Focus test question

[mszargar]

Hello forummates,

I have just bought a M6 classic, and have checked it all over according to the micro-bible of M buyers. All seems perfect. The camera has been used, I am sure of that, but the last three owners have apparently taken great care of it. You can't find signs of use on it without thoroughly searching...

The only thing I could not test when I bought the camera, obviously, was the focus. Now, I have a ground glass from a combo viewfinder (Zenit) and a 15x loupe. So, first things first, when I got home, I put the M6 on a tripod, focused on a lamp shade, put the camera on B, put the unpolished side of the ground glass on the film rail and the loupe on the backside (polished side) of the ground glass, to see whether the camera focuses accurately.

But it seems there is a difference between what the rangefinder sees as focused on what the lens sees as focused. This is a Voigtlander Nokton Classic 35 1.4 and I am focusing it on a shade 2-3 meters away. I have tried some Jupiter lenses I have for my Zorki, and they shows the same focus error, I supposed the mis-focus would be different as my Jupiters are not collimated for M yet. I tested the Jupiters on my Zorki, and there too, I get this difference. I have shot some rolls of film with the Zorki before, and sometimes used it for critical focusing, but I could never make sure whether the focus was accurate or there was a little bit of back focusing because the RF patch wasn't contrasty enough for me to trust my own focus.

I just wonder if I am doing the test the wrong way. I can send the camera for a CLA (it has been CLAd about 3 years ago in Germaby), but I want to make sure there are grounds for doing that. Also, it is important for me to collimate the Jupiter lenses I have (cheap, classic look), and for that, I need to master the art of rangefinder testing.

Please let me know if you have suggestions about my testing process. Please note that taking test shots is not really an option for now, because it does not allow for real-time lens adjustment.

 

[tunalegs]

Just to be sure, the ground side of the ground glass was laid across the film rails?

Have you shot a test roll with film yet?

 

[Dwig]

Just to be sure, the ground side of the ground glass was laid across the film rails?

Have you shot a test roll with film yet?

... and that's the "film rails" (the inner pair) and not the pressure plate rails (the outer pair). Still, this can lead to a very slight error since the film is not pressed against these rails but instead passes loosely through the channel left between these rails and the pressure plate. The channel width is controlled by the difference in the height of the two pairs of rails.

 

[baisao]

You might try running some film thru it first. I suggest Velvia 50 so you can get super fine detail.

 

[mszargar]

I had the exact same idea: I put a Velvia 50 in it and took it out for a stroll, but today the sky was rather covered in Montreal and I couldn't take many photos. On Monday we will have some sunshine, and I will take the camera out again. For now, I unloaded the film in the darkroom.

And about the ground glass, this is exactly what I did: The ground side laid on the rails that I have marked in the photo. I did it the same way Kim Coxon does it in his manuals.

 

[ottluuk]

Did you check horizontal alignment at infinity with a distant object (Moon, power lines a mile or two away)? The rangefinder images should coincide perfectly and the distance index on your lens should point at the infinity symbol. There may be a little focus travel past actual infinity depending on your lens but not much. If your infinity setting is off, close range will be off as well.

One reason for _small_ differences at close distances could be focus shift: the actual point of focus moves back and forth as you change the aperture. Depending on how your lens is optimized, it might focus accurately at f/2.8 but backfocus slightly when wide open. Both the 35 f/1.4 Nokton and the Sonnar-based Jupiters (J-8 and J-3) are known to exhibit focus shift. But these would be minor disagreements with the rangefinder - something that a lot of people don't even notice in real life use.

As for testing, the ground glass method should get you in the ball park - but that is still not exactly how the film will sit and record light. If your observed inconsistency was small (a couple of cm), it's quite likely that you can't see it on film. But if it is more like 20cm off at 2m, something will definitely need adjustment.

 

[ColSebastianMoran]

Accurate focus may be hard to see on a ground glass. I find there is nothing like shooting a test roll for checking these things.

Here's an example. Line up five boxes, 2 inch intervals, odd one in the middle. In this case, it looks like the actual focus plane is between the third and fourth items, a bit farther from camera than indicated (I tried to focus on the middle "200"). This is a Contax Aria at f/1.7, but same idea applies for a RF camera.

Fences can work well for this, if you mark the intended focal point.

And, all this is easier to see in the full resolution image.

 

[mszargar]

@ottluuk: horizontal alignment is perfect... This is actually the second time in my life I see such a clean M6 body (the first one being a collector's edition in the box). Your focus shift hypothesis makes a lot of sense to me: I have heard that the famous zeiss focus shift is most visible at middle distances, at about 3-4 meters. Tomorrow I will test again with real close ups at 0.7m and 1.0m and also at infinity.

@colonel: I did the test with 10 inch and 2 inch intervals (on velvia 50). I will have the film developed tomorrow and will report back in a couple of days.

 

[mszargar]

OK, the test film scans are in, and admittedly I haven't done an awesome job. In my defence: These are my first analog first shots, I did not have my tripod with me, we haven't had sunlight over here for more than a week, and the labscan I have access to really sucks quality-wise... Nonetheless, I think the photos can illustrate my point. All the pictures have been shot on Velvia 50, with my C/V Nokton 35 f/1.4 MC wide open. To see the photos, click on the title of each paragraph. Clicking on the photos zooms on them, so you can see the details.

1) Close focus. The boxes are arranged in two-inch intervals. (1) blue, (2) green, (3) red, (4), yellow. The green box is at the minimum focus distance (0.7), and is supposed to be in focus according to the RF of my M6. I had the horrible idea of putting a meter in the middle of the boxes, that you can see as a ghost.

2) Close focus with hand shake! 10 inches of distance between the boxes in the following order: (1) red, (2) greed, (3) yellow/blue, (4) white. Focus on green again @0.7m. Despite the handshake you can see that the green is in focus, similar to the last one. Although I should say I am not impressed by how much the VC lens can resolve.

3) Infinity focus. Either the lens is very soft at 1.4 or it is not really focused at infinity. Focus is supposed to be on the tilted roof of the furthest building.

4) Distant focus. This is the most importan shot IMHO. The lady in the photo is supposed to be in focus, but I don't think she is. The focus is rather on the edge of the green table which is about 10-15 inches away from her.

Voila, please let me know your opinion. Am I splitting hair here?

 

[ColSebastianMoran]

Sounds like a pretty good set of tests. If you intended to include the images, I didn't see them.

 

[mszargar]

Colonel, just click on the image titles (e.g., on Close Focus), and the photo opens in a new tab!

 

[ottluuk]

I think the lady in the #4 shot may have moved during the exposure - It's not really clear whether she is in focus or not. But what we are looking at is still hand-held, wide-open (?) stuff on ISO 50 slide film in murky light, apparently on the underexposed side and not scanned properly...

My advice, especially if the sunless days continue: buy some Ilford XP2 black and white film. It's C-41 process, like normal colour negative film: any lab can develop it. Set your ISO to 250 and expose as usual. Shoot things that you like. Use various apertures. Get 4x6 proofs from the lab. Choose photos that are interesting, either by themselves of for the focus check thing. Have a decent lab make 8x10/A4 or larger prints made from the interesting frames. Look at them. Do you like any of them? Can you tell that the focus is consistently off? Does it bother you at all?

A real-life test, if you will. Why XP2?

 

[gdi]

I have tested a number of lenses/cameras for focus issues and I can guarantee that you simply have to try to control the conditions.

The best way to do that is to use a simple test chart or even a ruler shot at about a 45 degree angle on tripod. Pick a hash mark near a inch/CM mark and focus on it very carefully; use a magnifier if you have it. You will be able to clearly see where the focus point is - whether it is in front or back of the target and by how much.

There is no need to shoot 50 speed film for this test, 400 will give you higher shutter speed and you will still be able to see the point of focus in a decent test. Here is a decent chart and instructions you can print out for testing.

http://focustestchart.com/focus21.pdf

Good luck

 

[ColSebastianMoran]

mszargar, I agree with you; these don't look as sharp as I would have expected.

If it's 50 speed, then tripod. Either that, bright light, or faster film.

And, are you getting good scans? (I should talk; my own test image above is a CVS scan.)

 

[mszargar]

Thanks everybody for the valuable feedback...

@Colonel: This week I will have my tripod, my lights and my darkroom access back. I will do a more decent test on B&W film and will make some 20x crops with the Besseler on Ilford Multigrade. I can't trust anybody more than myself when it comes to prints. I will probably have a second M6 body and a Summicron to compare as well. I will report back.

@ottluck: I could not agree more with you, except that your method does not help me discover my way towards collimating my russian lenses. I love the classic look these lenses give me for cheap (on Zorki, of course), and I can't wait to have them on a decent body. As you know, there are not many people that can help me collimate these lenses. I'd better do it myself. It's kind of a self-imposed challenge as well. I want to make sure of the accuracy of the body to use it as a tool for collimation.

@gdi: Nice idea, and nice link. I guess the A4 focus chart will be somehow small for testing lenses at infinity 😉, but I will certainly use it for close focus. I dont know why I didn't think of it first place.

 

[mszargar]

Reporting back on the focus test:

Did it again, much more rigorously. Used gdi's focus test target and other ordinary targets. Used ilford pan 100 and developed in ilfotec ddx. Used tripod mount this time, with proper lighting and 125-250 speed. Instead of scanning checked the focus with a pro-grade omega enlarger I have access to.

The results?

Nokton Classic 3.5 1.4 is generally soft wide open, that we knew. So, a part of softness is inherent to that aperture size. The lens is tak sharp at 0.7m and at infinity, both on negative and on ground glass. The lens definitely has a shift (ottluuk mentioned it earlier). This is a "feature" you get with this lens, as a part of its "classic" rendering. The shift is visible from 1 meter on, on the ground glass. 35mm being wide enough to provide some depth of field even at 1.4, the shift remains covered up to somewhere between 2-3 meters on negative, after which it becomes visible. People say on forums that the shift is visible between 3 and 5 meters, but I would say it is more like 2.5 to 7 meters.

By the way the ground glass method works perfectly if done carefully. I tested some Russian lenses with whacky focus using the same method and the results are always concordant with the results of negative test.