hall with a loupe, and I'm pretty sure that the focus shift will go away.
Harry Lime
Practitioner
Obviously focus shift exists, but the bigger question is how much of a problem is it in the real world? In my experience, it's a very small problem.
A few examples of focus shift:
Robert Capa - Zeiss Sonnar 1.5/50 - Focus shift central
http://www.fadwebsite.com/wp-content/uploads/robert-capa.jpg
Probably more concerned with not getting shot that morning than focus shift.
Nick Ut - Summicron 2/35 (v1, 8 element)
http://openlearn.open.ac.uk/file.php/1297/DD100_6005i.jpg
Too busy helping the girl after taking the shot to worry about focus shift.
Don McCullin - Nikkor 1.4/50
http://japanorama.co.uk/wp-content/uploads/2009/11/Don-McCullin-man-walks-to-003.jpg
"What's focus shift?"
I've never heard a working professional wring his hands about focus shift.
Somehow thousands, if not hundreds of thousands, of professional photographers produced spectacular images over the past decades, without ever uttering the phrase 'damn that focus shift'.
I've been shooting for at least 15 years and it's never even crossed my mind.
I own some of the lenses that are bashed on the net, because of focus shift (Lux 1.4/35 APSH v1, Ultron 2/28). It's never been an issue and I spend a lot of time shooting in the dark at 1.4 or 2.0.
It's another one of those subjects that the pixel peeping techno brigade loves to obsess about, when examining their cat shots at 800% in Photoshop. If these people spent as much time and energy obsessing about the quality of their photography, as they do about noise levels and purple fringing, the world would be awash in the likes of HCB.
A few examples of focus shift:
Robert Capa - Zeiss Sonnar 1.5/50 - Focus shift central
http://www.fadwebsite.com/wp-content/uploads/robert-capa.jpg
Probably more concerned with not getting shot that morning than focus shift.
Nick Ut - Summicron 2/35 (v1, 8 element)
http://openlearn.open.ac.uk/file.php/1297/DD100_6005i.jpg
Too busy helping the girl after taking the shot to worry about focus shift.
Don McCullin - Nikkor 1.4/50
http://japanorama.co.uk/wp-content/uploads/2009/11/Don-McCullin-man-walks-to-003.jpg
"What's focus shift?"
I've never heard a working professional wring his hands about focus shift.
Somehow thousands, if not hundreds of thousands, of professional photographers produced spectacular images over the past decades, without ever uttering the phrase 'damn that focus shift'.
I've been shooting for at least 15 years and it's never even crossed my mind.
I own some of the lenses that are bashed on the net, because of focus shift (Lux 1.4/35 APSH v1, Ultron 2/28). It's never been an issue and I spend a lot of time shooting in the dark at 1.4 or 2.0.
It's another one of those subjects that the pixel peeping techno brigade loves to obsess about, when examining their cat shots at 800% in Photoshop. If these people spent as much time and energy obsessing about the quality of their photography, as they do about noise levels and purple fringing, the world would be awash in the likes of HCB.
Last edited:
Freakscene
Obscure member
I've measured it on an optical bench, is what makes me think so. I don't follow your point on the cone projected from a point. When you print (or scan on a good scanner) you select where in the emulsion you focus on the paper or sensor.
That's partly it, but I saw a big difference when changing from Tri-X to TMY pre digital being a common medium. The f1 Noctilux really taught me about focus shift.
Not if the bride you're trying to photograph has sharp temples instead of eyes, visible in a 5x7, which is about how much the Noctilux shifts focus. The 75 Summilux is better, but harder to focus right in the first place.
Marty
Sparrow
Veteran
There was a program on the wireless, at the weekend, about that girl in Nick Ut's pulitzer shot, really interesting
http://www.bbc.co.uk/programmes/b00scvqh
http://www.bbc.co.uk/programmes/b00scvqh
rxmd
May contain traces of nut
Basically your assumption is that it's sufficient if the point of perfect focus falls somewhere within the film emulsion, and that because film is thicker, it's more forgiving because there is more room for perfect focus to fall into.
I'd say that this is a fallacy, because even if you project an image to some point of perfect focus within the emulsion, there will be a cone in front of and behind this point of perfect focus. Because you're still within the film emulsion, any halide crystals at these points within the cone will be exposed to light just as well. Obviously, crystal geometry and grain size plays a role here, too.
Printing is a completely different situation. When printing, you have an already-developed film with its grain. All you can hope to achieve when putting the focus somewhere in the emulsion is to get the grain on paper. It doesn't get any sharper than the grain. It's not like there is an even sharper image hidden somewhere within the film emulsion if you only find the right place.
I'm not sure whether you're talking from your own experience as a wedding photographer here. Getting moving subjects at f/1 into sharp focus has a pretty significant error rate anyway. Before starting to blame lost shots on focus shift, I'd suggest that it probably wasn't properly focused in the first place.
A good photographer should know their gear. This is really rule number 1. Someone who makes their living on getting people's eyes rather than temples in focus and who spends the $$$ on a Noctilux (plus a digital body?) even more so. If your salary depends on getting that shot of her eye in focus at f/1, with a 100% success rate because you get only that one shot and can't do bracketing, and you don't practice in your spare time until you get that no matter what optical quirkyness your lens exhibits, something is wrong with your job.
If you are sure that you are losing your own f/1 wedding shots to focus shift and not to something else, I guess it's time to learn to compensate - I find that I do fine focusing more by moving my body than by turning the focus ring on the lens, and if you know you have to compensate half an inch you can practice that half inch until you have it in the bones. But in all probably focus shift wasn't the problem to begin with; at extreme apertures people seem to consistently overestimate their own abilities, with moving subjects or at slow shutter speeds to boot.
Last edited:
Freakscene
Obscure member
Now I see what you mean, thanks. In practice, and I mean that, I think the main difference here is that a thicker film will be less sharp (we know this anyway). I doubt you could measure any practical difference in focus measured as resolved MFT from an image focused on the back or front of a ~200 micron thick emulsion. But I'm interested enough to look into seeing if I can figure out a way to test that.
This is probably a significant factor in digital needing less area and bits of data to transmit the same information when compared to film. It explains why some of the early wild overestimates of the number of megapixels needed to equal a 135 frame were proven wrong by viewing, quickly.
If you can resolve the grain entirely and decrease the depth of field enough, there is a sharper image in there. Some measured MFTs from film were made by shooting and developing frames and then viewing the negatives with a microscope, at enough magnification that the emulsion being viewed was only a few microns thick. These MFTs are misleading in terms of what you can get onto a print, but as you point out, prints are different. When you view exposed film with a microscope you learn a lot.
I know my gear very well. I do not overestimate my own abilities and usually try to hand-hold at no slower than 1/250 if I can avoid it. I use the M9 and the f1 Nocti only wide open. The problem is not at f1, it is from f2-4. By 5.6 the depth-of-field has caught up. It's focus shift, I've measured it. I tested the 0.95 and the control of focus shift and several other aberrations apparent in the f1 version is impressive, but it flared more wide open, which put me off. I'm learning to live with what I have. Sending the camera and lens to get returned to factory specs and matched to each other yearly has helped a lot.
Sure, I practice a lot. I've been doing it while I've been typing this, for instance. I've been doing it for quite a while now and my hit rate is good. I'd just prefer to be able to stick with the f1 version when I stop down but I find I can't easily compensate at a range of distances and apertures, so I use a different lens.
Marty
Ezzie
E. D. Russell Roberts
On a rangefinder, who (honestly) changes aperture after they set focus? I don´t, or very, very seldom. I´ve the CV35/f.14 and have never had any issues, even wide open and close up. When things are out of focus, it´s my fault.
Freakscene
Obscure member
Even if you do focus and then set aperture, it doesn't matter. The RF mechanism doesn't know what aperture you're using.
Marty
If you use a lens a lot, and get to know the amount of focus shift, you learn to compensate with the RF. Instead of lining the image up exactly, you line it up "slightly-off". For a Sonnar formula lens optimized at F1.5, you might focus slightly behind when using at F4.