interview with Dr. Kaufmann and CEO Kaltner on the SL

[CheshireCat]

Advertisement (gone after registration)

You said it is proven that shifted microlenses are useless. Would you care to teach us proven by whom and where? Can you link to the scientific research that proved it?

 

I said that shifted microlenses are not related to the "italian flag" issue.

 

This is a long but very interesting thread you should read, in which many good forum members are scientifically approaching the problem:

http://www.l-camera-forum.com/topic/153464-towards-an-explanation-of-the-italian-flag-phenomenon/

Spoiler alert  : at the end of the thread, you will find the explanation that makes most sense:

http://www.l-camera-forum.com/topic/153464-towards-an-explanation-of-the-italian-flag-phenomenon/?p=1758723

 

P.S. I also said that shifted microlenses are not a problem when using R lenses. This is quite easy to prove, but I don't have a M9 here.

[edwardkaraa]

I said that shifted microlenses are not related to the "italian flag" issue.

 

This is a long but very interesting thread you should read, in which many good forum members are scientifically approaching the problem:

http://www.l-camera-forum.com/topic/153464-towards-an-explanation-of-the-italian-flag-phenomenon/

Spoiler alert : at the end of the thread, you will find the explanation that makes most sense:

http://www.l-camera-forum.com/topic/153464-towards-an-explanation-of-the-italian-flag-phenomenon/?p=1758723

 

P.S. I also said that shifted microlenses are not a problem when using R lenses. This is quite easy to prove, but I don't have a M9 here.

You said:

 

No. It has been demonstrated that this is not the cause, and that shifted microlenses don't help at all.

Evidently I and you were talking about the color shading and you think shifted microlenses don't help to reduce or remove it. Leica engineers seem to disagree with this statement. I am referring to scientific articles written by mjh in LFI discussing this problem. I have the magazine somewhere in my archives but I would be grateful if he could chime in and clarify this issue.

 

PS. I just read the linked thread and I do remember having read it at that time. Luka is explaining how the shading occurs, but nowhere does it prove that shifted microlenses don't help. This is besides the fact that Sony introduced shifted microlenses in the A7RII along the BSI technology, and it seems to show a miraculous absence of color shading with M glass, even better than the M240, which in fact doesn't have the shifted microlenses.

[mjh]

Microlens shifting does help insofar as it allows photosites near the edges of the sensor to catch incoming light – non-shifted lenses would divert the rays to some other photosite nearby. But those rays still hit the photosite at a rather large incident angle, resulting in colour casts due to the complex and non-symmetrical structure of each pixel.

 

Starting with the M (Typ 240) Leica has adopted a different apprach, namely reducing the focal length of the microlenses. With the microlenses placed closer to the chip (which required reducing the height of the wiring on top of the silicon) there is less danger of rays hitting the wrong pixel. At the same time the incident angle of light on the chip is smaller and more uniform, reducing the problems with colour casts.

[edwardkaraa]

Thank you very much Michael!

[CheshireCat]

Microlens shifting does help insofar as it allows photosites near the edges of the sensor to catch incoming light – non-shifted lenses would divert the rays to some other photosite nearby. But those rays still hit the photosite at a rather large incident angle, resulting in colour casts due to the complex and non-symmetrical structure of each pixel.

 

Starting with the M (Typ 240) Leica has adopted a different apprach, namely reducing the focal length of the microlenses. With the microlenses placed closer to the chip (which required reducing the height of the wiring on top of the silicon) there is less danger of rays hitting the wrong pixel. At the same time the incident angle of light on the chip is smaller and more uniform, reducing the problems with colour casts.

 

Shifted microlenses serve to reduce optical crosstalks between pixels with light rays at very steep angles. This basically means that light supposed to hit one sensel, actually hits only that sensel, without leaking into neighboring ones, resulting in a sharper image.

In other words, the main purpose of shifted microlenses is a sharper image.

It is interesting to note that neighboring pixels have a different bayer-matrix color, however this has nothing to do with color vignetting.

 

There isn't much difference between shifted and non-shifted microlenses as far as color-vignetting is concerned:

http://www.l-camera-forum.com/topic/153464-towards-an-explanation-of-the-italian-flag-phenomenon/?p=1701091

 

And in fact, the cause of the color-vignetting has been explained in the post I cited above, and while microlenses could help a bit, they are not the solution (as clearly evident in the M9 and M240).

The only solution to color-vignetting is using a BSI sensor, because it avoids light interfering with CMOS wiring.

 

Unlike Edward, I am not sure the A7R2 uses shifted microlenses. Please cite source for this.

[edwardkaraa]

Sorry I'm not sure anymore if the mark 2 has offset microlenses. The mark 1 certainly did so it's possible I may have confused them. I was quite sure I read it somewhere though. In any case, this might prove that BSI is very effective against vignetting and color shading, more so than shifting microlenses, but there is plenty of evidence that the latter is much better than non offset ones on a FSI sensor.

[pico]

Advertisement (gone after registration)

Actually, my Biogon C 35 is a preferred lens on my Monochrom.

I love Biogons and use them for MF and LF. I cannot imagine why they would perform poorly for 24x36mm.

[jaapv]

Shifted microlenses serve to reduce optical crosstalks between pixels with light rays at very steep angles. This basically means that light supposed to hit one sensel, actually hits only that sensel, without leaking into neighboring ones, resulting in a sharper image.

In other words, the main purpose of shifted microlenses is a sharper image.

It is interesting to note that neighboring pixels have a different bayer-matrix color, however this has nothing to do with color vignetting.

 

There isn't much difference between shifted and non-shifted microlenses as far as color-vignetting is concerned:

http://www.l-camera-forum.com/topic/153464-towards-an-explanation-of-the-italian-flag-phenomenon/?p=1701091

 

And in fact, the cause of the color-vignetting has been explained in the post I cited above, and while microlenses could help a bit, they are not the solution (as clearly evident in the M9 and M240).

The only solution to color-vignetting is using a BSI sensor, because it avoids light interfering with CMOS wiring.

 

Unlike Edward, I am not sure the A7R2 uses shifted microlenses. Please cite source for this.

Confusion reigns. The reason: Colour shifting is not the same as colour vignetting. Colour shift (Italian Flag style) occurs because the sensor pixels are not a symmetrical design. Colour vignetting (Cyan) occurs because of the IR filter. Sensor vignetting in general occurs because the sensels have a limited acceptance angle. Crosstalk is in part to blame, but is certainly not the only issue.

[CheshireCat]

Confusion reigns.

 

You just added more confusion with your simplistic generalizations 

[t.kormicki]

In the contrary, listening to people who know better than you can be for your own benefit.

So you won't take it badly if I tell you that it's "on the contrary", not in. But as this is my first post, on to more important matters. Am I the only one (I have seen no mention anywhere in any of the comments regarding the future M) who read about the patent application from Leica for an EVF dual focusing patent on Mac Rumors. It kind of digitally mimics the range finder way of focusing with two overlapping images. This was about three or four weeks ago. 

[pop]

.... Am I the only one (I have seen no mention anywhere in any of the comments regarding the future M) who read about the patent application from Leica for an EVF dual focusing patent on Mac Rumors. It kind of digitally mimics the range finder way of focusing with two overlapping images. This was about three or four weeks ago. 

No, you are not: http://www.l-camera-forum.com/topic/248322-new-m-this-year-this-fall/?view=findpost&p=2913010

[CheshireCat]

Coming from a news background I would say that high MP cameras are a distinct disadvantage especially when wiring pics back to the office because of the time taken to transmit the files.

 

This is why professional high-MP cameras have optional low-MP RAW/JPEG formats you can use especially for this.

 

Your arguments about 16MP being enough for news are understandable, but there's not only news, and even for news, a high MP camera gives you more reach (i.e. "crop-zoom" and still get full 16MP).

 

The main reason I want a very high MP camera is I want to use light and compact primes, and then "crop-zoom" in post, if needed.

Seems instead that Leica is going backwards into the SLR + huge zoom lenses era, and I don't understand how they could ever compete with Canikon. Yes, the SL may actually be "the new R"... but we all remember how it ended.