I have been watching the detail discussion progress on the Pentax vs EM5.2 thread with interest as it derailed into a D810 vs EM5.2
Interestingly enough there have been people coming down on both sides... I think there may be a bit of confusion here about what constitutes detail in the comparisons. Simply put, the EM5.2 resolves to a greater level of detail but lacks the advantage of false detail.
Here are a few crops to demonstrate, with comments. View at full size.
[ATTACH alt="Here you can easily see the type is a fraction easier to see on the full size Olympus. They are both pretty bad however look at "side Walls and Running" 4 lines up on the left. There are plenty of words that are hard to read on both, however there are far more clues in the Olympus file."]1043519[/ATTACH]
Here you can easily see the type is a fraction easier to see on the full size Olympus. They are both pretty bad however look at "side Walls and Running" 4 lines up on the left. There are plenty of words that are hard to read on both, however there are far more clues in the Olympus file.
But of course this makes sense. You can see all 9 lines described to the end of the chart here, while the Nikon has smeared two together. In fact this smearing happens much earlier in this sample, while they start to blend in the Olympus file they do so evenly.
[ATTACH alt="And of course how could an image that can clearly "see" the concentric circles have poorer resolving power? Here you can see they again gracefully move out of resolution while the Nikon brings them in and out, creating and artefacts."]1043521[/ATTACH]
And of course how could an image that can clearly "see" the concentric circles have poorer resolving power? Here you can see they again gracefully move out of resolution while the Nikon brings them in and out, creating and artefacts.
[ATTACH alt="Ahhh, but you can clearly see the lines here... Not really. You can see "lines", they are most likely not the right number of lines. Here we see the advantage right beside the disadvantage of a lack of an AA filter. But look very carefully at the wall above the womans head... You can see vertical lines... while the Nikon has become a patchwork... How could the Olympus "see" these lines and yet have worse resolving power? It cannot. The detail further into the Nikon is false, but just so happens to work in that patch."]1043522[/ATTACH]
Ahhh, but you can clearly see the lines here... Not really. You can see "lines", they are most likely not the right number of lines. Here we see the advantage right beside the disadvantage of a lack of an AA filter. But look very carefully at the wall above the womans head... You can see vertical lines... while the Nikon has become a patchwork... How could the Olympus "see" these lines and yet have worse resolving power? It cannot. The detail further into the Nikon is false, but just so happens to work in that patch.
Here agin you can apparently see more detail in the Nikon. However everywhere where this detail starts to appear so do artefacts. Look at the bottom left corner of the painting frame. You can see the vertical lines. IN fact you can see a verticality across the entire frame. The Olympus image knows ther are there, but they are gracefully moving out of resolving power without and false detail "filling in". The girls face, again we see a patch of false information while the Olympus shows at least the direction of the strokes.
Those four images bring me to the last two. If the Olympus is showing a fraction more detail than the Nikon, surely it should show up? Well we lose the ability to reproduce false detail, and at times it is a good exchange, however we get more "real" detail which should become apparent with the right processing.
Here are two crops showing my processing of the EM5.2 file vs the D810 with the mark 2 downsized to 5000 pixels in height. These are now at 100%. I think most people can see the actualy detail advantage, as slight as it is.
So if you get slightly more detail than a 36mp file, with no aliasing, moire or artifacts (barring the obvious limitations of the technology and advantages of a native 36mp) what resolution would you say this leaves? I thought 40mp sounded about right
[ATTACH alt="Now at the same viewing size, is there any more detail int he "nikon" hair than the Olympus or are we seeing the perception of detail? While the moving in and out of resolution of the lines above would suggest it couldn't be more, simply a different "look"."]1043524[/ATTACH]
Now at the same viewing size, is there any more detail int he "nikon" hair than the Olympus or are we seeing the perception of detail? While the moving in and out of resolution of the lines above would suggest it couldn't be more, simply a different "look".
Hopefully all this very fine detail will help. Look closely at the left of the feathers where it all becomes very pale. Notice all the direction you can see, this carries through the entire feather... never mind the green beneath.
So to get a sense of how much resolution you can get out of the EM5.2 you have to reconsider how you process and downsize to your final output as these files are very flexible.
What do you think, can the EM5.2 match the D810 for real detail? Or do you prefer false details?
--
“You don’t take a photograph, you make it.” -Ansel Adams
blog.alatchinphotography(dot)com
