Leica M8 sensor=?

[keithwms]

So, like many of you I have been trying to follow the rumours of the M8.

I'd like to throw one extra thing into the mix. There apparently will be a new foveon-based camera, the sigma sd14, announced at Photokina. Yes it is an SLR, but wait! With non-Bayer sensor developments brewing, there could indeed be some substantial excitement at photokina. I'd love to see a next-gen foveon in a rangefinder. Or at least something like the Fuji S/R sensor. I just don't see why Leica would go retro and bring out a camera based on the same type of Bayer sensor one can get in many SLRs. Really, it's not going to be fun going from film to a Bayer sensor and learning how to avoid blowing highlights and all that.

Well maybe this post is more of a wish rather than a newsworthy post....

But do we have any actual details on the M8 sensor? If they have been posted here, sorry, I have missed them, I saw some speculation but no clear info.

Keith

 

[Bob Ross]

So, like many of you I have been trying to follow the rumours of the M8.

I'd like to throw one extra thing into the mix. There apparently will be a new foveon-based camera, the sigma sd14, announced at Photokina. Yes it is an SLR, but wait! With non-Bayer sensor developments brewing, there could indeed be some substantial excitement at photokina. I'd love to see a next-gen foveon in a rangefinder. Or at least something like the Fuji S/R sensor. I just don't see why Leica would go retro and bring out a camera based on the same type of Bayer sensor one can get in many SLRs. Really, it's not going to be fun going from film to a Bayer sensor and learning how to avoid blowing highlights and all that.

Well maybe this post is more of a wish rather than a newsworthy post....

But do we have any actual details on the M8 sensor? If they have been posted here, sorry, I have missed them, I saw some speculation but no clear info.

Keith

No actual details until Sept. 15th. The best speculation that I've heard is that it will be a derivative of the Kodak KAF-10010 found in the DMR. Differences probably are slightly bigger and new offset micrlenses to cope with the angles from the RF lenses.
With the M8 being a bit of a gamble to start with, I don't think that they wanted to go with new sensor technology. The Kodak KAF FFT sensor family is also found in many of the MF backs, their own DMR and the 4/3rds system. Bayer sensors are not exactly retro and you can certainly blow highlights with a Foveon sensor. Out of all the digital cameras in production there is only one that doesn't used a Bayer sensor, the SD-14 which has not been released.
Bob

 

[keithwms]

Bayer sensors are not exactly retro and you can certainly blow highlights with a Foveon sensor.

Sure, but it is very easy to blow out many DSLRs, and there are already good alternatives to Bayer on the market, i.e. the S/R sensor in the Fuji S3. And the Foveons do offer larger pixel sizes per equivalent resolution, so I expect that the newer foveons should offer lower noise levels and better dynamic range (fingers tightly crossed).

As I recall, the sigma sd-10 had wonderfully low and monochromatic noise and performed extremely well up to ISO 800. I didn't find any dynamic range specs for the foveons after some brief googling, but DR and noise are closely linked.

If this M8 is to appeal to film shooters, and perhaps a high-than-average proportion of b&w film shooters than exists in the DSLR consumer base... I think Leica can expect anger if what they deliver is the kind of performance you get from a mid-range cropped-frame DSLR!

I just think that the expectations are going to be substantially higher than in the DSLR market. As they should be. This thing is going to be made for some of the best glass there is.

Well we'll just have to wait and see how it all plays out.

 

[egpj]

It could be that because of the design (i.e. distance of the lens from the film plane) that a Foveon sensor would not be practicle. Foveon sensors rely on the light to penetrate the sensor surface to certain depths. Blue gets read at the top then green further down and then finally red at the deepest layer. Think that this would be very difficult to accomplish if the light was hitting the topmost layer at an angle and you would not get full penetration to the sub-layers of the Foveon. Unfortunatley, in this case, penetration however slight is not sufficient to commit the act.

They (Leica) were only able to use the bayer pattern sensor after using microlenses coupled with software tweaks for whatever lens is mounted.

 

[keithwms]

It could be that because of the design (i.e. distance of the lens from the film plane) that a Foveon sensor would not be practicle. Foveon sensors rely on the light to penetrate the sensor surface to certain depths. Blue gets read at the top then green further down and then finally red at the deepest layer. Think that this would be very difficult to accomplish if the light was hitting the topmost layer at an angle and you would not get full penetration to the sub-layers of the Foveon. Unfortunatley, in this case, penetration however slight is not sufficient to commit the act.

They (Leica) were only able to use the bayer pattern sensor after using microlenses coupled with software tweaks for whatever lens is mounted.

I'm glad you mentioned that. My take is a bit different. In fact, when I read the first rumour that Leica were coming out with a digital M8, my immediate thought was, well, that will mate very nicely to a foveon because the lens will be so close to the sensor. Best possible case for a foveon... or for that other multilayered sensing medium that we all know and love... because the light will come straight in, so there will be much less issue of angle.

Maybe I have it all wrong 😉

 

[egpj]

The closer you get to the recording media the higher the angle of the light, well, the higher the angle as you get farther out to the edges of the recording media. Got it?😉 🙄

 

[Bob Ross]

Hi Keith,
A couple of points that might be involved. There is a difference in the fill factor between the Kodak and Foveon sensors and that three layer arrangement requires larger pixels, so light can "go around" to get to the deeper layers. When light come in at high angles color aberations are quite likely. The Kodak FFT sensors have high fill factors and high saturation signals, so there is a larger margin to cope with the signal fall off in the corners. The new Foveon sensor is rumored to have gone from 3.4MP to 4.6MP with the same 1.7 crop factor, so the pixels are smaller. We'll have to see if that has any affect on the image. Foveon has been so quiet for so long that they could have rebuilt the whole sensor from silicon up, for all we know.
As for DR, I have an E-1 with the Kodak FFT sensor and DR is not one of its issuues. The Fuji SR sensor is the king of the DR heap, but it is processor intensive and considered very slow. It, of course has a modified Bayer RGB color array.
I suspect that Leica went after the sensor that could handle the rangefinder optical geometry problems the best. Images from the DRM are good indicators, but they are few and far between on the web compared to say, the D70 or Rebels. Comparisons that I have seen are against the Canon full frames.
Bob