Yet Another Rumour

[gdewitt]

Advertisement (gone after registration)

In a sense a full frame is necessary if one wants to push resolution beyond 10 Mp. As it is the pixel size on the sensor, when adapted to the diffraction limit of lenses is equal to the current Mp sizes of the different sensors, i,e approx 8 for APS-C, 10 for 1.3 and 16 for 35 mm. In other words, building a 1.3 20Mp sensor would not result in more resolution, just in two adjacent pixels sending the same information. So if technology would give us the option of a larger sensor in the future M9 (or ZI digital RF) it would result in a 16MP camera.

 

Diffraction varies with aperture, no? The M8 is diffraction limited somewhere between f11 and f16, I think. So increasing the pixel density would make the diffraction limitation appear at some wider aperture than now, therefore limited use of small apertures, but not limiting resolution, per se.

[Riley]

I thought the argument for a full frame sensor had been disposed of by the experts as being unnecessary on a 35mm camera.

 

there are some ways of tackling this

the first is do what Leica/Kodak did with lenses on the sensor

 

the next would be use longer lenses anyway, by that i mean longer register

for Leica that would have meant using only the R mount lenses

then the divergent light issue is mostly gone from 50mm and up

lenses like the Voigtlander 12mm are reported to have like 60 degree angular light

this should be impossible really, Im amased that Guy was able to make the 15mm work on the M

 

to deal with wide, ultra-wide

a fresnel lens mounted close to the sensor with a more favourable angle of acceptance

a sensor that in itself is less reflective would also help a good deal

 

in the balance, it would be more expensive than M8 to engineer

 

********************************************************************************************************

 

a four thirds would be an all round easier solution

the lenses are a lot more telecentric anyway

there are no old lenses to account for, so there are less engineering issues

it would enhance the FT range in status

[rvaubel]

The 1.3 format for a digital rangefinder may represent the "sweet spot" in the resolution arena. By pushing the technology of the sensor/IR filter/offset microlensing equation, Leica has managed to produce a camera that is comparable to the full frame 5D except for speed. Maybe, by refining the technology eventually a 12MP camera with fairly clean 3200 ISO and no filters could be produced. That would be a goal that would be almost impossible to produce if a full frame sensor was also required. I think that full frame sensors will remain the domain of the DSLR with their much longer back focus distances. But thats OK because the M lenses result in a much more compact format

 

Rex

[Riley]

i absolutely agree Rex

if you want M glass, it has to be cropped sensor

did they get it right ?

yes i think so

 

will 12Mp make a difference

probably not

you would need around 18 to make a substantial difference is my feelin

not to say that isnt possible

[rvaubel]

i absolutely agree Rex

if you want M glass, it has to be cropped sensor

did they get it right ?

yes i think so

 

will 12Mp make a difference

probably not

you would need around 18 to make a substantial difference is my feelin

not to say that isnt possible

 

I don't think 12MP will make any real difference over 10MP. It's just sort of the end of the road, technically speaking, for a 1.33 sensor. And it would allow Leica bragging rights over the 1.5, 1.6X sensor crowd. Basically a marketing thing. Personally I would prefer a higher ISO and less noise.

 

And please do something about the IR contamination. Not that I am not reasonable happy with the M8. I'm just thinking 3 or 4 years down the road when the M9 comes out.

 

Rex

and I'm not waiting

[Riley]

just on iso

what ever happened to low iso ?

seems there is a trend to raise the lower iso levels

[marknorton]

Advertisement (gone after registration)

Keep in mind that the sensor knows nothing about ISO. It only knows about absolute light exposure levels and it has its own saturation point, where an increase in exposure does not lead to an increase in signal output. At that point, or a bit less, you want to get peak white (RGB 255, 255, 255) output and circuitry/software are designed to do just that. The effective base ISO is established from the exposure (shutter speed/aperture) required to achieve it.

 

If you increase ISO, you are applying gain in steps of x2 to the measurement process so that peak white is reached at lower exposure levels, increasing the apparent ISO but in fact the sensor has not changed. You also lift the noise floor through this gain. On the other hand, it doesn't make sense to reduce ISO by reducing the gain in the measurement process to achieve an apparent reduction in sensitivity. If you do, you will no longer be able to get peak white out of the thing because the sensor will saturate long before your image is fully "exposed".

 

If you want to run a digital camera at a lower ISO, so that you can use slower shutter speeds and wider apertures, use neutral density filters. An ND-8 filter will take you down to ISO 0.6

[jaapv]

True-but I'm sure a number of other nasties would pop up when "over-resolving"a sensor.

 

If that were true, then there wouldn’t be any moiré in images taken with the M8, would it? Because the lens would take care of any antialiasing necessary for a 10 MP sensor.