Much has been made of the excellent color and files that come out of X-Trans sensors found in cameras such as Fujifilm’s X-T3 and X-H1, and there are plenty of people who rave about the images they make with these cameras. But can you tell the difference between X-trans and a conventional Bayer sensor?
Fujifilm also makes a series of lower-cost cameras that use Bayer sensors like you find in most other DSLRs. These include cameras such as the Fuji X-T100 and the X-A series of cameras. The website DCFever made a blind test of the X-Trans-based X-T20 and the Bayer-based X-T100 to see if users could really tell the difference (It’s probably a good idea to run the “translate” feature in Chrome is you’re not able to read Chinese). The folks over at Fuji Rumors talked about the differences they saw.
The primary difference between the types of sensors is the way the color filter is arrayed: whereas Bayer sensors feature a 2x2 pattern of pixels with larger areas of green, X-trans feature a 6x6 configuration. You can read more about the difference here on Fuji Rumors.
Many will say that this results in improved color science (though you can read why that term is a misnomer here), and others talk about how it can cause issues in certain types of photography, but on the whole, I haven’t seen major concrete differences in my few years of X-Trans shooting. I just know that I love what comes out of my X-T1.
Fuji’s marketing claims this eliminates the need for a low-pass filter to eliminate moiré and provides for more detail, but as this test shows, it’s sometimes hard to tell the difference. I’ve been a Fujifilm user since the X-100T and X-T1, and I wasn’t able to tell the difference easily. I got most of the test images wrong.
Were you able to tell the difference? Take the test and let us know in the comments below.
I picked out the x-t20 in all of them. I tried to pick the best dynamic range and color depth and. Reds and greens didn't appear as "uniform" in the ones I selected.
how so?
why would a 24mp APS-C sensor with an X-Trans color filter not be able to have the same amount of detail retention as a 24mp APS-C sensor with a Bayer color filter, if all other hardware within the image processing chain are identical?
would a monochrome image sensor of APS-C size thus be capable of more or less detail than that of a sensor with Bayer color filter?
I hear people debate the science of this every time I even mention a Fuji in an article, but in real-world usage, I just don't see it - I shoot an X-T1, Nikon D750, Canon 6D, Canon 80D, Panasonic GH3, etc. and given equal sensor sizes, it's about the same. It feels like often people come to these conclusions without ever using the camera.
But have you actually used any camera with an X-Trans sensor for some time?
5:1 for the X=T100.
missed 1. Actually surprised I got that many.
It really comes down to the converter you use, if you're using Lightroom then it's diabolically awful handling X-Trans sensors especially compared to Iridescent and CaptureOne.
From what I've seen X-Trans sensors have low high ISO color noise compared to Bayer sensors, but luma noise is pretty much identical. As someone has shown elsewhere here, the RAW processor can make mush out of the image quality. Probably the best bet is to stick with whatever software Fuji includes for RAW processing, assuming that software doesn't suck.
Unsurprisingly I use a X-Pro 2 and I can see the difference on how Lightroom renders out compared to CaptureOne. So yes, there is a difference in s/w. Here's a 4 year old article:
https://www.fujivsfuji.com/best-xtrans-raw-converter
Newer articles show similar results. There is also a big difference between X-Trans 1 and X-Trans 2. I find the images acceptably sharp and detailed, especially alongside the 5D II/III cameras I use.
There's no native RAW processing software from Fuji? Although for instance Nikon's NX-D is rather clumsy, it will yield the best color fidelity out of .NEF files, inasmuch as the latest versions are warmed-over Silkypix. I'm surprised if Fuji wouldn't include a native RAW processing app for what is a non-standard sensor that could use a manufacturer's custom-tuned RAW processor. Fuji has always gone off the beaten path when it comes to sensor design, with mixed results. Like Nikon, Fuji knows a thing or two about obtaining excellent color accuracy, so I'm surprised if they don't give their users a good RAW processor to yield the best image quality out of their maverick sensor.
You lose all credibility when you claim that software doesn't significantly impact how well a file from an X-Trans sensor is processed. As Daris pointed out, there have been countless articles doing direct comparisons between Adobe's awful handling of the files vs other raw converters. And I know that as someone who doesn't have, has never shot, and likely won't buy for the foreseeable future an X-Trans camera. No excuse on your end for dismissing such an obvious fact.
That doesn't mean that X-Trans is 100% equal to Bayer - but to dismiss it outright is blatantly wrong.
i got 5 out of 6 x-t20. X-Trans for the win
I have to admit that the X-Trans sensor has been around for so long now that I've forgotten why Fuji ever went this route in the first place. I know Fuji has experimented with non-standard sensor designs for many years now, and X-Trans was yet another design concept they went with. In the long run however I feel it has not really served them well, and I suspect there is some "Japanese pride" involved in Fuji's continual use to the technology. I respect Fuji for thinking out of the box in sensor design, but I think they have yet to really "hit a home run" with their obtuse designs. I would like to see Fuji evolve out of X-Trans into perhaps something that "brings home the bacon" and justifies their non-standard approach to sensor design. At the same time, I would like to see them also embrace Bayer sensors in their more sophisticated model designs.
"I feel it has not really served them well"
Really, How so?
From my experience and the interwebs in general, I hear and see nothing but praise for the X-trans sensor in the X line of their cameras. From their X100F, to X-T1,2,3 Fuji has killing it lately. I left my Nikon system for them. They Have the best out of camera JPEGs of any camera body in its class. Do you work for Sony?
You realize you sound like a Sony fanboy, right?
I'm a Fuji shooter, so not sure what you think i sound like.
BTW: good luck with Sony updating anything in a timely fashion. Sony is all bout putting out new hardware, they could care less about updating existing hardware.
Fuji for the win.
I don't shoot Sony, although I did once make the mistake of buying an RX100 III. Fuji does indeed make a better product than Sony in my eyes, but remember that it's Sony making Fuji's sensors. ;-)
To my own surprise I indeed picked them all right...
I have never used a Fuji camera, but I guessed that the X-trans would have slightly more saturated colours and that let me score 5/6. So a very slight difference is visible in a side by side comparison but I would not say that either is superior to the other. They are just a tiny bit different, and he difference is so little as to be irrelevant in real world photography.
This is disturbing. I picked the X-T100 with Bayer array filter as my favourite for 6 of the 6 images. A couple, side by side, were close but when you look at the images full size, the majority of the X-T100 images are crisper and more detailed than the X-T20 images and also have better colour. The X-Trans images just look a bit dull and detail muted by comparison.
Why disturbing? Well, I recently bought my first Fuji camera, an X-E4, based on marketing selling the benefits of X-Trans and the enthusiasm from the Fuji users. But after 4 months of using the camera and looking at the results, I've been questioning the marketing and user enthusiasm. It's all a bit of a have because what I'm seeing is that X-Trans gives worse results