Ever since Fuji released its first X-Trans camera, and Adobe added support, many of us who have shot Fuji over the years have been unhappy with how Lightroom handles Fuji files. There are issues with the way it handles fine details, certain repeating textures and so called “worm artifacts”. People have been hoping that Adobe would eventually fix the problem, and turned to other solutions, such as Iridient X-Transformer. Others have switched away from Lightroom together to something like Capture One. Today, Adobe has released a new version of Lightroom which finally addresses the X-Trans issue. Well, sort of, as it’s probably not the solution that many were expecting.

If you were hoping that there was a new process version which would automatically improve the rendering of all of your X-Trans files, then I’m afraid that you’re out of luck. Instead, Adobe has a new feature called “Enhance Details” which uses machine learning and AI to improve the quality of demosaicing in Lightroom. It creates a new linearised DNG file, which still has the edibility of a RAW file. If you are familiar with X-Transformer, it works in a similar way, in that it creates a DNG with improved demosaicing.

It should be noted that this isn’t designed specifically for X-Trans and is compatible with all raw files. Adobe claims that this process can extract an additional 30% of detail from a RAW file. When used with an X-Trans file, it creates an artifact free image, that doesn’t have any of the issues that Lightroom normally has. I’ve had access to an early version and based on my initial testing, it’s very effective. It produces extremely clean and detailed images from your X-Trans files, and is probably of a higher quality than X-Transformer in certain circumstances. On non Fuji files, there is also clearly a visible difference when examining details - at least in the files I've tried.

There is a caveat though. It’s very processor intensive, and therefore slow. It requires a relatively modern GPU, and it can take some time to process. Adobe doesn’t think that this is something that you would use to batch process images (although you can), but should be used on one image at a time. On my old MacBook Pro it takes around 1 to 2 minutes on a Fuji file. Compare that to X-Transformer, which takes around 15-20 seconds. It should be noted though, that this is doing far more than X-Transformer, and in my testing, for some images there is clearly a difference. I should point out though that my MacBook pro is quite old, and on newer systems with a better GPU this should be much faster. On my 2012 Mac Pro it doesn’t work at all as my GPU is too old.

Now, I know what people are going to say. Why did they go to such a roundabout way of addressing X-Trans files, and not just fix the original demosaicing? I did ask, and based on what I was told this is my understanding of it:

As I've already said, "Enhance Details" was not designed specifically for X-Trans files, and is doing some advanced computational imaging to improve the quality of all RAW files. Lightroom and camera RAW was designed to maintain backward compatibility and non destructive editing, all the way back to the original version and be fully non destructive. That means, if you edit a RAW file from 12 years ago in the latest version of Lightroom today and update to the latest process version, the look and feel of the image with be very close, and you’d have access to all of the latest tools.

Because "fixing" the X-Trans issues would have required significantly changing the most vital part of the rendering process - the demosiacing - It would have required significant engineering effort to change it for X-Trans rendering and not break the backwards compatibility, and may not have been possible without affecting the way the system works.

I guess that other applications can approach this differently because they either don't have this legacy or are willing to break backwards compatibility. I'm sure some will argue with this explaination but it makes sense to me.

Adobe believes that many X-Trans shooters are happy with the way Lightroom handles most RAW files (which I can sort of confirm given the backlash I got when I first tried to help people with the problem), and it’s only a small number of users and a small number of images that have signifiant issues. It wouldn’t be really justifiable reengineering the entire system just for a small number of users.

It’s also worth remembering that Fuji has a relatively small market share, and within that, a large percentage do just use Lightroom as it is and are happy with it. Ever since I started covering this issue, I’ve gotten swathes of hate mail over the years for even suggesting that anything was wrong, so there is a large percentage of users, for whom the exiting conversion is either good enough, or they just don’t care.

My point was always this: it’s not that it wasn’t good enough, it was that it was wrong. The existing conversions produced false details that are just wrong. I can give you any number of images that demonstrate this. You may not notice it in a large image, and it may be good enough, but it’s still wrong. Enhance details fixes this.

I haven’t done an extensive set of comparisons yet, but based on what I’m seeing, it does appear to produce results that are in many cases superior to most other converters. In my experience, when it's this close, people will see what they want to see, so it probably won't change any minds which are already made up about this. Having said that, at least based on the tests I've done, it appears to produce results that are sharper and cleaner than Iridient X-Transformer, especially when it comes to colour details. Edges are cleaner too, and Iridient tends to produce images with some colour effects around edges, but again, I'm really nitpicking.

At least there is a solution now that doesn’t require the purchase of additional software. I’m guessing this may impact on sales of X-Transformer, but given how much faster X-Transformer is, maybe it will still have a place, especially for those of us with older computers.

So what’s it like in use?

It’s pretty straight forward to use. You simply right click on an image, and choose “Enhance Details”. This brings up a dialog giving you a preview, and some information, and then you click ok. Thats all you have to do. You can batch process images with it too, but beware, it can take a lot of time. If anyone has a high end GPU sock as an iMac pro, or something equivalent on the PC side, please leave a comment with your processing times. I’m curious to know.

In terms of quality I’ve done a few comparisons for you to see. I’ve chosen these images as they demonstrate the issues, not because they’re good pictures. The first example is a crappy picture of a wall. I took this specifically, because Lightroom always had an issue rendering this kind of texture when shot on X-Trans. Here’s a 1:1 crop of the original Lightroom version:

Here’s the “Enhanced Details” version. As you can see, it’s clearly much better. It’s much more natural looking and it doesn’t have any of the weird smearing.

For comparison, here’s the X-Transformer version:

Here are a few more comparisons.

Conclusion

In my opinion this does a good job at rendering X-Trans files. I wish you didn’t have to do this extra step, and I wish it wasn’t so slow, but it’s a good solution. Quality wise, I think it is up there with the best. On some images, the differences are so negligible, I think people will see what they want to see. On others it is clearly better. Again, it’s worth repeating that this isn’t specifically designed for X-Trans files. I tested it on some old images from my Nikon D700 and the results were subtle but clearly visible to me. Fine detail was better - especially when there was moire - as this was almost completely emiminated. Round details that have aliasing appear more defined. Some would see this as very subtle, but I spend a lot of time analysing images and I can see the differences. I haven't done enough testing with non Fuji files yet thoghh, and I’m doing some more testing on other images from other cameras too, and I’ll have a follow up article discussing that.

I’m guessing that there will be something of a backlash against this. Bashing Adobe is quite popular, and this may be seen a s a “cop out” by some in the Fuji Community. I believe that Adobe is genuine in their explanation and this is a way to resolve the issue while also providing additional functionality. Is it an ideal solution? No. Of course not. But it is something and it does work. I’m sure with time, this will improve and it will get faster. In the mean time, if this doesn’t work for you there are alternatives out there now such as X-Transformer or Capture One.

June 2020 - Follow Up

In the year plus since I wrote this post, there has been a lot of commentary on this, and I’ve done a lot more analysis myself. The two biggest issues for people are that it takes up a lot of space, and it still isn’t as good as X-Transformer. The first is certainly the case. Since I wrote this a new version of X-Transformer has been released that includes new compression methods that allow for much smaller DNG files. The second is a little more subjective, but having had a lot more time to analyse this now, I think that in most scenarios, the difference is subtle, and personally I prefer the “Enhance” version.. Since this came out I almost never use x-transformer any more.

If you’re still struggling to find the best options for processing Capture One files, then I have posted another article about this, which covers what I think are the three best options and why. Capture One still offers the best quality, followed by Enhanced Details. You can find the full post here.

Help Support the Blog

This blog is pretty much my day job now and I work to bring you my own take on photography, both tutorials and tips, as well as inspiration from my own art. I support this work, and my YouTube channel entirely either via sales of my eBooks and Lightroom presets, or the kind support of my readers. (I also have some affiliate ads). Running all this isn’t cheap, and so If you like what you see and you want to help keep this all going, there are a number of ways you can do so:

If you’re interested in buying Capture One but haven’t yet bought your own copy yet, then by going through my affiliate link I’ll get a small commission!
You can buy any of my Capture One Styles through my online store!
If you’re buying anything on Amazon, you can help by using my affiliate link: Amazon US or Amazon UK.
You can support me on Patreon. If you like what I do here and find the information useful, then you can help by subscribing to me on Patreon. As well as helping keep this blog going with even more useful news, tips, tutorials and more, members also get special Patreon only feed, which gives you behind the scenes content, as well as monthly downloads (for the top tier), for only €2 or €5 a month Stop by and check it out.
Buy me a coffee / Tip Jar. If you want to support the blog but don't want to sign up for Patreon, you can buy me a coffee via PayPal with a one off donation to my PayPal tip jar.