Some Thoughts

[G Radda]

Advertisement (gone after registration)

As stated by Leica, thicker cover glass would degrade other performance factors. So that is probably out as a way to improve IR rejection. One possible way to improve IR rejection would be to use a multicoat on the 0.5mm cover, (similar to anti-reflection coatings technology which is well known). This would not add thickness but would require a replacement of the cover.

 

My background is in nuclear physics and optical sciences.

[alexr]

Having just read Jim's fantastically useful post, if i understand it right, there might be a solution. Change the characteristics of the red pixels bandpass filter in the bayer pattern, then you can keep the same thickness of IR glass filter which appears to be needed for IQ . Could this work? Ok it needs a new sensor, but we knew that. Kodak...kodak is it all your fault?

[G Radda]

Problem is not with Kodak, they just make sensors. problem is that Leica underestimated the effect of additional IR reaching the CCD when they opted for a thin cover glass of 0.5mm. They need to add a multilayer coating that targets the IR to the cover.

[alexr]

Understood. I thought Kodak made the sensor and the glass IR fillter which is i believe glued on it.

From your knowledge would you say the multilayer coating can be done? Will it be tough enough? And further, if a multicoating is all it needs (not an engineer, bear with me) how can Kodak-Leica not have thought of it? Or does the multicoating scatter the IR and introduce other problems?

Again, sorry about my ignorance, should have studied maths not fine arts...

[G Radda]

Hi Alex,

Multicoating can be very tough, same process as is on lens elements and filters, even plastic lens eyeglasses. Multicoating would essentially reflect the IR so they would have to be sure that there were no other problems, but the lens elements should not pose a problem since they are transparent to IR anyway. All other cameras I know of do not need to constrain the cover thickness to 0.5mm so they have plenty of room for a traditional bulk absorptive IR filter as part of the antialiasing filter. Finally, multicoatings are more expensive to design and apply and few if any other customers would require this from Kodak. I am sure that there is already an anti reflective coating on the cover, so maybe the IR blocking can be integrated in the layer stack. Some of these coatings can have over 50 different layers.

George

[alexr]

problem is that Leica underestimated the effect of additional IR reaching the CCD when they opted for a thin cover glass of 0.5mm.

I know it is an obsolete question right now, but don't you find it strange that they did not notice the problem. I mean a couple of days after the m8 gets into customers hands the world is full of pictures showing the problem. How could have they not know?

[Guest rubidium]

Advertisement (gone after registration)

Jim, that's very interesting. Does that mean that one could take a DNG and simply work the red channel on import to C1 and solve the issue -- ie., hold your real reds while removing them from where they are appearing because of the lack if an IR filter on the sensor?

 

I don't see any way to selectively manipulate red channel pixel readout values to preserve "real reds" while removing "false reds." By allowing IR and visible red to both pass through the red Bayer filters onto a wideband photodetector is analogous to allowing Evian and Perrier to mix together in the same glass. There's no going back. Once somebody hands you the glass of water you either have to drink it or discard it; there's no conceivable way of drinking the part of the mixture you prefer and throwing away the pary you don't.

[G Radda]

PS I think Kyocera of Japan makes the cover filters.

[G Radda]

James,

Exactly right. The sensor has no way of telling if the red sensor energy was from IR or true red wavelengths. Once the IR is through to the sensor, all further bets are off. I for one do want to try IR photography with the M8. I think with a little testing, the proper IR focus point can be found for various lenses. I still have some old Nikon lenses with the IR index on the barrel of the lens.

George