Digilloyd on RF Focusing Accuracy

[Bill Pierce]

You may find the following interesting. When I was using film Leicas as my workhorse cameras, it was almost automatic that when you bought a new M2 or M3, you sent it to someone like Norm Goldberg. First, he would relubricate the camera which had probably been sitting on a shelf for at least a year. Second, he would null-null the rangefinder mechanism; that is, to the limit of the accuracy of the measuring tools, go for absolute accuracy, no acceptable tolerances. In this way, all your bodies matched. You would also send him all your lenses. The rangefinder cams would be matched to your null-null bodies, which often meant the rangefinder cams had to be reground. Thus, whatever lens/body combination you used, you achieved a high level of focusing accuracy.

Unfortunately, Leitz will not make the necessary custom tools available to zero the M8/M9 rangefinders to top flight repairmen like Norm's son, Don. Their overworked in house repairman take a long time to deliver and, in my personal experience, often do a bad job that should have never left their shop. As far as I know, they do not offer any kind of null-null service to the general public. Those folks with multiple high-speed lenses and several digital bodies, the real Leica users, are hurt the most.

Here’s the interesting conversation from the Digilloyd site.

Correspondence with Erwin Puts on Leica M9 focusing.
Erwin Puts is well known for his Leica and other photography expertise, and in my exploration of the backfocus issues with my Leica M9, I decided to ask him what his Leica M9 focusing experience had been.
DIGLLOYD: I’ve been having multiple issues with calibration of focus for the M9 and lenses— backfocus in particular. Easy to repeat in the field or a target with various lenses, and that's using the 1.25X magnifier. How has your experience been with focus calibration? Do you take special steps to ensure calibration?

Hi Lloyd,
This is a difficult topic. When testing camera/lens combos with the M9 I often need to adjust focus distance to get best contrast and sharpness. The fact that the focus is not always accurate has different causes: the focusing cam on the lens might be out of tolerance or the accuracy of the rangefinder even in combo with 1.25 magnifier is not sufficient.

DIGLLOYD Too bad. I came to the same conclusion, and was hoping you had some "magic". My experience shows that getting perfect focus with more than one or two lenses for the M9 is a serious challenge.

I use two methods' one crude and one systematic. Crude: On a tripod I focus as best as I can, make the picture and enlarge on the display: if sharp OK if not sharp slightly defocus in one direction and if this is sharper continue. If it is not sharper try the other direction.

DIGLLOYD: I have used the same approach, which I deem “LCD zoom focus”. Alas, the M9 does not zoom in to actual pixels, or even sharp pixels, and the low-res screen further degrades the accuracy of that approach. Close, but no cigar. I discuss the whole composition and focus area at length in my review of the M9. I don’t know why Leica can’t offer a once-per-second Live View, it would help. Or a decent LCD screen. Or sharp actual pixel zoom.

Systematic: on a tripod I have a slider mechanism with 1mm distance scale. I focus accurately and have the slider in mid position. Then I take a series of pictures moving the slider in 1cm steps in both directions for a range of 5 cm. On the computer I select the best image.

DIGLLOYD: I sometimes use a slider setup with a 1mm engraving. It works. But it’s useful only for close up work, not at any distance. And of course some lenses don’t perform as well as when focused close, and some also exhibit shift focus when stopping down, so it’s very tedious for each aperture, and not very interesting for real world images.

The back focus issue is a sign that the rangefinder is at its limits. But so are many DSLR cameras. There you can individually calibrate a lens for the AF system.
— Erwin

DIGLLOYD: We are in agreement here in general. Yes, AF isn’t perfect, but it can focus off-center, and in conditions where a rangefinder is hopeless, and it can do so very quickly. Using Live View zoomed in with a loupe eliminates the focus issue for a DSLR entirely. I've rarely had to calibrate lenses for AF, but it’s a great feature. For precision work, Live View is mandatory anyway, as it eliminates a variable. Even contrast-detect AF can be unreliable, with both Nikon and Canon, observed by experience. The reason we don’t see Live View with the Leica M9 or S2 is that CCD sensors do not support it.
==============

 

[kbb]

I like DigiLloyd, but his assertion that "...Yes, AF isn’t perfect, but it can focus off-center, and in conditions where a rangefinder is hopeless,..." seems a touch off. Low light: no a/f but RF works. Low con scene: gimme an optical RF any day. Bright light + contrasty scene: I'd rather control the focus, but at least the gizmotron will work. -Keith B-

 

[Bill Pierce]

I was wrong when I said, "Unfortunately, Leitz will not make the necessary custom tools available to zero the M8/M9 rangefinders to top flight repairmen." No special tools are needed. Roger Dunham, whose thoughts brought about the Digilloyd post, corrected me.

 

[Bill Pierce]

It seems strange that someone would buy a $9000 camera body and fit it with ultra high speed lenses costing up to $6500 and not expect the rangefinder focus to be dead on when the lenses were wide open. This is no small order on a digital rangefinder camera. The acceptable tolerances on the lenses and the bodies are minimal. Achieving that dead on focus with a number of lenses and multiple bodies is exceptionally difficult.

Right now my own experience and what I’ve been told by others leads me to believe that Leica is not always meeting this demand. Hey, it’s a really tough job.

I’ve received conflicting reports about whether it is possible to match lenses to multiple bodies. If there is someone on the forum who has the answer to this question, I would love to hear it. If it is possible, hooray. If it isn’t, it’s going to tend to turn the camera into an excellent camera for someone who can get away with one body. But that’s not a photojournalist who has to use a number of focal lengths relatively rapidly or a documentarian stuck in the boonies and praying their camera won’t break.

Once again, does anyone know if it is possible to match multiple high speed lenses to multiple digital bodies so as to achieve accurate, wide-open focus at a variety of distances with all possible combinations of your equipment?

Thanks.

 

[sojournerphoto]

Once again, does anyone know if it is possible to match multiple high speed lenses to multiple digital bodies so as to achieve accurate, wide-open focus at a variety of distances with all possible combinations of your equipment?

Thanks.

Bill

A ppint well made and the asnwer is that it must be possible, but you need to find someone with the skills, time and wilingness to do it. That may make it impracticable. RF focus accuracy is often further out than we like to admit, ad the consistency with which we achieve inaccurate focus doesn't help🙂

Mike

 

[ferider]

I don't know about M8/M9, Bill.

But thanks for sharing your story about Norm's calibration of bodies and lens kits. I've had (individually and not null-null, obviously) calibrated film bodies and lenses, where one lens/camera combo would work, and the other one not, but all items were within specs.

Should also be an eye opener for other users of vintage film equipment, long and/or fast lenses, assembled together from ebay, KEH or our classifieds, maybe individually CLA'ed, certainly not together, and most likely not by a repairman who does null-null calibration ....

Best,

Roland.

 

[JoeV]

It seems strange that someone would buy a $9000 camera body and fit it with ultra high speed lenses costing up to $6500 and not expect the rangefinder focus to be dead on when the lenses were wide open. This is no small order on a digital rangefinder camera. The acceptable tolerances on the lenses and the bodies are minimal. Achieving that dead on focus with a number of lenses and multiple bodies is exceptionally difficult.

Thus lies the conceptual downfall of the rangefinder focusing method, its reliance on mechanical tolerances between components that have to slide and wear against one another in order to function, and the requirements of extraordinarily precise machined tolerances between lenses and bodies that may not have always been manufactured to those same standards.

I enjoy using rangefinder cameras, but also know that they have intrinsic limitations. Thus the high cost (for me) of a new Leica system is not justifiable. For professionals who can write off the cost, that may be a different calculus, however.

~Joe

 

[antiquark]

Thus lies the conceptual downfall of the rangefinder focusing method, its reliance on mechanical tolerances between components that have to slide and wear against one another in order to function, and the requirements of extraordinarily precise machined tolerances between lenses and bodies that may not have always been manufactured to those same standards

Agreed. If you apply some simple trigonometry, the rotation of the rangefinding prism has to be accurate to hundredths of a degree, to get spot-on focusing.

Nikon realized this was a fundamental limitation of the rangefinder mechanism, especially in relation to tele lenses. That's why they designed the Nikon F. (Or so I've read.)

 

[sojournerphoto]

Thus lies the conceptual downfall of the rangefinder focusing method, its reliance on mechanical tolerances between components that have to slide and wear against one another in order to function, and the requirements of extraordinarily precise machined tolerances between lenses and bodies that may not have always been manufactured to those same standards.

I enjoy using rangefinder cameras, but also know that they have intrinsic limitations. Thus the high cost (for me) of a new Leica system is not justifiable. For professionals who can write off the cost, that may be a different calculus, however.

~Joe

Of course, slrs aren't without their issues either. if the focus screen/sensor is not accurately aligned with the filmplane/sensor you will have a built focus error that is more apparent with wide angle lenses (because of depth of focus). With autofocus bodies, it's not unknown for the focus sensor, film/snsor plane and screen to be aligned differently. Modern bodies allow you to compensate for lens/focus sensor variation, but you're still at the mercy of accurate assembly for the viewing ground glass.

I just use what I like🙂

Mike

 

[splitimageview]

DAG has calibrated many M8s and M9s and their lenses. When I inquired as to his equipment and methods in this regard, this was the resultant thread. I am unaware if he can match bodies but I would guess he could, given enough time.

"Cost to adjust the M9 R/F is $80.00, no need to send lenses unless you want to, may not be a bad idea to at least send one lens just so we can both agree that the camera & that lens focus right and that's because so many lenses seem to be front or back focusing & I'm adjusting the focus perhaps to the thickness of scotch tape (or less) & that can really make a big difference with the M8, 9 cameras. Film cameras wouldn't even notice it."

"I don't understand why they can't do this at the factory. What are you adjusting, the cam? Shimming the lens mount? I am assuming you have special equipment to dial things in with high precision."

"Yes, proper collimators, shims & magnification optics. I have lenses made of tool steel by Leitz, Infinity targets, 1 meter test stand & 35 years of adjusting Leica R/Fs."

 

[Tom Rymour]

Does anybody know -- would a foveon sensor be less demanding on a rangefinder?

 

[dseelig]

rf focus

rf focus

I do not know maybe I am lucky but other then my 75 lux I ahve had very good lcuk focusing my rf lenses all luxes 24 35 50 and 75 . The 75 had to go to Germany for work but everything else was a breeze. My 90 cron is a little tougher in bad light but a lot of that is shooting moving people at 1/90 of a second.

 

[glenerrolrd]

I do not know maybe I am lucky but other then my 75 lux I ahve had very good lcuk focusing my rf lenses all luxes 24 35 50 and 75 . The 75 had to go to Germany for work but everything else was a breeze. My 90 cron is a little tougher in bad light but a lot of that is shooting moving people at 1/90 of a second.

Dave

An awful lot has to do with how critical the focus point is to your composition. I have the same lenses and have had multiple bodies . With the 24 even wide open I have a little DOF to work with but more important I rarely have a composition that demands perfection. The 50 and the 75 are where you see it because of the need for eye sharpness and the easy of seeing ..OH darn back focused. If you have 50 lux asph that focuses accurately then your are fortunate. (thats the lens that really hard to calibrate).

But your luck was getting a well calibrated body. The M8.2 and the M9 are both much improved over the M8 .

Roger

 

[Ranchu]

I do find slr's superior for accurate focusing, it's not as fast as RF's though. I rarely use AF anymore because I don't like dicking around with focus points, or focus and reframing which moves the focus plane. AF cameras are really almost useless for manual focusing, the screen is almost always not where it should be, and not matte enough to focus manually. The screen on my F100 actually hurt my eye to try to focus with it, that sucks. Another problem is that with AF cameras, the AF sensors are under the mirror, so it has to be semi-transparent, which makes a proper matte screen even darker.

Well, that was random. I think the bottom line if you want exact, controllable focus accuracy is to use an RF with lenses below a 50mm FOV and optimally only with one lens. Or manual focus slr. Or an AF slr with a really dark screen.

I mostly use an AF slr with a really dark screen because I love the F100's shutter and minimal mirror blackout. Plus my RF broke.

 

[Bill Pierce]

I think the bottom line if you want exact, controllable focus accuracy is to use an RF with lenses below a 50mm FOV and optimally only with one lens. Or manual focus slr. Or an AF slr with a really dark screen.

Actually, it would be to use a digital camera which has Live View, focusing with a magnified Live View. In this way you would be focusing a magnified view of the image falling on the camera's sensor.

 

[porktaco]

ccd doesn't support live view? wtf?

 

[Mister E]

Low light: no a/f but RF works.

You've clearly not used any of Nikon's newest DSLRs, the D3, D3X, D3S and D700 can all focus in light so dark you cannot hand hold shots even at 1/8th @ ISO 6400. That blows a RF away. I love my rangefinders, but I don't have any illusions about their limitations, nor should anyone else. You can almost always manually focus on an SLR as well which would be just as useful as manual focus on a rangefinder, but more accurate.

 

[Pico]

I'd never buy digital Leica M simply because of so very many reports of rangefinder bugs. I hope Leica gets a clue and fixes this.

 

[Soothsayerman]

To me, critical focus is an application specific issue, not something that applies to every situation in photography because to me, that is not the nature of photography. Photography is ruled by esthetics, not absolutes.

To me, a rangefinder is more about freedom, convenience and expression, an slr more about technical accuracy, production value, and most importantly, versatility. Two different tools for two different things. The marketplace decided a while ago which was more appealing to the masses.

To me, the question is really this: given a specific application, where in the heirarchy of what is important to produce the best result is the sharpest focus possible the number one priority?

That puts it in perspective for me.

I then choose the tool I need to use. Most of the time, sharpness is not my number one concern by a long shot, but sometimes it is and then I use an SLR and a big heavy tripod. Otherwise, I'll take a light small rangefinder any day.

 

[35mmdelux]

most shooters use the chimp technique, no?