Is It Worthwhile Using The High Resolution/Pixel Shift Image Feature In The OM-1 Or Just Upsize The Files? Also, How Do They Compare To The 40mp X-T5 Files?

This is from an in-camera 80mp JPEG.  The pixel dimensions are 10,668 X 7776.  It really looks pretty
good overall, but anything that moved, even by a slight breeze, is slightly blurred.  I guess what
surprises me most is just how good the JPEGs are!  (click to enlarge)
OM-1; 12-40mm f/2.8 II PRO lens @30mm; 1/1000th sec. @ f/5.6; ISO 200

I've come to a conclusion that I probably should have years ago.  We put too much emphasis on the importance of the number of pixels on a camera's sensor—and that includes me!  This emphasis has even translated to camera companies devising ways to increase the pixel count of their sensors without actually increasing the pixel count of their sensors.  Of course, I'm referring to the 'high resolution/pixel shift' modes now present in many cameras.  However, it has been shown over and over again, that what you think you are getting with those additional pixels is tenuous at best.  In the many tests and demonstrations I've seen on the Internet as well as my own experiments testing my cameras, the actuality is there is very little difference in that fine detail you so dearly want to see between, say 20mp or 24mp sensors and 45mp sensors.  Yes, it is there but only if you look so closely (100-300%) at your image files on your monitor (it really doesn't show up in prints) that you are no longer looking at the actual image but only a tiny portion of it.  I ask you, is that the way you want your photographs viewed?  I don't.  I want you to look at my images in their entirety which is the way I intended when I pressed the shutter button.

But still question for me then arises,  "Is it worthwhile using the 'high resolution/pixel shift' feature in my OM-1 or just upsize my images in one of the advanced AI-based software programs when I 'think' I need more pixels.  Furthermore, how do the OM-1's 'high resolution/pixel shift' feature files compare to the 40.2mp files in the Fujifilm X-T5?

For those of you who don't care, thanks for reading this far.  You don't have to stick around for the rest of this post.  For those of you who may be interested to know, I have spent a bit of time over a few weeks attempting to answer this question for my own edification.  Are the 50mp handheld high resolution image files as well as the 80mp tripod mounted high resolution files that one can create using the OM-1 have any or substantial benefit over upsizing the standard 20mp OM-1 image files using either the Enhancement feature in Lightroom or Topaz Photo AI?  Then, just to expand my experiments a bit more, I thought I would then look at the upsized 20mp OM-1 files using the Lightroom Enhancement feature (a 2X enlargement which means 4X the pixels, i.e., 20mp>80mp) and compare them to the 40mp image files from my Fujifilm X-T5.

First, I must note that these limited tests were not done in a rigorous scientific way but in such a way that would allow me to get an idea about whether or not employing the in-camera high resolution features would benefit me in the future.  Also, I didn’t spend a lot of time examining all aspects of the files super closely but looked at them in closely at 100% on my monitor.  I looked for what could easily been seen and obvious.

In short, here is what I did:

I made identical images of several different outdoor scenes (since I am primarily a landscape and nature photographer) with both the 20mp OM-1 and 40mp X-T5.  In all cases, as one might expect outdoors, there was a slight breeze—the bane of taking multiple exposures and combining the files, but unfortunately almost always present.  The Fujifilm images and 80mp HR OM images were made with the camera mounted on a heavy-duty Gitzo tripod, raw + JPEG, manual exposure, base ISO, IBIS turned off, two second delay and AF-S.  You know the drill.  I tried to set my cameras to make the best quality files and eliminate any variables.

The 50mp handheld 'high resolution/pixel shift' images was made by taking the camera off the tripod and holding it steady at the same level and making the exposures creating a composition that was as close as I could to the tripod mounted exposures.  A one-second delay was enabled to eliminate any unsteadiness caused by me pushing the shutter button.

I made several exposures with each camera to ensure at least one of every series was in sharp focus (turned out they all were).  However, I did refocus between each exposure.

Over a couple of days I spent about an hour looking closely at all of the photos and doing some upsizing and downsizing.  I have drawn a few basic conclusions.

I would consider both the 50mp & 80mp high resolution images as failures as anything that was moving was slightly blurred.  I even waited for the light breeze to die down as much as possible before pressing the shutter.  The parts that were stationary, like tree trunks, piers, etc., were sharply rendered but leaves, pine needles, wetlands grasses that moved a bit in the breeze were not.  

It seems the 50mp handheld high resolution files did do a bit better job in rendering the moving items better than the 80mp version.  I suspect that is because the two features create the high resolution files differently.  

I would hesitate to use those two features outdoors if there is any wind at all unless I a bit of blur was part of what I wanted in the final result as we often want moving water to blur in our images.

I did find one anomaly in some of the 80mp files.  In the 100% crop of the image below, I could clearly see what only I can describe as a sort of an oversized moiré in the sections of the vertical stalks of the grasses behind the dock.  Obviously, this shouldn’t be there but it is.  I don’t know if you will be able to see it after the algorithms do whatever they after  posting these images.  To describe them as best I can, they are short pieces of the vertical stalks that have a color cast (blue) and look differently from what the stalks should look like.  That said, if you aren’t pixel peeping, you won’t see them.  Also, I don’t see these in the 50mp handheld high resolution feature. Again, most likely as a result of how the 80mp files are generated as opposed to how the 50mp files are generated.

This is the overall scene I made, tripod mounted, using the OM-1's 80mp high resolution
feature.  Upon looking at the image below, you may be able to see the anomalies. (click to enlarge)
OM-1; 40-150mm f/4 PRO lens @ 75mm; 1/1250th sec. @ f/5.6; ISO 200

This is an approximate 100% crop of the image above.  I don't know if you can see them after the
image has been through Google's posting algorithms, but on many of the stalks of the grasses
are anomalies which I can best describe as being a sort of a large unusual moiré-type pattern.  It
shows up sort of blue  in color.  Also, focus was on the piling.  Has anyone else seen this?
(click to enlarge)

The 20mp OM-1 files were upsized in Lightroom Classic to 80mp then downsized in Topaz to 40mp to compare against the X-T5 files.  (I didn't try Topaz Photo AI for upsizing but maybe I should have) The upsized OM-1 files, to me, looked as sharp and detailed as the X-T5’s files.  In fact, they looked excellent as 80mp files also.  I wouldn’t hesitate to employ that feature if needed.

The OM files 20mp that were upsized to 80mp in Lightroom were also compared to the stationary subjects in the in-camera 'high resolution/pixel shift' 80mp files.  I ignored the parts that were blurred by the wind.  I think the Lightroom upsized files looked a little better.  They just looked ‘cleaner’ but I don’t know how to describe what that means.

I also compared the 20mp OM files that were upsized in Lightroom directly to the 50mp 'high resolution/pixel shift' handheld files.  I liked the upsized files better in this test as well.

All of this said, they are all close and if you don’t pixel peep, you won’t tell any obvious differences except for the slight blurring of moving subjects in the 'high resolution/pixel shift' files.

Also, I think the 'high resolution/pixel shift' files from the OM need additional sharpening over what I would give the native 20mp files.  Maybe a little more of the texture slider as well.  But that may be just my preference.

My thoughts right now are that I will just shoot the 20mp normal files and if I need a larger file for something, I’ll just upsize it.  The difference in detail is so minute, unless you are looking very intensely, you probably won’t seen any differences.  At least not a normal person like me. (My wife may disagree with my self-assessment of being normal! lol)  You "uber vision" photographers out there might see it differently.

As far as the 40mp Fujifilm files (demosaiced using DXO to squeeze out as much detail as possible) as compared to the upsized Olympus files, they look comparable as well.  Not much difference.  If you look at the files at 100-300%, the differences can seen, but in many cases, you really have to search for them.  Now, remember, I'm talking about the images I made.  You may make different types of images and your results may be a bit different.

Also, I wanted to mention that there is value in using JPEGs when using the 'high resolution/pixel shift' feature, as well.  To me, they look as good as the raw files for ‘average’ subjects, i.e., nothing that needs an extraordinary amount of manipulation in editing.  In some of my testing the JPEGs looked better (yes, I confirmed that in several images made of different subjects) than I could make the raw files look. However, they do seem to have a lower tolerance for overexposure, so be careful.  If you want to experiment as I did I would recommend shooting both raw and JPEG and look at both and see which one looks better to you.  It can vary by light, contrast, subject, etc.

Additionally, using the 'high resolution/pixel shift' feature will result in the reduction of noise in the files as well as produce greater dynamic range as measured by Bill Claff on his site Photos to Photons.  An increase of about a stop of dynamic range is gained according to his testing.  Like stacking images in Photoshop, noise is reduced as well.  Something to keep in mind.  So there is some benefit to using that feature.

I think I’m pretty satisfied as to how I will proceed in the future.

My overall conclusion is that I can be confident using the 20mp file in the OM-1 and upsizing without visible penalty in Lightroom if necessary.  In reality I don’t find much advantage in the 'high resolution/pixel shift' feature or the 40mp files.  I see a minor increase in very fine detail, again if you don’t look at the differences at 100% or 200% they won’t be obvious.  The differences are there, but you have to really look hard to find them and if I have to look that hard, I’m not really looking at the overall photograph as photographs are meant to be viewed.  But—I didn’t conduct these experiments to draw a definitive conclusion for all of you.  I just experimented for my own knowledge and am sharing what I found in my particular compositions.  I recommend you do your own testing and drawing your own conclusions.  

I do these experiments for the simple reason is that I want to know as much about my cameras' capabilities, operation, functions, etc., as I can.  The more I understand, the better chance I have at making better images using these tools and the lower the chances of being disappointed with the images I make.

Finally, as I’ve written before, my head says I don’t need 40mp or 50mp or 61mp, but my heart says it “yes!”  I want all those extra pixels.  I can’t rightly explain it but I suspect many of you feel the same way.  So there’s that.

Join me over at my website, https://www.dennismook.com.  

Thanks for looking. Enjoy!  

Dennis A. Mook  

All content on this blog is © 2013-2023 Dennis A. Mook. All Rights Reserved. Feel free to point to this blog from your website with full attribution. Permission may be granted for commercial use. Please contact Mr. Mook to discuss permission to reproduce the blog posts and/or images.