Lens detection affraid.

[fabianoliver]

Advertisement (gone after registration)

Hi all.

I'm just quasy duplicating a thread i wrote 30 min ago from M lenses, but i think, is not in the same league.

Here is more related to the 240 lens detection, and how drastic can be have it turned it off.

I've the MP240 from a year ago, and a 28/2.8 elmarit asph V2 too.

Today developing some color pictures ( i do mostly B&W) i've seen a magenta "zone" in the right part of the frame.

AB testing the 28 with a pentax takumar50/1,4, only the 28 showed the problem.

I thought the lens was bad!

But, i remembered that some month ago i turned off lens detection, because it was a pain in the ass when changing from the 28 to a summarit, to a sumicron R, to a takumar.

Point is that i turned it on automatic, and the problem was gone.

Why it has a so drastic consequence to have it off? some can explained to me?

I know, vignetting, ok, lets do some thing to correct it in the firmware, BUT, magenta!!

Uff

Here 2 pictures with lens detection off and the 28elmarit asph V2, and the one with LD on.

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[pippy]

My understanding (FWIW) is that it is a Colour-Fringing Aberration due to the acuteness of the angle at which light hits the sensor's microlenses at edge-of-frame. The angle is so tight that the 'wrong' information is being captured by the pixel underneath those particular microlenses.

Having the lens-detection switched to 'auto' allows the camera's brain to apply in-built corrections.

To see the effect 'in full' you might try fitting a 21mm f4.0 Super Angulon......😸......

Philip.

Edited August 17, 2024 by pippy

[jaapv]

Yes  - known since 2006 for M cameras... Just switch on lens recognition, or use Flat Field correction. The explanation is simple.  Some and most vintage rangefinder lenses have steep incidence angles which cannot be corrected completely by microlenses design. 

[jaapv]

Please do not start multiple threads on the same subject.  

[fabianoliver]

thanks for the explanation.

i'll keep reading

i wanted to erase or move to this forum the thread on M lenses but the system doesn't allows me.

thanks

[jaapv]

Already done. 

[adan]

Advertisement (gone after registration)

On 8/16/2024 at 6:54 PM, fabianoliver said:

But, i remembered that some month ago i turned off lens detection, because it was a pain in the ass when changing from the 28 to a summarit, to a sumicron R, to a takumar.

Point is that i turned it on automatic, and the problem was gone.

Why it has a so drastic consequence to have it off? some can explained to me?

I know, vignetting, ok, lets do some thing to correct it in the firmware, BUT, magenta!!

Why? Because -

1) the Bayer array of colored filters over the pixels is not perfectly symmetrical at a microscopic scale. There is a slight bias - red to one side, blue to the other. See image of four pixels below. And multiply by 6 million of these. tiled next to one another like a checkerboard, for an M420 sensor.

Keep in mind that silicon itself responds to any photon that hits it, regardless of "color" or wavelength. It has to depend on the light being filtered by color, and then correctly knowing which pixels are receiving only green, or red, or blue light.

(And yes, color digital camera firmware actually does specifically identify or map the "color detection" of every individual pixel on the sensor!   "The pixel 1287 pixels from the left side, on the 2013th row of pixels, detects green!" Times 24 million. Only way that color digital photography can work, excepting Foveon technology (which has its own problems).

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

2) For most lenses, that asymmetry does not matter, but with wider lenses, where the light hits the pixels at a shallow angle, and a progressively-more shallow angle approaching the edges of the sensor,  one color of light can "leak into" a neighboring pixel, and be detected as the "wrong" color.

In this example below (left side), the leaking red light will be counted as green light by the camera's circuitry, since it ends up landing on a green-filtered bit of silicon.

But the same can happen in reverse - green light counted as red or blue light, producing "excess red/blue, which equals magenta.

That effect will also be directionally biased, and build up over several hundred Bayer arrays as one gets closer to the sensor/picture edge.

Resulting in the fading in of a colored stain along the edge of the image.

As this image shows, the M (type 240) has an improved sensor - less of this effect than the M8/M9.

However, "improved" does not equal "perfect."

Therefore the M 240 firmware and CPU still have to "erase" some color stain from every picture made with wide-angle lenses, with processing of the image according to the lens used, in between the sensor and the SD card. In a very distinctive pattern for each lens type and focal-length.

If you turn off lens detection, the camera can no longer make that correction - and you get the results shown in your samples. An uncorrected pink stain (usually) on the right side of the picture. And sometimes a corresponding green/cyan stain on the opposite side.

It is sometimes known as "Italian Flag Syndrome 🇮🇹." 😉

Edited August 18, 2024 by adan

[fabianoliver]

Adan, thanks a lot for your post!

and the time you take it.

It helps me a lot to understand!