M8 on TOP

[jmr]

Advertisement (gone after registration)

All,

 

noticed this on The Online Photographer.

 

http://theonlinephotographer.blogspot.com/2007/02/leica-m8-introduction.html

 

Although it adds little to the sum of human knowledge, it may be of interest.

[Iron Flatline]

Pretty good article, thanks for linking.

[peterv]

Thanks for the link, interesting article. Scott is right to the point.

[barjohn]

Overall I agree with his assessment; however it ignores the responsibility that Kodak may share. The Kodak specifications for the sensor would not lead one to think that the camera had an IR problem. Their graphs show a veryy sharp cut off above 700nm and with the exception of red even above 650nm (red is at 0.1 QE. The graph on the cover glass shows a bandpass between 400 and 600 nm with very sharp attenuation on either side of the curve. Thus Leica engineers may have believed those specifications and spent time looking elsewhere for the problem. The link is here KODAK Image Sensor Solutions - KAF-10500 for the sensor specifications. The only other comment I would make is that it is generally bad business on a high tech product to relie on being just a specification producer and integrator. Without inherent in-house capbility you will get burned sooner or later.

[Vivek Iyer]

John, Kodak reports Q.E (sensitivity) for the green pixels and ignores the blue and red which cause the magenta madness. Flawed.

 

Why would anyone (with active brain cells) who wants to use the chip for any photographic application take that at its face value?

[scott kirkpatrick]

Thanks for looking at the article. I put quite a bit more stuff in the presentation that is linked to at the end, but I was already way too long for Mike Johnson's blog. I think the Kodak specs do show that IR will be a problem, since there is significant leakage above 700 nm, where there would be essentially no response in a fully corrected camera. (Look at the comparable numbers for the DMR's chip.) From what I could gather, Leica knew they had a problem, but didn't do the work to find out how serious it would be. And Kodak, who had seen this before in their camera (but not in their chips) didn't say anything.

 

On the last point, I don't think Leica had much choice. They had let their only digital people go after the S-1 failed (according to a forum member who worked on that product). Let's hope they build up firmware and conversion skills in house over time. Probably firmware is most critical.

 

scott

[carstenw]

Advertisement (gone after registration)

Hey Scott, noticed your name there as I was reading. Nice summary and rebuttal. I will read the full presentation next. Nice to see a picture of you too. Most people here remain anonymous, and I personally prefer to know how people look. I am curious

[stuny]

Carsten -

 

Some irony there -- wanting to know what we look like but using an avatar which is probably not quite how you look ;-)

[carstenw]

Haha, touché I have posted mirror shots of myself before though, to be sure. Here is a recent one:

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!

[scott kirkpatrick]

I posted that shot that Mike Johnson used in the same "8M" thread here on L-C-U of mirror shots that Carsten posted in about a month ago. I like it because I found a corner mirror and was able to actually get an "M8" shot next to an "8M".

 

scott

[sdai]

It's quite interesting to see how your pen pals' real looks are different from the images in your imagination ... the forum certainly could be friendlier if folks get to know each other "in person".

 

Hi, Carsten ... hi, Scott ...

[barjohn]

Vivek

 

The chart showed the QE for all three colors. I'm not sure which chart you were looking at. Look at the attched charts. Maybe I am not reading them correctly. Pre production samples may have met these specifications. Production runs may be different though this document is dated December 06.

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!

[J_Brittenson]

Their graphs show a veryy sharp cut off above 700nm

Wavelength Transmission

710-750nm < 10%

 

10% is about a 3 1/3 stop attenuation. That's not much. 10 stops would be much better. That's a little under 0.01%, nothing you'd discern on a graph. 14 stops would make it < 1 significant bit.

 

If the line doesn't hug the bottom axis on a linear scale, then it's not perceptibly attenuated... If the scale were logarithmic, showing attenuation in stops, then it would be quite acceptable.

[J_Brittenson]

And what the heck is up with a graph that has a value range of 0-1 indicating transmission in %? 0-1%? Or to 0-1x? But then why is there a 1.2 tickmark, clearly it can't transmit more than 1x (100%).

[barjohn]

J,

I think I follow what you are saying but look at the QE graph above it. The lines are hugging the zero axis (except red). I haven't looked at other sensor curves but to my untrained eye that would appear to have very low IR sensitivity and only in one channel.

[Vivek Iyer]

John, Thanks! Did not phrase it quite right. I have looked at the same pdf brochure and even got an email response from one of Kodak's people claiming no UV or IR response.

 

A photog who visited the photokina, at my request, did enquire with Leica reps about UV/IR response of M8. The reps had no clue about IR or UV.

[sdai]

I haven't looked at other sensor curves but to my untrained eye that would appear to have very low IR sensitivity and only in one channel.

I've checked most of their other sensors' spec sheets and they were all showing the absolute QE values in the charts - which are for the naked sensors I guess ... the one for KAF10500 was measured with the BS-7 coverglass, it appears to me that the KAF100500 was specifically designed for use with the BS-7 for a reason. I do believe that the KAF-10500 sensor was jointly designed by both Kodak and Leica ... remember how many times Leica has repeated that the choice was a compromise in order to achieve best available image quality? the naked sensor is probably very prone to aberration artifacts due to the variations in coverglass thickness and maybe IR coating thickness. They could have tested the whole thing together and found this might just be the best compromise they could came up with.

[rvaubel]

I've checked most of their other sensors' spec sheets and they were all showing the absolute QE values in the charts - which are for the naked sensors I guess ... the one for KAF10500 was measured with the BS-7 coverglass, it appears to me that the KAF100500 was specifically designed for use with the BS-7 for a reason. I do believe that the KAF-10500 sensor was jointly designed by both Kodak and Leica ... remember how many times Leica has repeated that the choice was a compromise in order to achieve best available image quality? the naked sensor is probably very prone to aberration artifacts due to the variations in coverglass thickness and maybe IR coating thickness. They could have tested the whole thing together and found this might just be the best compromise they could came up with.

 

I don't know about that. First of all BS-7 is a pretty non-aggressive absorbtive filter. Hoya and some others have developed some more modern formulas that may have been more effective. I realize that the use of a reflective cut filter may have been ruled out because of the low incident angle problem but monday might quarterbacking would indicate that maybe they should have looked a little harder. I mean if the alternative to one filter over the sensor with cyan vignetting problems vs one filter over each and every lens with cyan vignetting problems, why pick the latter?

 

My gut reaction is that they will eventially solve the IR contamination problems at the sensor level. Meanwhile we have the M8 with filters...not the end of the world, but definitely annoying,

 

Rex