16MP sensor improvements?

Started Apr 24, 2013 | Questions thread
Anders W Forum Pro • Posts: 21,468
Re: Rather significant, actually, although YMMV applies
1

jim stirling wrote:

Anders W wrote:

jim stirling wrote:

Anders W wrote:

Right. So why don't we continue exactly where we left off three months ago so that we don't have to repeat ourselves. Here's what I said in my latest reply to which you have yet to respond: http://www.dpreview.com/forums/post/50752068

Anders, you can quote all the abstract tests you want

I didn't quote any tests in my latest reply. I asked you some questions via the post from three months ago that I linked to. You didn't answer them. Enough said.

That thread reached its maximum post level  3 months ago Anders you had nearly 40 posts in it .For those of us who do not post as regularly it is quite difficult to go back a week or even a few days later and respond when the rate of posting in the mFT forum is so high.

I didn't complain about the fact that you didn't respond three months ago. I complained about the fact that you didn't respond when I asked you now.

You are asking me to take seriously a comparison between a (recently redesigned) 50/1.4 normal prime for FF at its peak aperture (f/5.6) with a 12-50 kit zoom for MFT with a so-so reputation for resolution and at an aperture already beyond the peak of what MFT can do at its best (due to diffraction). In your attempt to convince me that this is reasonable, you are using data from SLRGear for which no known basis for cross-format comparisons exist. The comparable data from LensRentals are as follows (lp/ih for center/average across the frame based on unsharpened output from RAW files) are:

50/1.4 at f/5.6 on D800: 1073/889

12-50/3.5-6.3 at 25 mm and f/5.6 on an E-M5: 830/670

The lens that should have been used for proper comparison with the 50/1.4 is the 20/1.7 or the 25/1.4 at their peak aperture. The figures for these are:

20/1.7 at f/4 on an E-M5: 1075/880

25/1.4 at f/4 on an E-M5: 980/850

Sources:

http://www.lensrentals.com/blog/2012/05/wide-angle-micro-43-imatest-results

http://www.lensrentals.com/blog/2012/05/standard-range-micro-43-imatest-results

http://www.lensrentals.com/blog/2012/05/wide-angle-micro-43-imatest-results

In an admittedly belated answer to your questions:

The 12-50 on the E-M5 is shooting at 1/60 sec;   f/5.6; {a shutter speed 2 stops slower than that on the D800}    ISO 1600 at 25mm, I used a central crop where it performs very well indeed.

From the SLRGEAR review of the 12-50

“” the lens offers excellent results for sharpness, corner shading and distortion. "

Their graph for the 12-50 at 25mm and F5.6 shows very good sharpness levels in the file  particularly good at the central area

The 50mm on the D800 is shooting at 1/250 sec;   f/5.6; { a shutter speed 2 stops faster than the E-M5}    ISO 6400. I used a central crop where it performs very well indeed.

From the SLRGEAR conclusion

"" it isn't until ƒ/5.6 that I would say the lens is completely sharp from corner to corner."

The car is not fully contained in the DOF of the 50mm on the D800 which is why I used a central are from both cameras, thus both areas are in focus both areas contain a high level of detail and black areas to demonstrate noise levels. I never claimed that they were at the same DOF . The central portion {which I used} in all sample files is sharp, in focus and at a good setting for each lens to deliver excellent image quality

The point is that the area you chose is not equally in focus in both shots. Apart from the greater DoF of the E-M5 shot, the focus is clearly placed differently, i.e., closer to the camera in the case of the E-M5 shot, which keeps the area you selected within the DoF but certainly not at peak focus.

I am not being unreasonable in any way in asking you to download a couple of RAW files and process them " all  "to get their best results , and posts 100% crop comparisons .These Focus Numerique samples have  the most consistent exposures of any of the review sites. In fact I would be happy for others to do the same and see a spectrum of results.

Of course you are being unreasonable when you ask me to be convinced by the clearly unsuitable data (the images from Focus Numerique) that you selected. I have already shown you (and others) what the situation is actually like in the comparison between the E-M5 at ISO 1600 versus the D800 at ISO 6400 two times around, i.e., here:

http://www.dpreview.com/forums/post/50310594

and here (click on "show signature" if the images don't show)

http://www.dpreview.com/forums/post/50462878

in both cases using data (images) that you rather than I had initially selected. What's wrong with those two tests and in what way would the clearly unsuitable source material you have chosen for your third attempt be any better?

I am simply asking you to download a couple of images and process them to their best , your download , your processing. I think we both know why you will not be doing it.It is not possible to sharpen the E-M5 images anymore to squeeze out detail without impacting the noise levels, and conversely while you can make the images less noisy it is always at the expense of detail.

It is the simplest thing in the world to categorically prove me wrong by showing me any two images comparisons from these cameras { fairly processed which I know you would do } at 2 stops apart where once processed the E-M5 has more detail and less noise. Of course the problem with that is you cant do it

I guess your stringent scientific approach doesn't stretch to experiments :-). Much easier to quote data sources none of which are impeccable and some of which are contradictory rather than simply process the images and look at them. The facts are that due to the extra detail { which I assume you do not dispute} it is a one click adjustment to to provide an image from the D800 at two stops higher that will have more detail and at worst the same degree of noise, giving the D800  a 2 stop detail/DR advantage  at base ISO along with an at worst same noise more detail file at 2 stops apart.

While not one of the test sites has 100% controlled lighting , Focus Numerique has by far the most consistent exposure settings between cameras. I chose an image with a good combination of details and textures.I used the RAW files,I used a central area of the image to reduce the impact of different lenses ,I processed all the images to achieve the best balance between detail and noise and I posted the results here at 100% crops. I provided links to the actual RAW files so that you or anyone else can give the images their best shot and post  the resulting images.

Jim

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +28 more
Post (hide subjects) Posted by
BJL
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow