Why coded lenses and IR?

[robsteve]

Advertisement (gone after registration)

I think I see where Leica is going with coded lenses and IR filtering. The increased IR filtering itroduces cyan in the corners of wide lenses like a vignette. The only way to correct this in camera is to map the color and density vignette of each lens when used with either a IR filter on the lens, or a better IR filter on the sensor. Below I am using a DMR as an example. It must have a decent IR filter on the sensor, because it does not have the magenta problems, but you can see the corners of a wide 19mm go cyan without the in camera ROM corrections.

 

From an earlier thread on ROM lenses and the DMR doing image correction, here are a couple examples illustrating what I mean.

 

Here is a 19mm shot done with the DMR and the ROM contacts disabled. This is a shot of my light table which is neither clean nor evenly illuminated. Just look for the corner vignette, not top to bottom even illumination. You can notice the corner vignetting of the lens and that it is cyan. I white balanced each image using the center of the image.

 

 

Here it is with the ROM enabled and the Camera doing the vignetting correction. This correction is even visible on the DMR display, so it is being done in the camera. Notice the corners are not going Cyan.

 

 

This is the first image that had no ROM induced correction. I applied some vignette correction in the DNG conversion. Look how the corners go cyan, while the in camera ROM corrected version doesn't have the corners going this way.

 

[Advertisement]

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

[Joseph S. Wisniewski]

Rob, thanks greatly for sharing those interesting examples.

 

I think I see where Leica is going with coded lenses and IR filtering. The increased IR filtering itroduces cyan in the corners of wide lenses like a vignette. The only way to correct this in camera is to map the color and density vignette of each lens when used with either a IR filter on the lens, or a better IR filter on the sensor.

Quite likely. I mentioned that a few days ago. I just hope they have a manual setting item on the menu, in addition to using coded lenses.

Below I am using a DMR as an example. It must have a decent IR filter on the sensor, because it does not have the magenta problems,

That's true of most modern DSLRs. Nikon only uses dichroic filtering on their higher level DSLRs (D2X, D2Hs, D200, and D80) but Canon uses it across the board, a few million DLSRs currently in the field.

but you can see the corners of a wide 19mm go cyan without the in camera ROM corrections.

 

From an earlier thread on ROM lenses and the DMR doing image correction, here are a couple examples illustrating what I mean.

 

Here is a 19mm shot done with the DMR and the ROM contacts disabled. This is a shot of my light table which is neither clean nor evenly illuminated. Just look for the corner vignette, not top to bottom even illumination. You can notice the corner vignetting of the lens and that it is cyan. I white balanced each image using the center of the image.

{/QUOTE]

I am surprised the Modul-R had such a cyan corner problem due to its internal filter. Being a DSLR, it shouldn't be dealing with large angles, and a dichroic filter should be very well behaved. You shouldn't see exit pupils closer than 52mm to the sensor, and that means light at most 16 degrees from perpendicular. The only way to see a cyan tint at 16 degrees is if the DMR sensor dichroic was tuned right on the edge of IR, or even slightly into the visible light, say 680nm. Then 16 degrees would bring it down to 650nm, which would be just enough to cause cyan corners.

 

Although I didn't take into consideration your f2.8 lens (which could add another 10 degrees, to bring it up to 26 degrees from perpendicular). That wouldn't be across the entire exit pupil, only the outermost part, so the effect should be much less...

Here it is with the ROM enabled and the Camera doing the vignetting correction. This correction is even visible on the DMR display, so it is being done in the camera. Notice the corners are not going Cyan.

 

This is the first image that had no ROM induced correction. I applied some vignette correction in the DNG conversion. Look how the corners go cyan, while the in camera ROM corrected version doesn't have the corners going this way.

 

I wonder if the camera is applying the correction to the DNG file, or if it's putting the color correction parameters into the DNG and the raw processor is picking up on them and doing the right thing...

[robsteve]

I did these conversions in Lightroom which is not the sofware provided by Leica for the DMR, so I suspect it is done in camera.

 

The in camera preview on the DMR's display also showed the difference in vignetting. I didn't shoot in JPEG mode, which would probably definately prove it was done in the camera.