Full Frame Is Here !!!!!!

[andym911]

Advertisement (gone after registration)

I have said it before and I'll say it again..

 

Fotokina 2008 will be where Leica launch their Digital Rangefinder Full Frame...

 

If they don't I will eat my Elmar...

 

Andy

[Jamie Roberts]

Kodak’s panchromatic pixel sensor isn’t quite the universal solution it is often made to be. Yes, the panchromatic pixels will raise the sensor’s ISO sensitivity (a sensor’s ISO sensitivity is equal to the ISO sensitivity of its most sensitive pixels), but at the same time, the raised ISO value implies that the less sensitive chroma pixels will get underexposed, increasing the amount of noise in the color channels. Also, the spatial resolution in the color channels is reduced, as just 50 percent of the pixels are color sensitive at all. And it’s not mere theory; evidence for this can be found in the example images provided by Kodak. I doubt we will ever see this sensor used for serious photography.

 

Interesting! Thanks Michael, for the correction!

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[adan]

Again I have to ask why so many continue to assume that technological limits have been reached, and that what we see available today in sensors is what we should base predictions on. (This time with reference to the Kodak hi-ISO pixel pattern).

 

"I doubt we will ever see this sensor used for serious photography."

 

Physicists could still "prove", as of 1895, that powered heavier-than-air flight was theoretically impossible - engines could not produce enough power to lift their own weight, and fuels' caloric content was too low to lift THEIR own weight.

 

Then the technogeeks of the time produced fuels with more calories per pound, and switched to aluminum for engine blocks instead of steel/iron - and voila - the Wright Brothers got off the ground.

 

Kodak's new patterned sensor may not be ready for prime time today (after all, they announced it less than 6 months ago). I suspect Barnack's Ur-Leica of c. 1913 wouldn't have gotten off the ground as a marketable product either.

 

But - things - change.

 

The head of the U.S. Patent Office in 1899 is famously supposed to have said "Everything that can be invented has been invented." Except that he never said any such thing - he wasn't that silly. (Charles H. Duell)

 

The best way to approximate the real technological near future is to double whatever we see today (or halve the downsides - speaking purely of the science itself, not necessarily the sociological implications).

 

When rumors of the Canon 5D arose in 2005, the conventional wisdom was CERTAIN that it was a hoax. "Can't possibly produce a full-frame sensor cheap enough to make money with a $3000 camera!!" was all over the Canon forum at dpreview.

 

Ooops! Especially since the 5d now sells new for $2150 devalued dollars (Double Ooops!)

[mjh]

Again I have to ask why so many continue to assume that technological limits have been reached, and that what we see available today in sensors is what we should base predictions on. (This time with reference to the Kodak hi-ISO pixel pattern).

 

"I doubt we will ever see this sensor used for serious photography."

 

Physicists could still "prove", as of 1895, that powered heavier-than-air flight was theoretically impossible - engines could not produce enough power to lift their own weight, and fuels' caloric content was too low to lift THEIR own weight.

That’s neither here nor there. You didn’t address my arguments – challenge them if you can.

 

By the way, I think it’s quite obvious that technological limits have not been reached. Still, Kodak’s new filter pattern(s) won’t get us anywhere. The excitement about this technology seems to be inversely proportional to the effort people took in examining Kodak’s claims.

[sdai]

When rumors of the Canon 5D arose in 2005, the conventional wisdom was CERTAIN that it was a hoax. "Can't possibly produce a full-frame sensor cheap enough to make money with a $3000 camera!!" was all over the Canon forum at dpreview.

 

The cost reduction achieved in the 5D's FF CMOS sensor is mainly due to:

 

1. the advancement in the wafer exposure process, Canon used to have to exposure 3 times to cover a 36x24 area now they only need to do it 2 times therefore the success rate has greatly improved.

 

2. the adoption of several new materials and simplified structure of the sensor assembly.

 

Canon has continued to improved in these two areas and mass adoption of FF sensors will soon wide spread in the industry.

[gravastar]

The cost reduction achieved in the 5D's FF CMOS sensor is mainly due to:

 

1. the advancement in the wafer exposure process, Canon used to have to exposure 3 times to cover a 36x24 area now they only need to do it 2 times therefore the success rate has greatly improved......................................

 

I was talking to a Canon R&D guy some months ago and he said they don't have to stitch exposures any more and can do it in one. They make their own steppers though.

 

 

 

Bob.

[adan]

Advertisement (gone after registration)

Michael: Specific issues, as you describe, require specific solutions. And I don't disagree that Kodak's idea introduces its own issues to be addressed.

 

Since I don't have insider knowledge of Kodak's imaging labs, I wouldn't attempt to guess what they are playing around with in terms of different approaches to demosaic-ing algorithms or chroma-only noise-reduction that avoids affecting the luminance signal's "sharpness". I can't even guess at what to guess at (if you'll pardon the Rumsfeld-ism).

 

I just look at the overall historical curve of technological advance - and the winning bet long-term is almost always AGAINST statements containing "never", "ever" or "doubt".

 

Maybe the new Kodak pattern will become just another "Foveon", relegated to the backwaters of digital imaging. Maybe not.

 

sdai and gravastar: Can't tell for sure if your were supporting me or taking issue (wink!) But sure, Canon obviously did SOMETHING that moved the technology ahead faster than the conventional wisdom thought possible ("lower cost" counting as an advance in my book).

[sdai]

I was talking to a Canon R&D guy some months ago and he said they don't have to stitch exposures any more and can do it in one. They make their own steppers though.

 

I was referring to the process adopted to make the 5D sensor, of course Canon could make it much simpler and cheaper now.

 

The Nikon stepper apparently wasn't up to the task, turning good ideas on paper into product really needs some great tools and I think that's why they're lagging behind. I haven't followed the industry closely enough ... but Canon seems to have some agreement with ASML, Zeiss in these areas, I think that we can expect some dramatic improvement very shortly.

[carstenw]

That’s neither here nor there. You didn’t address my arguments – challenge them if you can.

 

By the way, I think it’s quite obvious that technological limits have not been reached. Still, Kodak’s new filter pattern(s) won’t get us anywhere. The excitement about this technology seems to be inversely proportional to the effort people took in examining Kodak’s claims.

Michael, while your arguments are valid, I wonder if they will be relevant if Kodak decides to trade off pixel size and high ISO noise versus this new pattern. I wonder which factor will win, ie. will the improved sensitivity win over smaller pixels? If so, we could indeed see this sensor used in a camera.

[stunsworth]

Well being controversial <grin>. The 10 mp sensor in the M8 already allows me to print up to A2 provided that I've managed to handle to technical side of things ok, so a 16 mp sensor (or thereabouts) should let me do this with little interpolation.

 

I'd prefer Leica/Kodak to concentrate on lowering noise at the higher ISOs. After saying that, I know a lot of fuss is made about how poor the camera is at 1250 and 2500, but comparing the results from the M8 to what I obtained using Neopan 1600 and Delta 3200 the M8 is better, and I suspect that the _vast_ majority of file that was sold was ISO 400 or less - with most being ISO 100. High ISOs only seem to have become a 'holy grail' with the advent of digital.

[Olsen]

I have said it before and I'll say it again..

 

Fotokina 2008 will be where Leica launch their Digital Rangefinder Full Frame...

 

If they don't I will eat my Elmar...

 

Andy

 

I would do some training before the event. Like eating glass, coins and small stuff first. Then increase to spark plugs, knives and forks. There was a guy in America who ate up a whole automobile. - If he drank the, glycol, lubricant oil and petrol, I don't know, he is dead now. So, an Elmar should be no problem.

 

Good Luck!

[milyovsky]

Are there any advantages to a full frame sensor besides the wider angle of view?

[adan]

Getting back to the original question - while I will always bet my money on technology coming through with solutions in the long run, I would not bet money (or my tummy!) on a full-frame Leica M at PK08. The original interview quoted Lee as saying "when we can get it right".

 

The DMR debacle taught Leica that it's important to meet deadlines.

 

The M8 met an internal deadline Leica set for itself, at the cost of some imperfections (but as a newspaper guy, and recovering procrastinator-perfectionist myself, I think it was a step forward to actually meet a deadline imperfectly than continue the culture of "perfect procrastination" that was dragging the company down).

 

The R10 and the full-frame M (regardless of number) will be the tests to see if Leica can now balance the two goals smoothly.

 

I WOULD bet on a 16mm Super-Elmar-M f/3.4 (or maybe, maybe! a Super-Elmarit-M f/2.8, 16 or 18mm) appearing longl before a full-frame M - quite possibily at PK08.

 

But I won't eat anything if I'm wrong!

[adan]

Are there any advantages to a full frame sensor besides the wider angle of view?

 

Well, for the same number of pixels, you get bigger pixels to gobble more light, and thus less noise in the image - something the Nikon D3 and Canon 5D have proved very clearly.

 

Or you can get more pixels, and thus more resolution, for the same field of view.

 

Or you can get cheaper or faster wide-angles - the Leica 21 f/2.8 acts like a 28 on the M8, and cost north of $3000. Whereas with a full-frame M you could buy the 28 f/2.8 ASPH for less than half the price and get the same speed and field of view. Alternatively, one can use the 21 or 24 f/2.8 lenses at their original FOV and gain a stop of speed over the f/4 WATE.

 

On SLRs, the bigger sensor means a bigger ground-glass for viewing, and thus a bigger, clearer viewfinder (compare a Canon 1Ds viewfinder of any vintage to a 30D/40D - like IMAX vs. a 19" TV screen). This is not much of a factor with a rangefinder, where the viewfinder size is divorced from the sensor dimensions, since it is not a ttl view.

[marknorton]

The M8 met an internal deadline Leica set for itself, at the cost of some imperfections (but as a newspaper guy, and recovering procrastinator-perfectionist myself, I think it was a step forward to actually meet a deadline imperfectly than continue the culture of "perfect procrastination" that was dragging the company down).

 

I agree, we're further forward with the M8 than we would have been if they had held it back and endlessly fiddled with it - and still got it wrong.

 

I used to develop supermarket checkout systems for IBM and we always knew that the very first morning of operation in the very first live store was a better test than any amount of robotic testing in the lab.

 

Leica probably don't want to repeat the experiences of the DMR and M8 but, hopefully, they're amassing digital skills so that they don't have to. I'm encouraged by the new AWB, suggests they are building some real expertise.

[ho_co]

Whatever solution to the IR problem Leica might come up with, it won’t also solve the vignetting problem caused by the microlenses. Microlens-induced vignetting is different from both the ordinary vignetting caused by the lens itself or the cyan corner phenomenon.

 

... Sandy is absolutely right in that finding some filter material with a higher absorption factor wouldn’t be the answer. An absorption filter will also absorb some red, and a more effective filter would also render the red pixels less sensitive, increasing noise.

 

BTW, the dreaded internal reflections you mentioned would have been caused by a dichroic filter in front of the sensor – which would have been thin enough to avoid the issues with thick absorption filters, but was ruled out because of these reflections.

 

I lean more toward Andy P on this: Remember, just a few years before the M8 came out, a digital M was still "impossible" according to Leica.

 

Michael, I seem to recall that Leica rejected an absorptive IR filter like that of the DMR because of the sharpness reduction it would cause: Wideangles could produce shallow angles of incidence on the IR filter; refraction within the thicker glass would mean that nearby rays could end up in different pixel wells. Thus a stronger absorption layer could keep the thin glass. Couldn't concomitant problems with red absorption be compensated partially in firmware?

 

Your understanding is far better than mine. I'm just tossing out an idea for cross-fertilization.

 

--HC

[mjh]

I lean more toward Andy P on this: Remember, just a few years before the M8 came out, a digital M was still "impossible" according to Leica.

There’s a fine line between “impossible” and merely fiendishly difficult, and whatever Leica has been claiming, I think a digital M always belonged into the latter category. Same goes for a full-frame M now.

 

As far as I can tell – which doesn’t mean much –, Leicas reasoning leading to ruling out both a dichroic hot filter and a thicker/more effective absorption filter is still valid. I wish it weren’t so … It bears repetition to stress that the engineering decisions leading up the sensor assembly design, warts and all, were absolutely sound. It was a compromise, and an ugly one, but the best they could have come up with. Only at that point, the good people at Leica hypnotized themselves into believing everything was fine – some adjustments to the white balance perhaps, and nobody would even notice the issues this compromise design created …

 

I seem to recall that Leica rejected an absorptive IR filter like that of the DMR because of the sharpness reduction it would cause: Wideangles could produce shallow angles of incidence on the IR filter; refraction within the thicker glass would mean that nearby rays could end up in different pixel wells. Thus a stronger absorption layer could keep the thin glass. Couldn't concomitant problems with red absorption be compensated partially in firmware?

It could, in principle anyway, but there’s a catch. Correcting for vignetting in firmware means that the pixels near the edges and corners need to be multiplied by some factor to bring them up the level of the center pixels. This multiplication step will also multiply noise. On the other hand, it might be argued that with a more effective filter on the sensor, we would get rid of the filters in front of the lens, causing red vignetting / cyan corners for wide angle lenses. When we have to suffer some vignetting in the red channel anyway, at least for wide angle lenses, then maybe this is acceptable? Still, it should be quite tricky finding an absorption filter blocking IR effectively and at the same time leaving a healthy amount of red to hit the CCD.

[FastFashnReloaded]

PMA news: Leica to develop full-frame digital rangefinder camera news - Amateur Photographer - news, camera reviews, lens reviews, camera equipment guides, photography courses, competitions, photography forums

 

For the past two years people on these forums have been saying how good it is that Leica is using a smaller sensor since a larger one would vignette or other nasty things.

 

LOL

 

Two more years of pixel peeping at least here. The best thing I can see coming of this is that maybe some shrub will sell me his M8 for half price to buy something new.

[ho_co]

Michael, I agree on both points.

 

I think Leica has some of the most technically savvy designers around, and they will solve the problems.

 

But the solution may not include ending the need for lens-mounted filters.

 

Heck, that's no worse than some manufacturers do: With others, you buy a lens and then have to spring for a hood; with Leica, the add-on is the UV/IR-cut filter.

 

 

Fotokina 2008 will be where Leica launch their Digital Rangefinder Full Frame...

I'm with you on that! And the nifty thing is, we already (almost) own it!

 

BTW Andy, we all need to get the spelling down before the fair. It's "photokina," all lower case even sentence-initial.

 

photokina - Startpage photokina 2008 - koelnmesse GmbH

 

 

--HC

[Advertisement]

Hallo,

Look here: Leica M

[Guest stnami]

Yea I am waiting for that cheap B&W leica digital,,,,won't be long now

 

It would be ideal for you taking bnw pictures of the duke , .........hmmm still prefer the old silver 900SS