Retrofocus on 75 Summilux

[StuartR]

As I posted above, the sources are Leica's dof information in the lens information sheets. But in any case, you are right about .39 inches being a tiny amount. It is easy to move that much just pressing the shutter release or to screw it up by focusing on one point and reframing...the reframing could throw the focus off when it is that tiny...

 

[Hektor]

I would not be so sure about that, the noctilux has a closest focus distance of 1m, while for the 50mmf1.4 is 0.7m, maybe we should ask him about his sources and the formula.

I compared them at the same focussing distance, it is meaningless not to do so.

I'm not at all unsure, I'm certain. The 50 f1.4 asph has more dof than the 50 f1 with both wide open at the same distance.

 

[fgianni]

John, we really need to discuss this serious matter over a pint (or two)

 

[Hektor]

As I posted above, the sources are Leica's dof information in the lens information sheets. But in any case, you are right about .39 inches being a tiny amount.

I suggest you work it out for yourself.

It is easy to move that much just pressing the shutter release or to screw it up by focusing on one point and reframing...the reframing could throw the focus off when it is that tiny...

Thats the skill needed when using a fast lens wide open, - practise helps.

 

[fgianni]

the reframing could throw the focus off when it is that tiny...

Maybe is more correct to say: "the reframing WILL throw the focus off when it is that tiny"

Also if the DOF os 0.39 in on a film camera, it is 0.39/1.22 = 0.32 on in a RD-1,
proabably breathing will throw the focus off

 

[StuartR]

Hektor, why is it meaningless to compare them at different distances? The question was at the closest focus distance, not at 1m. The 50/1.4 WILL have less depth of field than the noctilux at .7m because the noctilux cannot be used at .7m! If it could, of course it would have less depth of field, but that was not the issue. If you are talking about small depth of field as an absolute property, the 50/1.4 is capable of producing thinner depth of field, but only because it focuses closer.

As for the wide open skills, I realize that you need to have good technique, but I am just saying that it is difficult. The way you phrased it was rather patronizing...

 

[J. Borger]

Now please back to practical sitiuations.
How often do you plan to shoot "only 1 eye in focus pictures at .7 meter" Fransesco.....
I have seen these type of shots over and over again to illustrate the possibilities of these ultra fast tele-lenses (the same with Canon 85 1.2 i used to own myself) . They pop up in threads like this but in practical day to day situations ..... how often do you plan to use the lens like this?
Like i said the 75 lux is reliable to focus at 1m. And even at .7m 1 out of 4-5 shots might be ok. Do you plan to use a steady tripod? Because the slightest body-movement wil ruin the shot even if the focus is spot on in a ruler test.


Han

 

[Huck Finn]

Here's a link, so you guys don't have to do DOF calculations by hand. (Yuck! ) There are others available if you do a search. And, yes, the 75/1.4 at minimum focus distance has the shallowest DOF of the three.

www.csgnetwork.com/depthoffieldcalc.html

Here's another link to Erwin Puts' discussion of Leica rangefinder accuracy, which applies to other cameras as well.

www.imx.nl/photosite/leica/technics/rfaccuracy.html

In this discussion, Erwin list the minimum effective baselength required for the 75 Summilux to be 40.2, which places the R-D1 just below this point. In his book, "Leica Lens Compendium", however, Erwin also offers a second more conservative number of 52.3 because not every focusing situation allows the rangefinder to align broken lines, which is the easiest way for the human eye to discriminate. So, there are some focusing situations for which the R-D1 will be woefully inadequate for this task.

I think that Stuart did an excellent job of describing the focus limitations of the R-D1. I especially agree with his observation that the baseline's physical length is more important than the magnification. In addition to the pure mechanical advantage, it also enhances the tolerances of the instrument. For example, a rangefinder that is off by 1 mm on a camera with a baseline of 37 mm like the R-D1 introduces an error of 2.7%. This same misalignment on a longer RF baseline of 74 mm (ZI) would only be an error of only 1.35%, or half as significant.

However, because of the consistent backfocus pattern to the errors, it sounds most like a focus shift problem to me, inherent in the 25 year old lens design, & which Leica corrected on the new 75 Summicron. But if this is the biggest problem with the lens, it's hardly significant, unless you're going to be shooting most of your photos wide open & at the minimum focus distance. Very unlikely.

Enjoy the lens!
Huck

 

[Sonnar Brian]

Skipped all the DOF stuff...

Your Jupiter-9 also focussed 2~3 inches behind what the RF of your RD-1 indicated, right? At 36" a lens made to the Contax standard (the J-9) will focus ~1" behind what the RF indicates. Your J-9 is off by a bit more than that, but it could have been manufacturing tolerance. The Leica lens is off by 1~2". That is two lenses that focus behind the RF. Are you sure the RF is exactly correct?

 

[CJP6008]

Have a read of this - it deals with RF accuracy and minimum base lengths.

http://www.imx.nl/photosite/leica/technics/rfaccuracy.html

Unless I am much mistaken, your camera has an effective base length of 37mm, short of the 40.2 that Puts recommends (and his calcs are a compromise - for better accuracy a longer base still is needed.) I asked him about his once re a 90mm f2 on a CV R3A. He said that it would work but that I would have to accept that some shots wide open would not be accurately focused. Even if all the cams etc are totally accurate (and this will help) it still will be an issue (if Puts is to be believed).

 

[Huck Finn]

Skipped all the DOF stuff...

Your Jupiter-9 also focussed 2~3 inches behind what the RF of your RD-1 indicated, right? At 36" a lens made to the Contax standard (the J-9) will focus ~1" behind what the RF indicates. Your J-9 is off by a bit more than that, but it could have been manufacturing tolerance. The Leica lens is off by 1~2". That is two lenses that focus behind the RF. Are you sure the RF is exactly correct?

Brian, my compliments on your attention to detail in your reading skills. Where did you find the Jupiter 9 comment? I missed that one.

Huck

 

[fgianni]

Skipped all the DOF stuff...

Your Jupiter-9 also focussed 2~3 inches behind what the RF of your RD-1 indicated, right? At 36" a lens made to the Contax standard (the J-9) will focus ~1" behind what the RF indicates. Your J-9 is off by a bit more than that, but it could have been manufacturing tolerance. The Leica lens is off by 1~2". That is two lenses that focus behind the RF. Are you sure the RF is exactly correct?

Well the Jupiter was FAR-OFF to the point of being unusable (as I said 2-3 in) while the Leica is only 1/2 to 1 inch so that some pictures even taken wide open at the shortest focus are useable, no other lens has any problem, and I guess the Nokton f1.4 at 0.7m must have a similar DOF to the J9 at 1m.

To be fair what I am seeing with the 75 LUX should be considered more a limitation of the camera-lens combo than a problem.

 

[jaapv]

My 20D autofocus must be brilliant! It consistently gets spot on shots with my 150mm at f2.8, and at 15 in distance, I wonder what's the DOF in that case, probably less than 0.4 inches, but I don't have the formula.

Not really surprising. Not only has Canon really made an effort after all the flak they got on the 10D AF (which,btw, is quite good really,in my experience), the focussing system in a SLR is principially different from a RF. A RF will decrease in accuracy as the focal length gets longer, a SLR will increase in accuracy, the turning point being, in the case of a Leica M 135 mm.This is the scientific reason Leica still makes R and M camera's side by side(apart from customer demand, I mean). I guess the turning point (my math is not up to calculating this exactly) on a RD1 would be about 90 mm effective, this being 60 mm real focal length or maybe 75 considering the larger DOF because of the crop-factor. So that bears out the impression of being at the limit of the design.

 

[ezio gallino]

cron 2 is better, lux is heavy and only on theather could do his part. but cost a lot of moneyì

 

[fgianni]

Not really surprising. Not only has Canon really made an effort after all the flak they got on the 10D AF (which,btw, is quite good really,in my experience)

I never understood the fuss around the 10D AF, for me it was very good, the reason I moved to the 20D are ETTL2 (Really ETTL1 was rubbish) and fast startup time, the 2 extra Mpixel may be nice but unless you print bigger than A4 6Mpixel are good enough.

 

[vizioneer]

I'd really like to see a comparison of lux and cron, .75m, f/2. The 75 lux is high priced new, but popflash.com has them for cheaper. Also, the used market holds a sufficient number of them for around $1600 it seems. It's a sweet lens but not everyone's cup o' tea, although the imagery is unique.

Francesco, I will let you know how my sample is when it arrives, it should be here sometime next week!

 

[fgianni]

cron 2 is better, lux is heavy and only on theather could do his part. but cost a lot of moneyì

I like taking pictures in (low) available light, one extra stop is nice for me, even if probably the cron is sharper.

 

[Hektor]

Most of you are missing the point that Fgianni is making, and the thinking has become muddy.

He says he has repeated the focussing effort many times and always gets the same result, i.e. approx 2cm behind the subject.

This is not indicative of an inadequate rangefinder, (i.e. too short), it simply suggests a rangefinder which is incorrectly adjusted, or a lens which is similarly maladjusted.

This is precisely what Fgianni correctly surmised and asked for guidance. His 75 can be checked by comparison with a known good example and similarly so can his RD-1.

Much of the other comments are not relevant, and several are inaccurate, e.g. there is no focus shift on the 75F1.4. If you find that difficult to accept I suggest you try it for yourself.

The question of Dof is a matter of precise mathematical calculation, not conjecture, and depends on making assumptions, amongst which are focussing distance, aperture, and reproduction ratio.

That it is why it is irrelevant to make comparisons, unless similar assumptions are made.

Fgianni knows already that his combination produces a repeatable error, and he has only three options........

1. put up with it.

2 change the lens.

3. change the camera.

Since he had an option to return the lens within the next few days, he was simply wondering whether it was most likely to be the lens or the camera at fault.

I think the camera is most likely in need of adjustment, and that is fairly easily tested with another 75F1.4 (a 75F2 is not good enough) or a 90F2. It's not rocket science, (which is probably what Fgianni does for a living )

 

[vizioneer]

Francesco- yes I like the luxes as well, and the extra stop can of course result in a sharper photo in the real-world than a slower shutter speed cron.

 

[Huck Finn]

Most of you are missing the point that Fgianni is making, and the thinking has become muddy.

He says he has repeated the focussing effort many times and always gets the same result, i.e. approx 2cm behind the subject.

This is not indicative of an inadequate rangefinder, (i.e. too short), it simply suggests a rangefinder which is incorrectly adjusted, or a lens which is similarly maladjusted.

This is precisely what Fgianni correctly surmised and asked for guidance. His 75 can be checked by comparison with a known good example and similarly so can his RD-1.

Much of the other comments are not relevant, and several are inaccurate, e.g. there is no focus shift on the 75F1.4. If you find that difficult to accept I suggest you try it for yourself.

The question of Dof is a matter of precise mathematical calculation, not conjecture, and depends on making assumptions, amongst which are focussing distance, aperture, and reproduction ratio.

That it is why it is irrelevant to make comparisons, unless similar assumptions are made.

Fgianni knows already that his combination produces a repeatable error, and he has only three options........

1. put up with it.

2 change the lens.

3. change the camera.

Since he had an option to return the lens within the next few days, he was simply wondering whether it was most likely to be the lens or the camera at fault.

I think the camera is most likely in need of adjustment, and that is fairly easily tested with another 75F1.4 (a 75F2 is not good enough) or a 90F2. It's not rocket science, (which is probably what Fgianni does for a living )

Hektor, well said & well thought out.

Regarding the question of focus shift which I raised, you are correct & I was wrong. Although a focus shift would give results such as those described & would occur at close distance, it would NOT occur wide open & I forgot about that. Focus shift only occurs when the lens is stopped down. According to Erwin Puts, all longer high speed Leica lenses are susceptible to focus shift at closer distances, but it would not be the issue in this case. Erwin wrote this comment before the introduction of the new 75/2 Summicron, which I would assume is not susceptible to this problem because its floating elements should correct for this.

I agree with your conclusion that the problem is most likely in the camera's rangefinder. When you say: "This is not indicative of an inadequate rangefinder (i.e. too short), it simply suggests a rangefinder which is incorrectly adjusted . . .", in my opinion, you underestimate the problem. Nothing simple about it. Before we know whether the RF is incorrectly adjusted, we need to know the tolerances that are spec'd. It may well be that the RF is correctly adjusted according to the manufactureer's spec's. However, the short baseline on this RF requires such minute fine tuning to meet the demands of focusing this lens at its widest & closest, that you are likely asking the impossible. There may well be some R-D1's that meet these focusing demands, but I woudl think that it's largely a matter of luck whether the calibration fell at just the right point. Another sample of the camera, also calibrated within spec's may fall just outside the requirements for the lens under these conditions.

In his "reveiw of fast lenses for the R-D1" on www.luminous-landscape.com, Sean Reid's tests also found the focusing precision of the R-D1 to be inadequate to reliably focus the 75/1.4. Asking this rangefinder to do this is asking it to do something that it was just not designed to do. Although the results in this lens test found that the lens consistently focused behind the subject as you said, it was still not otherwise consistent in its focus point as the "8" was sometimes covered by DOF & in focus and sometimes not as the focus point was set a little further back. Erwin's mathematical calculations make it clear that even under the best of circumstances, a longer effective baselength is needed for this task than the one on the R-D1.

Huck