Sorry, I think Leica as something wrong.

[Mornnb]

Advertisement (gone after registration)

If you own an M9 and M240, turn off lens detection for a wide-angle lens or a very fast normal lens, see which one vignettes more without any type of lens corrections applied. The "Tricks" played to get the sensor to work include using special offset microlens arrays and thinning the sensor. For CCD's, sensor thinning means less material to store electrons. The sensor in the M9 has a lower saturation count than the one in the M8, both use the same width pixels, I can only assume the M9 pixels are not as deep. With CMOS, it means giving up material in the pixel and bringing the processing portion of the chip in closer to the light gathering sections. Reduced saturation count and a possible source of noise. SLR's do not have this problem as the angle of incidence allows use of deep pixels. The saturation count on the CMOS sensor used in the Nikon D4 and Df is double that of the M8, triple that of the M9 and M240. ISO performance in terms of noise compares with the M Monochrom, I have had both cameras out at the same time.

 

Back-Side Illuminated CMOS is promising. 

 

The CMOS chip in the M240 is not using offset micro prisms. What they are doing is using shaped micro lenses which pick up extreme ray angles common in rangefinder lenses and avoid spill over into neighbouring pixels as per the diagram below. You will notice the M240 handles wide angle lenses with less vignetting and colour fringing because of this microprism design.

 

Canon has a different design, using gapless micro lenses to pick up as much light as possible. To increase the high ISO performance. 

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!

[BerndReini]

 

....... I note a Belgian Company (sounds familiar ?) has been working on a CCD based sensor with CMOS electronics bolted on to the back to produce a hybrid chip with the advantages of both ...... currently more designed for imaging from space .... but you never know .... technology advances very rapidly these days ....

I will put that in the category of a completely different type of sensor then. I'm a big fan of trying different sensors. I gladly compromise in some areas if I like what I'm getting. I own a Sigma DP2 Merrill and it is like owning a little medium format camera, but with all its shortcomings unfortunately it would never be my main camera. Of course this is what other people would say about the M9, which is my main camera.

 

If Leica really did come up with a new sensor technology, it would certainly be good for the company, given it does give superior results in at least some areas. So in this respect I agree with the OP. Where I disagree is that Leica doesn't understand this problem, they just may not have any options. If you go back and read what people on the forums wanted from the new M before the 240 came out, it was higher ISO, live view, more dynamic range, more resolution. Leica listened and offered almost everything users collectively asked for in the M240, and the only way to do it at the time was with a CMOS sensor.

[Lenshacker]

The CMOS chip in the M240 is not using offset micro prisms. What they are doing is using shaped micro lenses which pick up extreme ray angles common in rangefinder lenses and avoid spill over into neighbouring pixels as per the diagram below. You will notice the M240 handles wide angle lenses with less vignetting and colour fringing because of this microprism design.

 

Canon has a different design, using gapless micro lenses to pick up as much light as possible. To increase the high ISO performance. 

I wonder if Kodak, (later Truesense, now ON Semiconductor) holds the patent on the offset microlens array. I've seen one comparison of the same lens on an M9 and an M240 with lens detection turned off on both cameras, the M240 showed much more vignetting than did the M9. I've used a Jupiter-12 (with Zeiss elements in it) on the M8 and M9- was amazed how little vignetting occurs.

 

On Semi has published the spec sheet for the KAF-18500 with Quantum Efficiency as a function of angle of incidence.  It would be interesting to see the numbers for the CMOSIS sensor and Canon sensor. 

[jdlaing]

I hear you keep saying the spec sheet for the KAF 18500 where it relates to Leica's use of that sensor. I hope you understand Leica did not use the off the shelf sensor as it was modified for their use. The sensor wafer in itself is a mindless brick without the accessories. That spec sheet is a nice place to start but the end all for what Leica bought.

[Lenshacker]

Read the spec sheets.

 

The Leica M8 used the KAF-10500 and the Leica M9 used the KAF-18500. The M8's spec sheet was published years ago, the M9 spec sheet was not. The M8 KAF10500 spec sheet is not on the On Semi website, I have the long-sheet downloaded. These are the specification sheets for the sensors used in the cameras, including the IR cover glass used, QE for the Red, Green, Blue of the Bayer filter, and the offset microlens arrays.

 

http://www.adorama.com/alc/0011872/blogarticle/Kodak-statement-about-Leica-M9-sensor

 

http://www.onsemi.com/pub_link/Collateral/KAF-18500-D.PDF

 

http://www.kodak.com/ek/uploadedFiles/Content/Small_Business/Images_Sensor_Solutions/Datasheets%28pdfs%29/KAF-10500ProductSummary.pdf

 

Leica used the KAF-10500 in the M8, and it's performance is accurately given in the spec sheet.

 

Leica used the KAF-18500 in the M9, and it's performance is accurately given in the spec sheet as published. On Semi seems to be free to sell the KAF-18500 now on the open market, this was not the case before.  It's easy enough to assign different part numbers for a specific product to avoid confusion in spec sheets. The KAF-18500 for the M9 is now available through On Semi.

 

http://www.onsemi.com/PowerSolutions/product.do?id=KAF-18500

 

I found the specification sheet for the KAF-18500 and KAF-10500 very useful for writing the raw processor, needed to know how many calibration columns were in the data. The Spec Sheets as published match the files as stored for the M9 and M8 storing files in service mode.

[jdlaing]

It doesn't quite work that way. Leica does not walk in to a sensor manufacturer like WalMart and buy off the shelf.

They work with the manufacturer over, in the case of this one, 25 months get what they want. Different chip architecture, proprietary cover glass coatings, guarantees on percentages of power on/off cycles, tempest levels and some other little niglets. 

 

You cannot buy the same package that Leica buys as it is proprietary by contract.

 

Then the third party programmer has some input on minor hardware changes. 

 

The spec sheet is only a starting point.

[jdlaing]

Advertisement (gone after registration)

The conversation goes like this:

 

Hello. I am Joe Bob Schwanz, sensor manufacturer.

 

Hello. I am Billy Bob Butt, Camera manufacturer. We are interested in one of your products. Can you make it do this, this and that and can you facilitate this and that to get the result we want?

 

Joe Bob: Yes we can, and how many would you like.

 

Billy Bob: We need 12 prototypes with a reservation for an order of 85,000 with a parts reserve of 8,500 maintained by you for a period of ten years. Our lawyers and contract department will be in touch.

 

Billy Bob's Contract Department: These items are proprietary in nature and may not be sold to anyone else for a period of twenty one years and then only with our permission. You may market whatever you like from the original design but not in the package we specified. Please sign here.

 

Billy Bob's Third Party Programmer: We have a design change proposal. Can we get this item changed?

 

Joe Bob: No problem. 

 

This is how it works.

[Lenshacker]

Right.

 

Unless you post the technical differences between the sensor as delivered to Leica and the differences with the published spec, I do not believe it.

 

"4H2058 KAF-18500-NXA-JH-AA-08 Special Color, Aperture, Enhanced, ESD, LOD, Microlens, Sealed IR Cover Glass" , "S8612 with MAR coating (0.8mm) Transmission", from the spec, right down to the type of IR cover glass and thickness used in the M9.

 

And I have called Kodak's sensor Division to have an off-the-shelf sensor revised for use. Kodak sold it as a standard product afterwards.

 

I don't know what your experience is dealing with sensor manufacturers, but I do know the long-sheet as published reflects the image being recorded in my M9. I've had several custom parts made for projects, by allowing the company to retain commercial rights I saved a lot on the NRE cost. I've had the company put up 80% of the NRE to implement a design, deliver prototypes and production to me, then go on to market the parts. That's how I work it.

 

A sensor design that was  derivative of the proprietary design, but different enough not to be bound by such an agreement, would be assigned a different part number. The part number for the data sheet linked to is the "KAF-18500". That is the part used in the M9.

 

There is more to a digital camera than just the sensor, especially an Analog sensor like the KAF-18500. It was called a "Data Acquisition System" when I worked on them.

[Mornnb]

I wonder if Kodak, (later Truesense, now ON Semiconductor) holds the patent on the offset microlens array.

 

They don't, as this is what Sony is doing on the A7R. Here's Sony's sensor, and it doesn't work any where near as well as the M9 or M240.

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!

[mjh]

Unless you post the technical differences between the sensor as delivered to Leica and the differences with the published spec, I do not believe it.

 

"4H2058 KAF-18500-NXA-JH-AA-08 Special Color, Aperture, Enhanced, ESD, LOD, Microlens, Sealed IR Cover Glass" , "S8612 with MAR coating (0.8mm) Transmission", from the spec, right down to the type of IR cover glass and thickness used in the M9.

The specs refer to a sensor with an IR cut filter, unlike the absorption filter as used for the sensor of the M9. The spec sheet vaguely refers to microlenses without giving their precise specs; it is unlikely that any customer other than Leica would have a use for the special microlens shifting employed in the M9.

 

The samples offered, on the other hand, specify an IR absorption filter. I guess On Semi hasn’t sold this sensor to anyone else so far and the only samples available are those of Leica’s version.

[jdlaing]

Lenshacker,

 

I am afraid you will depart the planet before you get that information and I personally could care less if you believe it.

It is proprietary information.

 

[theendlesshouse]

I applaud Leica for battling through some very hard times and flourishing in the digital age, where far to much emphasis is placed on mega-pixels and sharpness, this versus that! They provide unparalleled tools so we can do what we love with confidence! That they also happen to be aesthetically exquisite and very discreet is a wonderful bonus.

[Lenshacker]

The specs refer to a sensor with an IR cut filter, unlike the absorption filter as used for the sensor of the M9. The spec sheet vaguely refers to microlenses without giving their precise specs; it is unlikely that any customer other than Leica would have a use for the special microlens shifting employed in the M9.

 

The samples offered, on the other hand, specify an IR absorption filter. I guess On Semi hasn’t sold this sensor to anyone else so far and the only samples available are those of Leica’s version.

They state that the sensor shown uses the S8612 glass, which is the problem with the M9 sensor. S8612 absorbs Infrared. See page 28.

 

The Microlens array is shown as "special", same as in the KAF-10500 specification.

 

Just put up the QE as a function of angle of incidence for the M9, that will be definitive.

 

This is the data sheet for the KAF-18500. The KAF-18500 was stated to be the sensor used in the M9. Same with the KAF-10500 data sheet and the M8. Had the design specifications been significantly different, the part number would also be different. That's why we have part numbers and data sheets. If the engineer responsible for sensor selection on the M9 wants to read the spec and announce that this is different from what he used, I'd tend to believe it. Otherwise, I believe Part Numbers and Spec Sheets that have that part number on them.

 

I think Leica might have bought some of these. Had they used a different IR cover glass, they would not be having the problems. Let's hope a revised sensor comes out with a different cover glass, I'll keep an eye open on the ON website.

 

Kodak would have done better to keep these data sheets and the techniques used in their sensors more "proprietary". As it was they tended to act as a research lab and publish their work. I think in the end it benefited the industry as a whole, but gave up an advantage. 

[Lenshacker]

http://www.optcorp.com/pdf/FLI/31Mp_and_39Mp_Full-Frame_CCD_paper.pdf

 

This paper is typical of Kodak describing their sensors; the 39MPixel CCD used by Phase 1.

[dant]

I think it is more to a camera than the sensor.

 

 

Yes concur. I shoot Leica for the controls. The MM sensor is top rate except for the highlight burning issue. The M240 sensor is just OK. I could shoot better sensor cams for less, but the controls are crap...like the Sony.

[Lenshacker]

My main objection to the Sony is that it does not store Raw data; it uses a lossy image compression scheme in raw mode. This is a function of the firmware, not the camera. Maybe someone will discover a service mode in the Sony to store true raw images, as is now available with the M8. I did my own demosaic routine for the M9 using a yellow filter, converted the image to linear-monochrome DNG and used 'AFFF'x (45055) for "white" after doing some max/min measurements of the converted images. The M Monochrom uses '3FFF'x (16383). Jaap was right about Highlights with a color camera converting to monochrome.

 

I did find some of the Kodak patents for gapless micro-lens arrays, including offset arrays. The patents included details of fabrication, so nothing as broad as using microlens arrays on sensors. I've been reading Kodak data sheets since the KAF-1300 was new, ~1991 or so.

[Herr Barnack]

 

Traductor

 

 

 

Leica is not just a camera to take pictures and a goal, Leica is a filosifia.

 

I note with concern a course is wrong.

 

I do not like having a photographic camera that costs 7,000 euros a CMOS inside, that is already a canikon, a CCD was different, something that NASA used in missions to Mars.

 

My Monochrom is a product that is not seen in other manufacturers, it's different, but CMOS is equal to others.

 

When Leica launched the digital way many shouted: NO !!

 

Now we are on the same path as the rest, I do not like this.

 

I think I'm not alone.

 

I will not buy another camera Leica CMOS inside, I prefer any other manufacturer, canikon.

 

My Leica Monochrom and my Leica Mp will remain with me, but no new ones.

 

It is my thought.

 

Regarding the price, the Monochrom 246 will sell for $500 USD/446 € less than the original Monochrom did three years ago when it was released.

 

Personally, I do not object to the price of the M/M 246.  I object only to the fact that I do not have a spare $7450 USD at present to buy one. 

[jdlaing]

There is no such thing as a SPARE amount of money. If you had SPARE money you'd blow it on something anyway.

 

Leica equipment is pocket change anyway. I spend more per year on cigars than that.

[xvarior01]

We have been through this many times, and I understand there are various tricks to get the most out of CCD sensors at high ISO. I am perfectly happy with my M9 and just added a CCD MM to my arsenal. I do not own any CMOS cameras with the exception of my iphone and a Sony RX100 point and shoot that I let my kids use. However, the greatest, most competitive engineers in photography and cinema have not been able to produce CCD cameras that can compete with CMOS cameras in high ISO. If your CCD camera can do ISO 10,000 then the CMOS version can do 25,000 at the same noise level and without banding.

 

I have been shooting video with professional Panasonic CCD cameras for a long time, and my camera just got replaced with a new model with a CMOS sensor (Varicam 35 and Varicam HS). My camera's base ISO is 320 and I love the colors and rendering from it. The new camera has a dual ISO (!!!) of 850 and 5,000. I don't know exactly how this is done, but in a demonstration ISO 850 and 5,000 were completely free of noise and in many examples hard to distinguish from each other, while anything in between ISO 850 and 5,000 was noisier.

 

There is at least a stop and a third gain in high ISO between my m9 and the M240. I use my M9 up to ISO 1,000 and I would use the M240 up to ISO 2,500 without reservations. Whether any of that gain matters to you and me personally is not the question. The fact is that unless you, or some other specialty company builds another CCD camera, then the CCD sensor in cameras for general photography is dead. Done. Leica will not build another camera with a CCD chip. Anyone betting against that might as well put their money on Qatar winning the next soccer world cup.

Very well said and true except for the Qatar analogy. They already bought many of Fifa officials to get competition played on their land, I won't be surprised if they buy the world to win the cup :-)

[Advertisement]

Hallo,

Look here: Leica M

[Herr Barnack]

 

 

...Leica equipment is pocket change anyway. I spend more per year on cigars than that.

Congratulations, I guess...