M-M Higher resolution, Really? Thinking this through a bit...

[mjm6]

QC issues. The Achromatic+ digital back is expensive because they have to pick a "flawless" sensor (1 in 10 or so?) Pixel mapping and such can not be done there to mask the flaws.




Yes, it is a monochrome sensor with no Bayer dyes but there is interpolation going on to get a RGB output. I will even categorically say that it is a M9 sensor without the Bayer dyes.

I am not going to post anymore on this. But, eventually the truth will come out.

OK, I think I see what you are saying, and I'll bet you are correct. No way to avoid the hot pixels and other flaws without some overlayment mapping to ensure they don't stand out.


---Michael

 

[etrigan63]

Considering recent history (ie: Nikon's claim that the D800e had no AA filter when, in fact, it has a "neutered" one), I can see why Vivek thinks the way he does. To produce a true monochrome sensor would require more than simply turning off the Bayer Filter Insertion step on the assembly line. Engineering considerations, changes in electronics, changes in the layering of the sensor, the new file spec, a whole new doping process, etc, etc, would greatly increase the cost. Perhaps Kodak/Platinum Group found a way to "neuter" the bayer filters like Nikon did to the AA filter in the D800e? As he said, either way the truth will eventually come to light.

 

[docmoore]

Bob,

Do you understand how optics systems are defined for performance testing? 'Sharpness' isn't really defined to my understanding, but 'resolution' is a definable term.

Sharpness in a print is often considered a combination of resolution and acutance, and is more about perception, not about performance. Hence the purpose for unsharp masks, for example.

At least that's how I've always interpreted when people use the term sharpness.

Look here:
Understanding resolution and MTF


---Michael

Which is why Leica chose the term sharp....they could put ANY number they want on it and make it stick.

Bob

 

[mjm6]

Bob,

Yup...

Thanks everyone for the comments and information... This forum is one of the best our there; highly knowledgeable and friendly people. It's always a pleasure to learn from others with a passion for this, and more knowledge than I.


---Michael

 

[docmoore]

the claim of higher resolution. We all know that the 100% higher resolution that Leica is claiming is nonsense (why do they do that, I wonder?). I could see 15% better, some people have said 30% better, but it's going to be in that range, not 100%.


Does anyone have any insight into this?


---Michael

Here Erwin Puts agrees with you....

Blog

Bob

 

[Brian S]

Considering recent history (ie: Nikon's claim that the D800e had no AA filter when, in fact, it has a "neutered" one), I can see why Vivek thinks the way he does. To produce a true monochrome sensor would require more than simply turning off the Bayer Filter Insertion step on the assembly line. Engineering considerations, changes in electronics, changes in the layering of the sensor, the new file spec, a whole new doping process, etc, etc, would greatly increase the cost. Perhaps Kodak/Platinum Group found a way to "neuter" the bayer filters like Nikon did to the AA filter in the D800e? As he said, either way the truth will eventually come to light.

All of the detectors used with Digital cameras are monochrome in nature, and Silicon based. All of the commercially available detectors use color dye in front of the photo-sensitive elements to produce color. Most do it with a mosaic filter, most using a 2x2 Bayer pattern. Foveon layers three monochrome planes with different color filters in front of each, more like film.

The alternative would be to fabricate detectors using different material, each with a different spectral response and use a series of beamsplitters for each. No one is doing that. Fabricating a single detector with individual photosensitive sites using different material on the same chip to eliminate the need for a mosaic filter- anybody know how to make one?

The Leica M9M uses a monochrome detector. It's easy to do, and Kodak has been offering detectors like this for decades. That is the truth, been there, done that with Monochrome cameras and with "color" (multi-spectral) cameras using beam-splitters and detectors with different spectral response.

The Kodak Sensor Group has offered monochrome sensors for decades.

http://www.kodak.com/ek/uploadedFil.../Datasheets(pdfs)/KAF-16803ProductSummary.pdf

http://www.kodak.com/ek/uploadedFil...s/Datasheets(pdfs)/KAF-6303ProductSummary.pdf

http://www.kodak.com/ek/uploadedFil...s/Datasheets(pdfs)/KAF-1603ProductSummary.pdf

No need to change the drive electronics when these devices were used in the color and monochrome version of the same camera. I have both.

 

[Brian S]

A different 16MPixel CCD from Kodak that was offered in Color and Monochrome. Part numbers supplied., but the part was recently discontinued. Maybe the KAF-18500 Monochrome did it in.

Kodak Image Sensor Solutions - KAF-16802

KAF-16802-CAA-DD-AA
Color (Bayer RGB), No Microlens, CERDIP Package (sidebrazed, CuW), Clear Cover Glass with AR coating (both sides), Standard Grade

KAF-16802-CAA-DD-AE
Color (Bayer RGB), No Microlens, CERDIP Package (sidebrazed, CuW), Clear Cover Glass with AR coating (both sides), Engineering Grade
KAF-16802-CA

KAF-16802-AAA-DD-AA
Monochrome, No Microlens, CERDIP Package (sidebrazed, CuW), Clear Cover Glass with AR coating (both sides), Standard Grade

KAF-16802-AAA-DD-AE
Monochrome, No Microlens, CERDIP Package (sidebrazed, CuW), Clear Cover Glass with AR coating (both sides), Engineering Grade
KAF-16802-AA

 

[bensonga]

Kodak 760m Review

 

[jlm]

so out of curiosity, why is Vivek saying there is RGB output from the MM?
you could create a neutral gray (black and white) composed of RGB values and use the Bayer filter (M9 method), but if you eliminate the Bayer filter and the interpolating and simply record luminance values at each pixel, no need and no native ability to make RGB;
If the luminance levels need to be represented as RGB values, that would be a simple chart of equal R,G and B values

 

[D&A]

Kodak 760m Review

Ben,

Interestingly I read Pete Myers review when it was first posted on LuLa some time ago It was then that I became so intrigued with the Kodak 760M Monchrome that a number of years later when one of the Kodak-Nikon Varients was actually put of for sale (used), I serious considered it and debated with myself it's purchase for a considerable time.

Now after the Leica M9M has been announced, I too was perplexed with Leic's proclamation that its monochorme camera is 100% sharper. I was trying to understand exactly what they meant by that and if there was some matamatical or technical explanation how they derived at this figure.

Less than a week ago I re-read Pete Myers article on the 760M and one line he wrote caught my eye, where he said the following:

>>>"Without an anti aliasing filter and no Bayer color matrix, the resolution of a 6 mega pixel monochrome camera is astonishing. In monochrome, 6 mega pixels effectively does what it takes 12-24 mega pixels with a color matrix."<<<

I won't specifically provide conjecture for exactly what he meant and whether it had anything to do with the figure Leica came up with, but I've been mulling over both his and Leica's statement and trying to get a handle on all of it.

I appreciate the extremely interesting discussion by everyone in this thread. I've personally have found it both enlightning, informative and yes, even when there is disagreement, it provides a vehicle to grasp a better understanding of some of the technology behind the camera and it's development.

Dave (D&A)

 

[Brian S]

Well, this little guy thought the monochrome camera looked cool.



Resized in Photoshop. R60 filter, Kodak DCS200ir. Circa 1993.



8-Bit ADC, I am looking forward to the M9M. 6-bits extra. And $4,000 cheaper.

 

[cunim]

All of the detectors used with Digital cameras are monochrome in nature, and Silicon based. All of the commercially available detectors use color dye in front of the photo-sensitive elements to produce color. Most do it with a mosaic filter, most using a 2x2 Bayer pattern. Foveon layers three monochrome planes with different color filters in front of each, more like film.

The alternative would be to fabricate detectors using different material, each with a different spectral response and use a series of beamsplitters for each. No one is doing that. Fabricating a single detector with individual photosensitive sites using different material on the same chip to eliminate the need for a mosaic filter- anybody know how to make one?

The Leica M9M uses a monochrome detector. It's easy to do, and Kodak has been offering detectors like this for decades. That is the truth, been there, done that with Monochrome cameras and with "color" (multi-spectral) cameras using beam-splitters and detectors with different spectral response.

The Kodak Sensor Group has offered monochrome sensors for decades.

http://www.kodak.com/ek/uploadedFil.../Datasheets(pdfs)/KAF-16803ProductSummary.pdf

http://www.kodak.com/ek/uploadedFil...s/Datasheets(pdfs)/KAF-6303ProductSummary.pdf

http://www.kodak.com/ek/uploadedFil...s/Datasheets(pdfs)/KAF-1603ProductSummary.pdf

No need to change the drive electronics when these devices were used in the color and monochrome version of the same camera. I have both.

Brian, just to clarify - 3CCD color cameras have similar spectral responses in silicon, but filter wheels, interference filters and/or beamsplitters create the discrete RGB data streams.

The downside of the beamsplitter is that it eats light and some MTF. Filter wheels, while inconvenient and failure-prone, retain sensitivity because a large interference filter can be very efficient. However, there are some spectral shifts across the CCD arising from the way in which interference filter respoonse characteristics interact with incidence angle to the detector.

All that aside, there is no comparison between a bayer image and one made by a 3CCD camera. I think that is why a discussion of multi-image technologies arose in another topic.

The M9M is really neat. To tell the truth, I am not a fan of M9 images. To my eye, however, the M images are much better at monochrome than the bayer CCD is at color. I do like the M9M (not enough to buy one). However, thinking of how Leica could create a truly revolutionary camera - put in a microstepper and do multishot. One compact camera could yield color images with the quality of the M9M (which could then be converted to monochrome), or just normal color images (in single mode). That would get me into Leica right quick.

 

[Brian S]

The sensor that I worked with in the 1980s used different material for the two-arrays, so the arrays themselves had different spectral response.

Minolta made a 3 CCD camera in the 90s, all three CCD's were Silicon, color filters in front of each, and beamsplitters used for the image. The Minolta RD-175.

1995 I

Hey! I need to send these guys an Email.

1992

They left off that Kodak offered the DCS200 in Infrared as well as color and monochrome.

 

[Shashin]

All that aside, there is no comparison between a bayer image and one made by a 3CCD camera. I think that is why a discussion of multi-image technologies arose in another topic.

Here is a comparison of Bayer color and tri-color non-interpolated. Both 100% crops. See the difference? The question is also whether you are seeing a difference in resolution or sharpness/contrast in the M9M.

 

[cunim]

The sensor that I worked with in the 1980s used different material for the two-arrays, so the arrays themselves had different spectral response.

Minolta made a 3 CCD camera in the 90s, all three CCD's were Silicon, color filters in front of each, and beamsplitters used for the image. The Minolta RD-175.

1995 I

Hey! I need to send these guys an Email.

1992

They left off that Kodak offered the DCS200 in Infrared as well as color and monochrome.

Wow, neat site. Kind of off topic, but I remember going down to Akihabara in the 1990's and picking up a 640x480 Fuji digital camera. I thought it was amazing and everyone marvelled at it when I brought it home. "Look, easy digital!" At least until it died a week later. I think a multishot Leica would give me the same feeling of wonder.

 

[Tim Gray]

Now after the Leica M9M has been announced, I too was perplexed with Leic's proclamation that its monochorme camera is 100% sharper. I was trying to understand exactly what they meant by that and if there was some matamatical or technical explanation how they derived at this figure.

I think the key is they said "100% sharper" not "100% more resolution". Resolution is something that is relatively easy to quantify. I don't think sharpness is. Isn't sharpness essentially perceptual, and based on both resolution and acutance?

At least that's how I'm reading it The M-M should resolve more than the M9, but not 100% more.

 

[jlm]

my understanding:

simplified by assuming equal distribution of R.G.B filters, interpolation scheme, blah, blah:

at each pixel location, luminance is set by that pixel's reading (this is color blind)

The R,G,B levels are averaged over adjacent pixels and then that average is assigned to the main pixel and output as separate R,G and B values for that pixel. This is the interpolation and since the adjacent pixels are spread out, the actual imaging data is spread out and averaged and therefore reduced in resolution from the total pixel count. there are still the same number of pixels, but, here you go again...the image is not as sharp!

Now if the MM only uses the luminance value at the main pixel, the resolution would be equal to the pixel count and the image would be sharper than the interpolated version

 

[Brian S]

What surprises me, lots of people end up resizing the color interpolated images to smaller images.

Is there any software package that combines the 2x2 bayer site into a real RGB pixel? This would effectively make it a 4.5MPixel image without artifacts from moire.

Take the average of the 2 green pixels, the red, and the blue- get an RGB pixel without interpolation. Of course, this is for the M9 and M8- not the M9M.

 

[Tim Gray]

Is there any software package that combines the 2x2 bayer site into a real RGB pixel? This would effectively make it a 4.5MPixel image without artifacts from moire.

I'm pretty sure that's what Raw Photo Processor does in one of it's modes.

Raw Photo Processor (RPP)

 

[Hosermage]

Brian, why would you need two green pixels instead of one? This link comes to my mind: sonyalpharumors | Blog | (SR5) New Sony RGBW sensor coming in compact camera by end of the year!, which to me sounds reasonable to allocate 1 of the green to basically BW.

On the same idea, doesn't that make the single pixel 4 times as big, which I'm sure will lead to other bad things