Backfocus and Focus Shift: The Plot Thickens

[tashley]

Advertisement (gone after registration)

Tim, that second 100% crop (with the 28) is done at a subject magnification which is 50/28 = c. 1.8 x larger than the first one (with the 50). This is not a fair comparison. It would get you thrown out of any decent physics department. Do compare apples with apples. It would be a miracle if the first crop didn't look sharper than the second!

 

The old man from the Age of the Wet Print

 

 

Lars, I am well aware of this, and referred to it in the post itself. Nonetheless the 50 lux shot is downrezed, the 28 cron shot is more than 1.8 times as fuzzy, and there is nothing in the full sized file that is in focus. I will be posting more rigourous examples later but I think that anyone who looks at those two shots can see there's a problem with the cron, despite the mildly unfair comparison.

 

Nonetheless, to half-prove my point, here's another 28 cron shot taken this morning, also at F4 as per yesterday, different light. Here I took five shots, one with the lens set to infinity, one with it focussed on this crop point, and then three with focus pulled progressively closer until for this one, the sharpest, it was set at ten meters. The building is perhaps 80 meters away. Here the cron is pretty much as sharp as the 50 lux if you allow for the less contrasty light and the wider FOV.

 

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!

 

 

 

Tim

[Advertisement]

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

[BerndReini]

This is already sharpened for printing with NikSharpenerPro at 25%, also Alienskin Exposure Grain applied. The print looks beautiful, so save the comments on the sharpening and grain please.

 

Downrez or not, I can definitely see that there is a problem with one of Tim's lenses since I also own bot the 28cron and 50lux asph. They should both be piercingly sharp. This is what the 28cron should look like and if it doesn't, your rangefinder or lens are off. It is probably the sharpest lens out there.

 

Full picture, center crop, and edge crop.

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!

[tashley]

Thank you Bernd, you eye-poppingly sharp shot shows quite exactly why I purchased the lens and what I'm not getting!

 

Thanks for rubbing it in!

 

;-)

 

Tim

[wattsy]

even my late model 135 tele-elmar (now sold) where I would think I would notice even more discrepancy than with a wide lens (methinks you have that backwards).

 

No, I believe I have it right. It does sound counterintuitive but any error in the mount to focal plane distance will be more apparent with shorter focal length lenses than longer ones.

[rob_x2004]

Well if they arent right out of the box then Leica deserves a kick in the pants.

My question is have Leica made optical compromises that sees a certain percieved out of focus at infinity or any other place. Tims definition of infinity is different to Berns version of infinity.

I am jsut asking. Leica cant build cameras to save themselves, that is already a given, but I trust their optics.

[Guest stnami]

Leica can't build cameras to save themselves, that is already a given, but I trust their optics.

............ well waving and drowning at the same time is a bit of a handicap.................

[tom0511]

Advertisement (gone after registration)

The main problem in my opinion is the following:

Even if you r lenses and M8 are calibrated fine, the problem of the thin ensor surface (compared to the thicker film layer) does lead to a different sharp-unsharp-transition.

It also leads to the fact that minor focusing inaccurancies are recognized (which might not have been the case with film).

This means focusing digital rangefinders totally accurate is much more important than it was in film times.

Hyperfocal shooting gets more difficult. Shooting moving subjects gets more difficult.

We might need to close the f-stops more often.

Overall -in this regard-a digital rangefinder might be more limited compared than it is when shooting film.

 

Cheers, Tom

[rob_x2004]

Yea? What happens when you go buy a second body?

Anyway besides the point. are the lenses right or wrong out of the box? I figure they are. I'd like a 28.

[ChrisC]

It does sound counterintuitive but any error in the mount to focal plane distance will be more apparent with shorter focal length lenses than longer ones.

 

Ian - That's always been my understanding too; that the depth of focus can be very shallow with wide angle lenses [whilst having a concomitant large depth of field, of course], whilst longer focal length lenses have a deeper depth of focus, and narrow depth of field. Therefore; getting coincidence of focus plane and image plane with wide angle lenses demands fine tolerances. If I am wrong on this; I'm game for a tweak to my ignorance base.

 

Wretched luck Tim, it's a sorry tail - especially as you have been feted by Leica.

 

.........Chris

[Peter Branch]

There are at least three different effects being discussed and poured over. I shall ignore inaccurate equipment calibration which should simply be corrected.

 

1) Film was never “flat”. There was a seminal paper published in the British Journal of Photography, (“BJP”), by Zeiss about 30 years ago regarding the limitations the curvature of the film posed for the performance of fast wide angle lenses such as 35mm f/1.4. This inevitably leads to some compromise. Lenses may have “deliberate” field curvature, c.f. the Curved Field projection lenses made by Leica is an extreme example. The setting up of the lens in manufacture may position the point of best focus slightly behind the plane defined by the camera’s engineering. This will nearly always give a more “optimum” result on film.

 

Now we have a near perfectly plane sensor. Some lenses will cope better with this than others. E. Puts has to his credit repeatedly pointed out that some non-Leica lenses have less Curvature of Field. They will now gain some advantage. Leica are likely to respond with their new designs but there is, as yet, little or no evidence that Leica has deliberately flattened the image plane of their lens designs for the digital age; the possible exception being the new 28mm f/2.8 ASPH.

 

2) Most of the films used in Leica cameras are quite incapable of showing up the small differences in performance that are now being easily seen by those with an M8, some software and a computer. Film resolution and grain simply masks many of the subtleties now being discussed whether the film is viewed by projection or is scanned. These effects were always there but it took specialist films used under laboratory conditions to be able to reliably reproduce them. Why would anyone other than film manufactures bother?

 

3) Focus shift with aperture. This has always been a fact of life. If the system allows for the “optimum” image to be identified, (M8 or specialist film), then this effect is easily seen. The Noctilux is a very fine example where if an image is in focus at f/1 and you stop down, leaving everything else exactly the same, except exposure time obviously, then at f/2 the image will appear to be significantly out of focus – actually back-focused. This focus shift continues with further stopping down but is masked by the rapidly increasing Depth of Focus so the image appears to be getting better again.

 

What we are witnessing are some of the many effects that the M8 has brought about and with which a company like Leica, dedicated as it is to the highest possible image quality, will have to come to terms. In the medium to long term only new lenses optimises to different parameters will keep them ahead – in the mean time we have the best that is available and it is surprisingly good.

[Guest guy_mancuso]

I have not read what anyone said here which is good so it does not sway what i think and have done. Like everyone I was bouncing around on this subject with some of my lense for infinity and such . My conclusion to all of this is and what i adjusted for is pretty simplistic but it works . I have on all the M8's that passed through my hands is optimized my focusing for around 10 ft. and made sure every lens i own can hit a critical focus around that length wide open, what i gave up thing was infinity had to be sharp wide open. Why on earth would anyone shoot wide open at a infinity mark. Would most landscape shooters shoot at 1.4 when there shooting the Grand canyon, never would happen . Think about this almost every infinity shot is not shot wide open but what is people at 10 ft or thereabouts and what would be the most critical length for wide open shots , the people shots. This is a comprimise in my book i rather have everything critically good before infinity for wide open shooting but at infinity i am going to stop down anyway so why sweat it. Now i am talking a ever so slight difference here but i optimized my system for critical sharpness before infinity. Now i am nailing everything in sight with all my lenses. So you ask hwo did i do this , first get the fastest lens like a 50 lux or 75 lux and tune for that than try everything else. Also if you turn the camera over so the hex screw is facing up than i tuned it going left with the allen wrench so it is just hitting the infinity mark with lenses but used a 75 lux as my final word and tuned for that and around 10 ft or so than treid every lens i had and they all it the same area wide open. Now this is what i found that works for me and it maybe completely wrong by Leica wisdom and such but it seems to be working

[tashley]

I have not read what anyone said here which is good so it does not sway what i think and have done. Like everyone I was bouncing around on this subject with some of my lense for infinity and such . My conclusion to all of this is and what i adjusted for is pretty simplistic but it works . I have on all the M8's that passed through my hands is optimized my focusing for around 10 ft. and made sure every lens i own can hit a critical focus around that length wide open, what i gave up thing was infinity had to be sharp wide open. Why on earth would anyone shoot wide open at a infinity mark. Would most landscape shooters shoot at 1.4 when there shooting the Grand canyon, never would happen . Think about this almost every infinity shot is not shot wide open but what is people at 10 ft or thereabouts and what would be the most critical length for wide open shots , the people shots. This is a comprimise in my book i rather have everything critically good before infinity for wide open shooting but at infinity i am going to stop down anyway so why sweat it. Now i am talking a ever so slight difference here but i optimized my system for critical sharpness before infinity. Now i am nailing everything in sight with all my lenses. So you ask hwo did i do this , first get the fastest lens like a 50 lux or 75 lux and tune for that than try everything else. Also if you turn the camera over so the hex screw is facing up than i tuned it going left with the allen wrench so it is just hitting the infinity mark with lenses but used a 75 lux as my final word and tuned for that and around 10 ft or so than treid every lens i had and they all it the same area wide open. Now this is what i found that works for me and it maybe completely wrong by Leica wisdom and such but it seems to be working

 

 

That's good pragmatic logic Guy, and for those handy with a hex wrench I'd say go for it for tuning issues.

 

I just flash tested my 28 cron on my second body and first impression is that it is exactly the same on that one. In other words, the lens is focussing way past infinity. It would take Prof Stephen Hawking to define the exact nature of the place where it is focussing!

 

The trouble with all this, as other posters have noted, is the too-large number of variables. I am pretty confident that both my bodies (one 'hexed' and one 'virgin') are correctly calibrated or thereabouts, since they both give great results with most of my lenses. The two lenses that have 'issues' are the 35 cron (which we all know about) and the new 28 cron. But they have different issues. What is confusing is that their issues look similar at first sight. However, it is clear to me that the 35 has classic 'focus shift with stopdown, but not saved by DOF' whereas the 28 cron has 'backfocus at all distances and apertures, far worse on objects further away, subtle but noticeable close up.'

 

The problem is, you need a lot of experience and testing to ascertain what problem you have and it is too easy, IMHO, to miss the subtle issues with an individual lens and reach for the hex, only to risk screwing up the others.

 

It is all worth it of course, because when it all works the results are so very fine - but I think people should do a LOT of careful testing and thinking before they start playing with their rangefinder calibrations: if they have more than one lens that works perfectly from 1 metre to infinity, as do I, especially if this is on two separate bodies, then they should see the faulty lens as the item for adjustment and not the camera.

 

I have had two 35 luxes and one 35 cron with focus shift on stopdown. I have had one 24mm elmarit which was smeary only on one side. I had a 90mm F2 that backfocussed. I now have a 28 cron that does the same. BUT I have also had a big enough sample of other lenses that are crisp at all apertures and distances to know

 

a) what a really crisp shot should look like and

how to ascertain (slowly, painfully) what the problem with a given lens is

 

The one thing I would add is that I have bought a lot of brand new glass. I wonder if that is one of the variables to consider?!

 

All the best

 

Tim

[Jamie Roberts]

Well, I have to weigh in here...

 

Tim, it seems you've got a lens problem, as I suspected all along. Your RF might be out as well.

 

First, I want to say categorically that film depth versus sensor flatness has nothing whatsoever--in this universe OR the next--to do with the effect that Tim is seeing.

 

I've shot enough film--and enough with the M8--to know this is, in practice, complete and total "hooey," as they say (though curvature of field is not... but that affects the shape of the field, not the in-focus focal point).

 

BTW--I really don't care what Puts or anyone else has to say about this, since my own prints prove to me that the M8 is as sharp--or sharper--than any other 35mm form factor digital I've ever shot.

 

So, for Tim, who has even more of sympathy than he did before, once again (just like the 35 ASPH), we have an effective difference on a very sharp lens--of several FEET off.

 

Yikes.

 

I'm glad Mark mentioned this salient point: you can't fix a lens adjustment with a RF adjustment. That's not true for any other camera maker too, RF or not (which is why when Canon CPS fixes a focus problem they usually request the lenses as well).

 

The mount and lens both need adjusting, and of course with the M8 the RF does too (and on an SLR even a mis-aligned screen can mess things up!)

 

Once BOTH camera and lenses are adjusted, if necessary, you will get interchangeability.

 

It's not a matter of needing to adjust when changing lenses.

 

Tim obviously has the very bad fortune of some misadjusted LENSES. As a consequence, his rangefinder might be out as well, or the lenses may be masking an RF problem.

 

Consider, for a moment, if the 50 Lux and 90 cron he has are at the very edge of spec--and are effectively out in the same way. If that's the truth, then all his other lenses will show some slight effect--and the amount of that effect will vary depending on the lens itself.

 

I'm just saying what Mark already said, so I'll shut up .

 

However--there are a certain number of M8s that DO backfocus... I've heard by up to about 4 inches at close focus. These can be adjusted with the hex fix (like Brad's, like Carsten IIRC, and like mine).

 

The question is 1) do you have a backup strategy if you mess up and 2) are you confident you have the specific backfocus issue?

 

So on my own system, all of mine were backfocussed by about 2-3 inches wide open at 5 feet. That's a lot. Fixing the infinity focus (with the hex key)--and NOT touching the close focus--fixed my system for every lens but my used 75 Lux (canadian version).

 

To me that means the 75 is out, and I'll get it adjusted.

[bsmith]

There are at least three different effects being discussed and poured over. I shall ignore inaccurate equipment calibration which should simply be corrected.

 

1) Film was never “flat”. There was a seminal paper published in the British Journal of Photography, (“BJP”), by Zeiss about 30 years ago regarding the limitations the curvature of the film posed for the performance of fast wide angle lenses such as 35mm f/1.4. This inevitably leads to some compromise. Lenses may have “deliberate” field curvature, c.f. the Curved Field projection lenses made by Leica is an extreme example. The setting up of the lens in manufacture may position the point of best focus slightly behind the plane defined by the camera’s engineering. This will nearly always give a more “optimum” result on film.

 

Now we have a near perfectly plane sensor. Some lenses will cope better with this than others. E. Puts has to his credit repeatedly pointed out that some non-Leica lenses have less Curvature of Field. They will now gain some advantage. Leica are likely to respond with their new designs but there is, as yet, little or no evidence that Leica has deliberately flattened the image plane of their lens designs for the digital age; the possible exception being the new 28mm f/2.8 ASPH.

 

2) Most of the films used in Leica cameras are quite incapable of showing up the small differences in performance that are now being easily seen by those with an M8, some software and a computer. Film resolution and grain simply masks many of the subtleties now being discussed whether the film is viewed by projection or is scanned. These effects were always there but it took specialist films used under laboratory conditions to be able to reliably reproduce them. Why would anyone other than film manufactures bother?

 

3) Focus shift with aperture. This has always been a fact of life. If the system allows for the “optimum” image to be identified, (M8 or specialist film), then this effect is easily seen. The Noctilux is a very fine example where if an image is in focus at f/1 and you stop down, leaving everything else exactly the same, except exposure time obviously, then at f/2 the image will appear to be significantly out of focus – actually back-focused. This focus shift continues with further stopping down but is masked by the rapidly increasing Depth of Focus so the image appears to be getting better again.

 

What we are witnessing are some of the many effects that the M8 has brought about and with which a company like Leica, dedicated as it is to the highest possible image quality, will have to come to terms. In the medium to long term only new lenses optimises to different parameters will keep them ahead – in the mean time we have the best that is available and it is surprisingly good.

 

Well said Peter, finally intelligent comments to the M8 focusing issue.

[bsmith]

I have not read what anyone said here which is good so it does not sway what i think and have done. Like everyone I was bouncing around on this subject with some of my lense for infinity and such . My conclusion to all of this is and what i adjusted for is pretty simplistic but it works . I have on all the M8's that passed through my hands is optimized my focusing for around 10 ft. and made sure every lens i own can hit a critical focus around that length wide open, what i gave up thing was infinity had to be sharp wide open. Why on earth would anyone shoot wide open at a infinity mark. Would most landscape shooters shoot at 1.4 when there shooting the Grand canyon, never would happen . Think about this almost every infinity shot is not shot wide open but what is people at 10 ft or thereabouts and what would be the most critical length for wide open shots , the people shots. This is a comprimise in my book i rather have everything critically good before infinity for wide open shooting but at infinity i am going to stop down anyway so why sweat it. Now i am talking a ever so slight difference here but i optimized my system for critical sharpness before infinity. Now i am nailing everything in sight with all my lenses. So you ask hwo did i do this , first get the fastest lens like a 50 lux or 75 lux and tune for that than try everything else. Also if you turn the camera over so the hex screw is facing up than i tuned it going left with the allen wrench so it is just hitting the infinity mark with lenses but used a 75 lux as my final word and tuned for that and around 10 ft or so than treid every lens i had and they all it the same area wide open. Now this is what i found that works for me and it maybe completely wrong by Leica wisdom and such but it seems to be working

 

bizarre

[Jamie Roberts]

Well said Peter, finally intelligent comments to the M8 focusing issue.

 

No, I think they're mostly a red herring.

 

For example, if what Peter says is true, I have

 

1) A magic camera

2) A magic set of lenses

3) and a very Magic Noctilux, where I shot 4GB of test shots on the M8 with a Noctilux from f1.0 to f5.6 . They were all in focus at the focus point. IOW, the lens did not go "out of focus" by f2.

 

Since I don't believe in magic, but I do believe in internet nonsense, I've made my own conclusions

[charlesphoto99]

Well, as I've said, my 28 Summicron, which had a major overhaul at Solms last year due to my fumbling fingers, is unsharp compared to all of my other lenses. Don Goldberg, who is not a repairman unfamiliar with Leicas (ha!) said it was in proper adjustment. So obviously there is some sort of disconnect between the M8 and that lens. One of my concerns is having lenses adjusted specifically to the M8 - does that render them useless on film bodies? (yes, I plan on still using my M7 on occasion).

 

Anyway, I'm going by my dealer today in the hopes he has a 28 Summicron and can try and get to the bottom of it.

 

I think one of the reasons some have resorted to diy repair like myself, is due to the expectation of buying a 5K camera, not to mention $500 worth of filters, etc (I still have several hundred $ of lens codng ahead of me) the thought of the dreaded 6-8 week visit to Solms (not to mention $ for shipping) and not sure if when it returns it won't have to go straight back is a bit much. Mine still has to go back for a row of high iso dead pixels but am waiting to see what other issues come up first. One trip is enough!

[Jamie Roberts]

{snipped}One of my concerns is having lenses adjusted specifically to the M8 - does that render them useless on film bodies? (yes, I plan on still using my M7 on occasion).

{snipped}!

 

Not if film is thicker and more receptive to mis-focus! (just kidding folks!!)

 

A proper lens / rangefinder adjustment shouldn't render anything unusable on your film bodies. As Mark said, both need to be adjusted to an external reference and be within tolerance.

 

If your film body is also within that tolerance, it should be fine.

 

The only question then, is whether the M8's closer tolerances (according to Puts) are within a subset of the film tolerances. And of course, properly adjusted, they are.

 

And just to show I'm not anti-Puts, here's what he says about the M8 and focusing:

 

"In reality the [increased manufacturing tolerances over film M bodies] factor is about three to four times. This makes the M8 the most accurately machined and assembled M camera in history. This is done to compensate for the lack of image capture thickness.

 

When focussed accurately the focus plane is spot on on the sensor surface."

 

So even if I disagree about the whole film thickness thing (and I'm willing to be wrong about the theory), practically speaking with a properly adjusted M8 there is no difference.

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[tashley]

Thanks to Jamie, Charles, Bill, Peter Chris and all for comments and thoughts.

 

I just did some more tests. I chose a distance of close to exactly 28 metres (which is theoretically infinity on the 28 cron) and shot the same scene with a 1.25 magnifier, ISO 160, on a tripod with the cron first on my new black body (which has always behaved and never had its hex nut touched) and then on my old Chrome body (which I adjusted very minorly with a hex). Then, on the chrome body still, I nudged the focus three little shifts closer for the third shot. This equated to a start point of about halfway between 10M and infinity on the barrel, as chosen by the RF's observed focus, to an end point of exactly 10M. F4 throughout.

 

The two bodies when focussed according to their RF's both give pretty identical results. The last shot, 'guess focussed' and therefore not as good as the lens might be capable of, is clearly in much better focus.

 

I am not seeing spherical changes in focus here as I did with the 35's and I have yet to do the ruler test to check specifically for focus shift. Either effect may exist as well, but the main thing we are seeing here is that the lens is simply backfocussing, and focussing past infinity. This is made clear by the fact that peripheral areas on one side of the frame are well into the infinity area, but only in focus in the third shot where the focus was pulled forwards. And I have to assume, since both bodies behave the same and give me no trouble at any focal length or aperture with all my other glass (35 cron excepted but that has special needs!) that my RF's are both pretty well adjusted. BTW I have also tried Jono's 28/35/50 tri-elmar, another 'trusted' lens, on both bodies with perfect results.

 

Other 20 cron users take a look at your lenses. Don't rely on 1 metre to 3 metre shots, where the effect is subtle, go for close to or past infinity. And check if your RF shows exact object coincidence on something five miles away. Mine shows 'twins' when the cron is mounted, and aligns perfectly with all other lenses.

 

Best

 

Tim

 

On each of the first two shots I focussed on the chrome hose stand. The third was was guess focussed by shifting focus forward.

 

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!

 

 

 

Finally the full frame for reference:

[BerndReini]

Tim,

 

I've been following your post about your 35lux (love your Venice shots btw.) and have since been very concerned about backfocusing etc. So much so that every time I buy a new lens, I will take my camera body to the dealer and give the lens a quick test run. I've never had a problem. When I bought my M8, I tested it at different distances with a 50lux. I have had no backfocus problems whatsoever with any of my six lenses. However, I did try three different versions of the 35lux asph for a friend of mine and all three of them showed the backfocus problem, yours suffered from. I myself own the old handmade 35lux aspherical and it is just perfect. I tried three different 28crons though and have not seen a problem with any of them. This makes me think that it might be your M8.

We're almost back to the days of the Linhof Technika when the rangefinder cam, body, and camera had to be perfectly matched to each other for sharp focusing. My recommendation to everybody would be to test every lens before you buy it whenever that is possible. Of course, I understand that with the 30% off lens etc. this is not possible.

 

Anyway, the point of my picture post was to ensure you that there is definitely a problem with your M8/28cron combination and don't let anybody tell you otherwise.