[falconeye]

Well, exactly as I anticipated, my point of view is a minority view :ugh:

Why I wrote "Let me offer a different opinion" in the first place


No need to repeat my point, I think I presented my humble opinion already.


Therefore, let me just make a comment regarding this quote:

Like I said, they can make any corrections they like in JPEG, but leave them out of RAW please.

The RAW data have been heavily processed when you see them for the first time: Bayer demosaicing, hot pixel removal, color profile conversion (e.g., for gamma 2.2) etc. The actual raw data on file aren't altered but the converter does all this in order to make the image visible to the user. Adding correction for distortion, vignetting and CA to the process if done correctly wouldn't make such a big difference.

Just one example: the Bayer filter colors don't correspond to sRGB tristimulus standard values and are always and silently corrected by the converter. Nothing anybody ever complained about.


But to concede an argument to the majority in this discussion: If a lens requires electronic correction then it has to excel in some other aspect which wouldn't have been possible or affordable otherwise.

[nostatic]

For the target audience, do you think they want a lens that is 3x as big, 3x as heavy, and 3x as expensive as the Panny kit zoom? Because if you want everything that some of you are asking for, that's what it would take.

Some people don't want to haul around a huge lens, especially on a compact camera. But they do want a reasonable FOV. Enter some creative engineering and software correction. Of course the 800lb gorilla in the room is that you have software correction going on pretty much everywhere through the system. And much of it you cannot have control over. So is giving up "control" in order to get a small/light/decent performing lens worth it? Well, evidently given the G1/GH1 (and soon GF1) sales, the market seems to think so.

[johnmflores]

Overall, your thinking is old thinking (which isn't necessarily a bad thing). But the engineering task is to provide a best possible image, not a best possible pinhole lens emulation. To enforce an arbitrary constraint to stop optimization at some arbitrary boundary (the raw data in your case) is just poor engineering. Believe me, the professionals (lens designers) know it better and thank Panasonic to pave the way (and Leica actually).

+ 1 for sure. I never understood the ascetic devotion to the film plane. Leica/Kodak handling of edge-conditions is considered ok yet Panasonic's handling is blasphemy? Is this because one is physical and one is not? What if two competing systems - one physical and one not - produced identical results, the same exact raw file? Would there still be such an outcry?

Maybe I'm ok with this kind of digital manipulation because I'm an engineer by education and can appreciate the cost-effective elegance of the solution. Or maybe it's because I'm not as much a craftsman as other here are. Whatever the case, I can't afford the Leica solution and can afford the Panasonic one!

[*isteve]

+ 1 for sure. I never understood the ascetic devotion to the film plane. Leica/Kodak handling of edge-conditions is considered ok yet Panasonic's handling is blasphemy? Is this because one is physical and one is not? What if two competing systems - one physical and one not - produced identical results, the same exact raw file? Would there still be such an outcry?

The judgement I made was between micro 4/3 and normal 4/3.

The hypothetical situation that you and Falk describe is NOT the situation we are talking about in reality. In reality, there ARE visible issues with the RAW file which cannot be removed, including visible noise in the corners of the frame which have had to be lightened because of heavy vignetting.

This is done to cope with a fundemental issue of using a large sensor and small registration distance.

Maybe I'm ok with this kind of digital manipulation because I'm an engineer by education and can appreciate the cost-effective elegance of the solution. Or maybe it's because I'm not as much a craftsman as other here are. Whatever the case, I can't afford the Leica solution and can afford the Panasonic one!

I dont remember saying either solution was perfect. All I know is that on a 4/3 or APS SLR I dont have the same issue.

So even if I do manipulate the image, which one will be better? The one taken with the 4/3 camera and the 14-40 Oly lens, or the one taken with the micro 4/3 camera and the Panny 14-40 lens.

If you are going to apply your theory that its OK to digitally enhance images, then I say you still benefit from the best possible input to start with. If you compromise the input (and micro 4/3 is just such a compromise) then you WILL affect the quality of the result.

[*isteve]

But to concede an argument to the majority in this discussion: If a lens requires electronic correction then it has to excel in some other aspect which wouldn't have been possible or affordable otherwise.

Or, that a system that requires electronic correction because of its own, inbuilt limitations, will never be as good as a system that imposes fewer limitations.

I never had anything against digital manipulation, I do it all the time, I have something against a system that forces quite heavy manipulation as necessity just to level the playing field.

But if you can achieve enhancements to a system using electronic correction, then I say fine, but I would like to be able to apply it myself at the appropriate time in the post processing cycle.

[*isteve]

For the target audience, do you think they want a lens that is 3x as big, 3x as heavy, and 3x as expensive as the Panny kit zoom? Because if you want everything that some of you are asking for, that's what it would take.

Some people don't want to haul around a huge lens, especially on a compact camera. But they do want a reasonable FOV. Enter some creative engineering and software correction. Of course the 800lb gorilla in the room is that you have software correction going on pretty much everywhere through the system. And much of it you cannot have control over. So is giving up "control" in order to get a small/light/decent performing lens worth it? Well, evidently given the G1/GH1 (and soon GF1) sales, the market seems to think so.

I am now more firmly convinced that the compact DSLR still has some advantages. It is only slightly bigger (eg. "deeper") but it appears to be considerably less compromised.

What other people buy is up to them. I have always been a supporter of the micro camera idea, but under the circumstances I dont think I will bother just yet.

I have to say that I'd love a small camera body with K mount lenses and a sensor no smaller than APS-C. I could sacrifice the optical VF if that's what needed to make the camera body small. I'm personally not a very small person and I suppose my hands are at least the average male size, but I've found that I often prefer to take the smaller K200D and not the slightly bigger K20D with me. And as I'm slowly building a collection of decent K mount glass, it would be great to put them to work on an EVIL camera.

[eurostar]

There was a very old Pentax prototype of a EVIl camera with true K mount. The mount was at the end of a cone shaped "ring" where many controls were located. I guess that not-too-fast superwideangles could still be designed for such an arrangement by simply having half of the lens protrunding behind the mount, without needing a new, short registered mount.

[*isteve]

There was a very old Pentax prototype of a EVIl camera with true K mount. The mount was at the end of a cone shaped "ring" where many controls were located. I guess that not-too-fast superwideangles could still be designed for such an arrangement by simply having half of the lens protrunding behind the mount, without needing a new, short registered mount.

You may as well stick a converter to use SLR lenses on a micro camera - the issue is you have not really reduced the size by much.

The challenge for micro and rangefinder cameras appears to be the low angle of incident light hitting the edges of the sensor. In my opinion, this will make it quite difficult for such cameras to truly compete with SLRs for ultimate IQ, but if you are not a purist and you dont mind software being used to mask the issue at the cost of some post processing flexibility, then the Panasonic solution is cheap at least.

The other solution of course is to build a camera with a well corrected fixed lens designed specifically for that sensor (like Sigma and Leica have done).

But what I cannot for the life of me figure out is why there is no sensor size somewhere between P&S size and 4/3.

[gazonk]

But what I cannot for the life of me figure out is why there is no sensor size somewhere between P&S size and 4/3.

I was just going to write the same!
There's a large gap between the largest P&S sensors in the best P&S cameras, e.g. the 1/1.63" sensor in the LX3, and the 4/3 format.

Imagine a camera with a 1:3 crop sensor with 10mp, that would give it approximately half the pixel density of the LX3.

[falconeye]

I was just going to write the same!

I was puzzled as well.
But look at, e.g., the Leica D-Lux4 (the LX3 mentioned above) and Leica XR1 (APS-C). The size difference is small already (124 x 59.5 x 32 mm vs. 109 x 59.5 x 27 mm).

And the XR1 is APS-C and not even FourThird. A hypothetical FourThird XR2 may be almost indistinguishable in size from the D-Lux4.

With an intermediate sensor size, the gain in miniaturization may just not be large enough to justify the loss in image quality.


Therefore, I believe a better question is why almost all FourThird or APS-C-sized sensor cameras are so bulky, E-P1 and NX included. The Pentax 110 showed that P&S-sized cameras with a 22mm sensor/film are possible, even as a system camera with exchangeable lenses.

[gazonk]

I was puzzled as well.
But look at, e.g., the Leica D-Lux4 (the LX3 mentioned above) and Leica XR1 (APS-C). The size difference is small already (124 x 59.5 x 32 mm vs. 109 x 59.5 x 27 mm).
.

Yes, but the LX3/D-Lux4 has a wide-angle fast aperture zoom... (24-60, not exactly "super zoom", but it's almost the same as my DA16-45 which I find very useful in range) ! The EP-1 is also small when you mount a pancake on it. That's what the XR1 is: An APS-C camera with a fix-mounted pancake lens...

[falconeye]

That's what the XR1 is: An APS-C camera with a fix-mounted pancake lens...

Maybe, but the E-P1 is larger than the XR1, even without a lens at all (120.5 x 70 x 35 mm vs. 124 x 59.5 x 32 mm which is 25% larger in volume). And still it has the significantly smaller FT sensor. I am sure, the XR1 could be even narrower if made by a larger engineering team.

I tend to insist: current mirrorless system cameras are bulky. That's the problem, not that there is no intermediate sensor size.

[MikePerham]

I would snap up a Pentax compact like the Leica X1 with an APS-C sensor, fixed 24mm lens. This camera does not have to be an EVIL ...as a carry around camera I would be quite happy with a fixed 24mm. In fact Pentax could use the same sensor as the K-x that has the same pixel density as the Leica X1 and I am sure could compete with that camera in IQ at a considerably lower price.

[gazonk]

I tend to insist: current mirrorless system cameras are bulky. That's the problem, not that there is no intermediate sensor size.

I agree - I'd say that the current DSLRs are still bulky, too. I bet we'll eventually see FF DSLRs as thin and low as the Pentax ME - and less wide! There's no absolute physical reason the sensor, even with shift mechanism, should take much more place than the film back plate, and no reason electronics should take more place than the film itself. It's just a matter of miniaturisation - i.e. time...