Distances to the M9 sensor...

[DFV]

Advertisement (gone after registration)

This one is for the techies... I recently shared a problem with this forum regarding problems with my Summilux 50 and my M9. I was certain the problem was the lens since on my Elmarit 21, Summicron 90 and WATE I have had no problems. Yet now, looking back in detail, these lenses are not 100% sharp on the left ether.

 

Recently, after the request of a member to compare the CV15 and the WATE I decided to get the CV 15 out of retirement and carry out the test. To my surprise the CV15 looked horrible on the same spots where I had problems on the LUX 50. The WATE did well on the other hand. So what could be the common denominator? It was obvious that it cold be the M9 by now. To me the main difference it the distance of the last element of the lens with the sensor of the camera. I know the CV15 is very close while the WATE is further away.

 

So if the problem is the sensor that is not aligned properly my question is this; Could it be that out of the lenses I mentioned before the two that are closest to the sensor are the CV15 and the Summilux 50? They are both closer than the CRON 90, the Elamarit 21 and the WATE?

 

Can anybody confirm this and solve the mystery?

[carlmuck]

I don't have the number handy, you can probably google it. The mount to film plane distance is the same on all M's. I doubt you can measure any differences yourself without special tooling. If you seriously think the camera is out of whack send it to Leica.

 

c.

[Advertisement]

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

[jaapv]

27.95 mm is the register length.

[delander †]

I think he means the rear element? In which case WATE is closer than the summilux but not by much and the CV15 is probably too close for comfort.

 

Jeff

[jaapv]

You could measure the protrusion of the rear element to the mount flange and deduct that from the register to get the exact distance to the sensor.

[adan]

Point 1: distance from the rear element really isn't a factor in focus questions - distance from the rear nodal point (or exit pupil) of the lens is what counts. The point light rays apparently come from, as seen through and modfied by the rear element(s).

 

Which is somewhere ahead of the actual rear element surface - the "somewhere" depending on the precise optical formula: whether the lens is a "straight" symmetrical design (the 50 'cron), a telephoto design (the 90), a retrofocus design (the 15, and probably the WATE - although as a zoom, its nodal point may vary with the chosen focal length). The 50 'lux ASPH takes a whole new approach, especially in the rear optical group, which is borrowed from the 35 Summilux ASPH (in broad outline but not necessarily exact construction) and which also floats.

 

Without access to Leica's or Cosina's computerized descriptions of the lenses, one is really up a creek in trying to measure anything meaningful.

 

Point 2: with a digital sensor, the image plane is NOT the front surface of the sensor, as it is with film* - it is behind the front surface. Probably at the surface of the microlenses (thus behind the IR/cover glass). So while the 27.95mm register distance applies, WHERE it applies to is a different question. Certainly one could not micrometer from the lens flange to the front of the cover glass, which is LEAST 0.8mm in front of the image plane (and perhaps an undetermined small amount more, if there is an adhesive layer between it and the microlenses).

 

*some might even quibble with that: since film coatings have some - microscopic - thickness to them, it has been theorized that Leica's actual chosen image plane is .00x mm BEHIND the film front surface surface to provide best average focus throughout the depth of the film gelatin.

 

However - IF the question is regarding the sensor being installed at a slant and causing a focus problem on one side only, then it is just a question of relative distance. The exact amount doesn't matter, so long as it is (or isn't) the same on each side of the frame. Lenses are also irrelevant.

[DFV]

Advertisement (gone after registration)

However - IF the question is regarding the sensor being installed at a slant and causing a focus problem on one side only, then it is just a question of relative distance. The exact amount doesn't matter, so long as it is (or isn't) the same on each side of the frame. Lenses are also irrelevant.

 

Well, What I have seen is that the problem magnifies itself on especially two lenses. The CV15 and the LUX 50. The WATE and CRON 90 did results that where acceptable at first site and if not looked at inn depth would have been hardly noticeable. That is what led me to believe that there was something wrong with the LUX.

 

I have sent both in to calibrate. I suspect by now that there is indeed nothing wrong with my LUX. However both together will give them the best way to see the problem. (I would have loved to send the CV15 as well just to see their faces. )

 

In any case, my question was if the CV15 and LUX 50 had something in common that made them much more sensitive to a misaligned sensor. As far as I have come to know is that the closer to where the image is captured the worse the sensor will handle the edges due to the angle where the light is coming from. Even with the angled micro-lenses of the sensor if you get any closer you will get in trouble. Somebody already mentioned that the CV15 was too close and therefore was not acceptable on digital while it was great on film.

[adan]

Just for the record, it's probably better to say the CV 15 is not acceptable on the M9 - it performs quite well on M8s. It may perform acceptably on an M10 someday. Mine actually performs quite well on the M9 - so long as I avoid color shots.

 

"As far as I have come to know is that the closer to where the image is captured the worse the sensor will handle the edges due to the angle where the light is coming from."

 

Generally correct - but the "closer" refers to the exit pupil, which is not the same thing as the physical back of the lens. And the "worse" effect on the image will not be blur or fuzziness (as I read your original post) but vignetting or red edges.

 

It is also not the ONLY factor.

 

So the fact that your 50 and 15 extend the closest to the sensor physically, and also may have been the troublesome lenses, is an interesting coincidence. But unless the exit pupils are also the two closest to the sensor - which just won't be true with a 15mm and a 50mm, much too different optically - nothing more than coincidence.

 

It IS possible that your 15 CV itself is not correctly centered (it has been known to happen). That can lead to fuzziness on one side but not the other. But has nothing to do with digital sensors. The off-center fuzziness would show up on film, too. Just an plain old manufacturing flaw.

[DFV]

Just for the record, it's probably better to say the CV 15 is not acceptable on the M9 - it performs quite well on M8s. It may perform acceptably on an M10 someday. Mine actually performs quite well on the M9 - so long as I avoid color shots.

 

Yes, See it that way too and hope that sensors over time will become more "forgiving". It is clear that with the results I got from the M9 with the CV15 color shots are out of the question.

 

 

It IS possible that your 15 CV itself is not correctly centered (it has been known to happen). That can lead to fuzziness on one side but not the other. But has nothing to do with digital sensors. The off-center fuzziness would show up on film, too. Just an plain old manufacturing flaw.

 

I do not rule out this possibility yet it would be a very BIG coincidence. I will first make sure the LUX and the M9 are perfect and then take it from there...

[Jerry_R]

It is clear that with the results I got from the M9 with the CV15 color shots are out of the question.

Why? Have you used CornerFix?

Recent version does solve color shift.

[DFV]

Why? Have you used CornerFix?

Recent version does solve color shift.

 

Am going to look into it tonight. Will tell you if and how it works on the M9 and CV15 shots.