X-Trans 26Mpix vs 40Mpix, differences in resolving fine details

BeatX · Apr 16, 2024

Hello,

Inspired by a user from FM Miranda forum, I conducted a simple test to see clearly what are rally differences between X-Trans IV and V sensors.
I had in mind real world usage with using one of best possible lens for Fuji X, and I was focusing only to spot any differences in resolving fine details.
X-Trans V 40Mpix sensor (in theory) should benefit more in resolving fine details, than previous X-Trans sensors generations.
For this reason it should (in theory) let users crop more in post, and it should allow to print in bigger size than native A3 (@300dpi) from 26Mpix sensor.

I used for this test my wife X-T30II and my X-H2, plus one of the sharpest standard primes available for Fuji X system - Viltrox 27/1.2
Sorry for poor align of both test shots.. I was trying to do my best, to put both cameras on my tripod and center them ideally - but it's not as simple as it seems

Files have default C1 settings, I didn't touch anything apart from setting identical WB in both files.
Last image is original X-T30II file upscaled in C1 to 40Mpix

X-H2

X-T30II 26Mpix

X-T30 40Mpix

Crop of both files from X-H2 and upscaled to 40Mpix file from X-T30II
X-H2

X-H2

X-T30II @40Mpix

X-T30II @40Mpix

Link to raw files:

https://drive.google.com/file/d/1lUo8tiSyfoJXlhtIeYAOsiH3RLwH1xzq/view?usp=sharing

There are very tiny differences between those sensors - I admit.
But it is worth for upgrading? I'm not so sure.
For me those differences are symbolic, and they are seen only on 100% magnification (how many times I've read here and there, that pixel peeping is wrong?)

40Mpix sensor creates problems like bigger file size, and diffraction starts earlier - offering in return almost no benefits in real usage.
My theory for such tiny differences between sensors, is because lenses are weak chain.
I think, that lens which could "infuse/resolve" more resolution from such high pixel density sensor like X-Trans V 40Mpix, probably would be very difficult to design and produce.
And it would be for sure gigantic, heavy and very, very expensive - hardly noone would buy it.

My conclusion: 26Mpix is sweet spot for aps-c size sensors

Now, when I know exactly how very little "improvement" I have in fine details from my X-H2 compared to my previous X-S10, I would definitely choose X-H2s (if I would stay in Fuji X system)

--
My gallery: https://www.flickr.com/photos/maciej_k/

GMacF · Apr 16, 2024

Thanks for sharing your findings. I too have done testing on and off and my conclusion is pretty much the same - there's a difference but it's tiny.

There are perhaps tangible benefits when it comes to printing however I've never had any issues with printing up to my usual size (A3+) with the 26mp sensors.

I always said, if someone upgrades to the newer sensor bodies just for the extra megapixels then they are going to be disappointed. I did upgrade to the X-T5 and in all honesty the 40mp was probably 4th or 5th on my list of features.

a_c_skinner · Apr 16, 2024

Lenses and most importantly technique. Even if the lenses are first rate the latter is crucial. Even so you'd expect the difference to be very modest as the linear resolution only improves by 24%.

Erik Baumgartner · Apr 16, 2024

Default import settings for both isn’t a fair comparison as you’re not going to be getting the most out of the 40MP files. As the pixel density is far greater at 40MP, some fine-tuning is going to be required for the best results - any sharpening applied at 40MP needs to act directly upon far finer detail.

toktik · Apr 16, 2024

I have an X-H2 as well as an X-H2S and X-T3. I generally agree with your findings. I believe one substantive benefit of the 40MP sensor is that it is less likely to suffer from Moire than the lower resolution sensors. X-Trans does a good job of reducing Moire, but it still occurs sometimes. My understanding is that Moire will be much less likely to be an issue with a higher megapixel sensor. This holds true with the GFX series as well: the appearance of Moire is much less with the 100MP sensor vs. the 50MP sensor.

Krummj · Apr 16, 2024

I was not so impressed with the resolution on my new XH2 compared to the XH1... until I bought the Fuji 18 1.4 WR. Then it's like the sensor can finally breathe. With my other lenses (35 1.4, 23 f2, 27 2.8) they looked a little soft at 200 percent, good at 100 percent on my Mac Studio Display. My 18 1.4 shots look sharp at 200 percent, across the frame. Next I will likely get the 33 1.4, for the same reason. These top lenses make the Fuji much closer to my Nikon Z7.

Morris0 · Apr 16, 2024

For my taste after opening both images, I like the X-T30 II best.

Never upscale as the algorithm is what you are looking at. If you want to crop to show what's there, crop from the original in both cases.

Morris

BeatX · Apr 16, 2024

Erik Baumgartner said:
Default import settings for both isn’t a fair comparison as you’re not going to be getting the most out of the 40MP files. As the pixel density is far greater at 40MP, some fine-tuning is going to be required for the best results - any sharpening applied at 40MP needs to act directly upon far finer detail.

It is possible that it is as you write. Unfortunately, I lack the knowledge and competence on how to effectively sharpen files from 40 megapixel X-Trans V sensor. Could you download my raw files, sharpen them and show if the differences have increased between those two examples, when they have been processed in Your way?

Erik Baumgartner · Apr 16, 2024

BeatX said:
Erik Baumgartner said:

Default import settings for both isn’t a fair comparison as you’re not going to be getting the most out of the 40MP files. As the pixel density is far greater at 40MP, some fine-tuning is going to be required for the best results - any sharpening applied at 40MP needs to act directly upon far finer detail.

Click to expand...

It is possible that it is as you write. Unfortunately, I lack the knowledge and competence on how to effectively sharpen files from 40 megapixel X-Trans V sensor. Could you download my raw files, sharpen them and show if the differences have increased between those two examples, when they have been processed in Your way?

BeatX · Apr 16, 2024

Erik Baumgartner said:
BeatX said:

Erik Baumgartner said:

Default import settings for both isn’t a fair comparison as you’re not going to be getting the most out of the 40MP files. As the pixel density is far greater at 40MP, some fine-tuning is going to be required for the best results - any sharpening applied at 40MP needs to act directly upon far finer detail.

Click to expand...

It is possible that it is as you write. Unfortunately, I lack the knowledge and competence on how to effectively sharpen files from 40 megapixel X-Trans V sensor. Could you download my raw files, sharpen them and show if the differences have increased between those two examples, when they have been processed in Your way?

Click to expand...

Yeah sure, but it will be with Lightroom sharpening, not C1.

No problem. Hardly noone is using other software than LR, so this comparison should be interesting.

Erik Baumgartner · Apr 16, 2024

BeatX said:
Erik Baumgartner said:

BeatX said:

Erik Baumgartner said:

Default import settings for both isn’t a fair comparison as you’re not going to be getting the most out of the 40MP files. As the pixel density is far greater at 40MP, some fine-tuning is going to be required for the best results - any sharpening applied at 40MP needs to act directly upon far finer detail.

Click to expand...

It is possible that it is as you write. Unfortunately, I lack the knowledge and competence on how to effectively sharpen files from 40 megapixel X-Trans V sensor. Could you download my raw files, sharpen them and show if the differences have increased between those two examples, when they have been processed in Your way?

Click to expand...

Yeah sure, but it will be with Lightroom sharpening, not C1.

Click to expand...

No problem. Hardly noone is using other software than LR, so this comparison should be interesting.

Will do when I get home in a few hours.

kdarwin-photo · Apr 16, 2024

Rather than upscaling 26 MP to 40 MP, I'd have thought it better to go the other way, shrink the 40 MP down to 26 MP.

Or better yet, shrink them both to 12 MP or less based on what will actually get printed or displayed.

Does a 40 MP image or 26 MP image look any different when rendered as 12 MP?

Erik Baumgartner · Apr 16, 2024

kdarwin-photo said:
Rather than upscaling 26 MP to 40 MP, I'd have thought it better to go the other way, shrink the 40 MP down to 26 MP.

Then you're throwing away 14MP of image resolution from the 40MP file.

kdarwin-photo said:
Or better yet, shrink them both to 12 MP or less based on what will actually get printed or displayed.

You can print and/or display 40MP.

kdarwin-photo said:
Does a 40 MP image or 26 MP image look any different when rendered as 12 MP?

Yes, the 40MP image will have fewer demosaicing errors and less moire/aliasing.

Ap0ll0n · Apr 16, 2024

To Erik's point, both processed with DXO Pure Raw and zoomed in at 50%. 26MP on the left.

After processing with (latest) Topaz Photo AI:

To me, the 26MP clearly looks over-sharpened and inferior. Topaz helps narrow the gap a bit but personally if I care a lot about IQ, at least when when zoomed in, I would stick with 40MP. To each their own of course and thanks for posting.

The full DXO pics below for reference.

26MP

40MP

--
Apollon

Apollon+

Explore Apollon+’s 515 photos on Flickr!

www.flickr.com

RWN Photo · Apr 16, 2024

BeatX said:
My conclusion: 26Mpix is sweet spot for aps-c size sensors

Seems fairly accurate since a lot of people have come to the realization that 20MP is the sweet spot for M4/3 sensors....

dccberry · Apr 16, 2024

I have both the X-H2 and X-H2s, and the X-H2s is always my go-to camera. I have sold 36" x 36" prints made from the X-H2s that look beautiful, and printed at 16 x 20, I doubt that anyone could tell the difference between the two cameras. The X-H2 requires really careful technique and the best lenses to extract any meaningful gains in quality.

Morris0 · Apr 16, 2024

kdarwin-photo said:
Rather than upscaling 26 MP to 40 MP, I'd have thought it better to go the other way, shrink the 40 MP down to 26 MP.

Or better yet, shrink them both to 12 MP or less based on what will actually get printed or displayed.

Does a 40 MP image or 26 MP image look any different when rendered as 12 MP?

Don't do either. Just crop showing the size you want. Anything else introduces the software algorithm.

Morris

BeatX · Apr 16, 2024

Ap0ll0n said:
To Erik's point, both processed with DXO Pure Raw and zoomed in at 50%. 26MP on the left.

After processing with (latest) Topaz Photo AI:

To me, the 26MP clearly looks over-sharpened and inferior. Topaz helps narrow the gap a bit but personally if I care a lot about IQ, at least when when zoomed in, I would stick with 40MP. To each their own of course and thanks for posting.

The full DXO pics below for reference.

26MP

40MP

Thanks for taking Your time and edit my sample photos.
I think I mostly share Your conclusion.
Still.. it does not change the fact, that in my eyes - those differences are super small, visible on pixel level in 100% magnification

--
My gallery: https://www.flickr.com/photos/maciej_k/

Ap0ll0n · Apr 16, 2024

BeatX said:
Thanks for taking Your time and edit my sample photos.
I think I mostly share Your conclusion.
Still.. it does not change the fact, that in my eyes - those differences are super small, visible on pixel level in 100% magnification

You're welcome. I normally ignore the 50%, or 100%, magnification too but I've come to realize that the top Fuji primes, including your Vitrox one, serve as pretty darn good "zooms" up to about 100% magnification or so, viewed in normal screen sizes and helped by Topaz. Case in point below:

--
Apollon
http://www.flickr.com/photos/apollonas/

Erik Baumgartner · Apr 16, 2024

Here are my comparisons with Lightroom/IXT (No A.I. processing) and my default sharpening settings for both cameras.

40MP (L) vs 26MP (R) at 100%

40MP Lightroom/IXT (L), Capture One (R)

26MP Lightroom/IXT (L), Capture One (R)

A huge difference between 40 and 26 MP here? No, not really, but we're looking at dark/distant detail with these examples which is likely limiting the potential for the finest detail to a significant degree. I suspect most scenes like this will show similar results with both sensors. With closer optimally lit fine detail with optimal exposure/processing I think the difference could be quite a bit more obvious, but probably still not enormous. While 40MP is better and will offer some real advantage in some situations to some photographers, 24/26MP really is quite adequate for APS-C in most situations for most people. I have little doubt that someone who really has their shooting/processing dialed-in with the 24/26MP sensors could probably produce a better result than someone with the 40MP sensor who doesn't. There is so much more to great image than just maximum resolution.

X-Trans 26Mpix vs 40Mpix, differences in resolving fine details

Senior Member

Senior Member

Forum Pro

Forum Pro

Senior Member

Leading Member

Forum Pro

Senior Member

Forum Pro

Senior Member

Forum Pro

Member

Forum Pro

Senior Member

Senior Member

Senior Member

Forum Pro

Senior Member

Senior Member

Forum Pro

Similar threads

Keyboard shortcuts