A front-focusing 50mm f/1.5

[loccdor]

I want to get more use out of my Jupiter-3 50mm f/1.5 lens. I know the ideal is to get it shimmed by a professional, let's leave that aside for a moment.

What I learned about it so far:

1) The lens at f/1.5 front-focuses by approximately 5 cm at the minimum focus distance of 0.9 m with the Kiev 4 rangefinder.
2) The lens at approximately f/4-f/5.6 is focusing accurately (focus shift).
3) I mostly only want to use this lens wide open, for its character. I have a Jupiter-8 when I need a more technically perfect shot, which has no focusing issues.
4) Each f-stop down shifts the focus at the MFD away by approximately 1.5 cm.

The markings for a 5 cm focus shift align with the DoF scale mark at f/4.

I'm thinking I'll focus in the rangefinder, look down at the lens and focus out until the f/4 mark. Then things should be accurate for a wide open shot. Kind of like how you'd adjust for the infrared mark.

That would imply, by these pictures, that when the camera is focused on infinity wide open, it's actually at 15 meters.

Does all this seem to be sound logic?

 

[Hassasin]

And how did you determine the 5 cm focus "shift" at minimum distance ?

 

[brbo]

Yes, you can learn how to "focus out" on the lens or perform "chicken head" move. I do the latter with my Sonnar that is optimized for f2.8-f4.

 

[Rumbo181]

Just an idea.

I have a number of Soviet lenses with LTM mount. When I use one of them on a Leica and focus with the rangefinder on a target at 2 m, I might read 1.8 m on the lens scale — just as an example. If I repeated the process at different distances, I could get a correction curve for the lens. Actually the error is always predictable.

For instance, imagine that at short distances you get a −10 % error. Then:

1. Focus using the rangefinder.
2. Read the distance indicated on the lens.
3. Refocus the lens to the previous reading +10 %.

This way you can still use the camera’s rangefinder without making any physical modification to the lens.

 

[loccdor]

If I repeated the process at different distances, I could get a correction curve for the lens.

Thanks. Since the distance markings on the lens are already on a "curve", wouldn't determining a constant offset mark (as exists on infrared lenses) be the same as adding a % to the distance?

how did you determine the 5 cm focus "shift" at minimum distance ?

Test shot wide open at minimum focus distance. The lens focused on something approximately 5 cm closer. My other lenses with this rangefinder focus well.

 

[Rumbo181]

Thanks. Since the distance markings on the lens are already on a "curve", wouldn't determining a constant offset mark (as exists on infrared lenses) be the same as adding a % to the distance?

Actually, I should thank you for your post.

At first reading, I took for granted that focus shift was just a generic way of referring to a mismatch between rangefinder and lens information.

Then I read it more carefully and realized that “focus shift” is actually the name of a known and documented effect that happens with some fast lenses at wide apertures, shifting the focus as you stop them down. I certainly didn’t know that — so, thank you.

Furthermore, I realized that I should read more carefully. This is not a language problem, but rather a too-much-Internet problem. Thank you again.

What I proposed was the centuries-old method of recording the difference between an observed measurement and the real one, keeping that difference in a table so it can be used to find the true value from the observed one. If I understood your initial post correctly, this is essentially what you did.

I used a constant difference — a percentage — just as a way to illustrate the method, that’s all.

And now, two considerations:

I suppose that focus shift depends on the difference in f-stops (which could be fixed, let’s say from f/1.5 to f/8), but it’s not certain that it has a fixed relationship with the lens–subject distance. I believe it diminishes with distance.

Actually, I did the classic thing: I asked AI, and it brought back this little table:

Cause	Type of change	Δv in image plane	Δu in object space	Fixed mark valid?
Infrared light	Difference in wavelength (constant for λ≈800 nm)	Constant	Variable but proportional	Yes
Focus shift	Spherical aberration (dependent on distance and aperture)	Approximately constant, but varies with u	Tends to 0 at infinity	No

Δv is the variation in the image plane (the negative), and Δu is the variation of focus in the lens–subject distance.

To summarize: a fixed mark is not really valid (hence the oddity of your last two pictures), and the effect of focus shift becomes less worrying as the distance increases. I suppose that at around 5 m it would be hardly noticeable — though of course, it’s always possible to do a test to find the distance at which the effect can be safely ignored.

I hope this helps.

By the way, I have two Kievs with their Jupiter 8. Is Jupiter 3 a good complement?