Friday, November 10, 2017

I've Been Experimenting With Iridient X-Transformer And My Fujifilm X-T2 RAF Files

This is the full image used for one of my Iridient X-Transformer tests.  Below are 100% comparison images
showing editing only in Lightroom and using X-Transformer to convert the X-Trans RAW file.
X-T2, 16-55mm f/2.8 lens @ 55mm; 0.9 sec. @ f/11; ISO 200; tripod mounted
In their latest release of Lightroom Classic CC, I was hoping Adobe would have put a bit more work into Lightroom's demosaicing algorithms when it comes to converting Fujifilm X-Trans RAW files.  Unfortunately, I don't see any difference in image quality (a very subjective assessment, of course) from the last version.  Going from Process Version 3 (2012) to Version 4 (2017) made no difference that I can detect.  

Generally, for the majority of my images, Lightroom does a "good enough" job of converting Fujifilm RAW files.  However, there are some images that: 

a) either have some false content artifacts that I have detected and those bother me as the LR conversion algorithms shouldn't be allowing those, 

b) the green foliage, although looking pretty detailed, is not quite what it should be, or 

c) I have an image which, because of its nature, I want to extract maximum detail and sharpness.  

The instances in which I want something better than Lightroom's RAW converter, I think, are "occasional."  As a result, I thought I would try experimenting with Iridient's X-Transformer Beta 4 RAW conversion program—again.  The first time I experimented with it, I didn't see enough of a difference to make it worth my while to incorporate X-Transformer into my workflow.  But in this new Beta 4 version and in short, so far I like what I see.  

The obvious questions that come to mind next are "How much detail and sharpness does one really need versus how much more can be extracted with a lot of extra work?"  At normal viewing distances do I really need to extract maximum detail from any image or am I chasing a rainbow and only looking to satisfy my own ego and trying to justify my purchasing decisions (Fujifilm gear)?  Do I need extreme detail in all my images?  Will anyone else except another "pixel peeper" even notice extreme detail or do almost all who will look at our images view the image as a whole and enjoy what they see?  Maybe its only us photographers who are obsessed with extreme detail?  Do you think?

I think an appropriate analogy would be spending $4000 for the absolute best Zeiss lens just to make images with ultimate sharpness versus buying a $1500 Canon or Nikon lens that is only a tad less sharp.  Do you really need to achieve ultimate sharpness in your images?  Does it matter?

In the last couple of releases, I have been generally satisfied with Lightroom's treatment of Fujifilm RAW files, in most cases. Lightroom does a pretty good job converting Fujifilm X-Trans RAW files (I specify X-Trans as Fujifilm has and does make cameras with Bayer type sensors).  They are not the best conversions by any measure, but okay.  I apply that assessment to most of my photography, i.e., my keeper images, routine photography, record shots or family events or the like.  But when I started my experimentation with X-Transformer last week, this latest Beta version of the program took the files I tried to another level when it comes to detail and sharpness.  Either I'm seeing more resolution or I'm seeing more sharpness.  I'm thinking a bit of both.  But I'm getting ahead of myself.

I had experimented with X-Transformer sporadically since Beta 1 came out.  At that time, I wasn't thrilled that the differences I saw were worth the extra steps necessary to incorporate X-Transformer into my workflow.  I also took a swat at Beta 2 and my opinion hadn't changed.

Recently, however, and as a follower of Thomas Fitzgerald's blog, which can be found here, he has continued his extensive work with X-Transformer and Fujifilm RAW files.  He has now published an e-book highlighting his work.  I bought Mr. Fitzgerald's e-book containing his explanations and his recommendations on how to best use X-Transformer. Beta 4 appears to be significantly better than its predecessors.  I can recommend his book if you, too, are thinking about utilizing X-Transformer to convert your Fujifilm RAW files. The book provides a very good understanding on what X-Transformer does, how to set it up, processing individual images as well as batch processing and he has developed some sample recipes he recommends depending upon certain factors.

I did some extensive testing on a number of different types of files, with fine detail, with and without green foliage, with and without lots of color and came away thinking I will use X-Transformer in the future for what I believe are the images that I deem worthy of the extra time and effort.  Again, I don't think I need extreme detail and sharpness in the majority of my image files, only select ones for specific reasons.  One of those images I worked with is the one I posted here.

For my testing, what I did was take a Fujifilm RAW file, edit it and sharpen it the best I could using Lightroom.  Then, I would take the original RAW and convert it using each of Fitzgerald's four different recipes he has created, edit each of those files the best I could, then directly compare the five resulting images.  I was delighted with the extra detail X-Transformer drew out of my files.

The images below are direct comparisons of the same file as shown above.  The left side of the below images show the X-Transformer conversions and the right side show conversions and editing only in Lightroom.  I don't know how well you can see the differences in this rather small (1600 pixel), low resolutions comparisons, but at 100% on my 27" monitor, the differences are very clear.  I could easily see detail in the logs, leaves, plants and roofing tiles in the first image and the fencing, rocks, leaves, tree bark and vines in the other two images.  I did not apply any export sharpening to any of these images.

I'm not sure how much you of the enhanced detail in the image on the left you can see, but it is clearly visible in the full
size image on my monitor.  The small green plant just behind and to the left of the rock in the foreground shows detail
in its leaves in the X-Transformer converter image but none in the LR image.  Same with the roof tiles. The logs in the 
cabin show significantly more detail.  However, if you are not careful, the X-Transformer conversions can appear oversharpened.  These, I think are right on the verge. (click to enlarge)
The X-Transformer image is on the left.  There is clearly more definition in the individual leaves, tree bark, etc. 
(click to enlarge)
 I see much more detail in the wood, the leaves, vines, rock, etc. in the Iridient conversion on the left. (click to enlarge)
My initial impression was that when directly compared, the Lightroom images appeared as though they never had been sharpened or not sharpened enough, when in fact they had been sharpened significantly!  They looked great until you compared them to the same image processed in X-Transformer.  I guess it is all relative!  I couldn't match the Iridient conversion for detail and sharpness no matter what I did in Lightroom.

So this brings us back to my initial questions and one additional one.  Do you really need this amount of extreme detail and sharpening, for all practical purposes, or "wanting" this level of detail and sharpening more of an OCD (Obsessive/Compulsive Disorder) or egotistical thing?  Are other aspects of your image more important than detail?  Finally, does extracting maximum detail make your images better?  That is the bottom line.  You have to decide for yourself.

As for me, I have not yet finished my experimentation.  I see a couple of things that I need to do a little more work on so I can decide how I want to set up the conversion menu in X-Transformer.  Also, I find the X-Transformer files tend to be a little more saturated and a little more contrasty than the LR file.  Also, as I mentioned in the caption above, it is easy for the files to take on a look of being over-sharpened, maybe even a tad crispy.  In my mind, I think that is due to the fact that the extreme detail now visible makes the file look different than what I am used to.  Also, this program is still in Beta.  I want to see what the final version does before I buy it.

If I do eventually buy the program, I think I will only use Iridient X-Transformer on certain images as I stated above.  I don't think I need the extra detail and sharpening on the vast majority of my images.  As I said, only we pixel peepers are going to notice anyway.

I hope you find this valuable.  I suggest you conduct experiments with your own images as the types of photography you practice may be much different than mine.  You can download a free trail copy of X-Transformer for Windows of Mac here.

Thanks for looking. Enjoy! 

Dennis A. Mook 

All content on this blog is © 2013-2017 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.

7 comments:

  1. Dennis,
    thank you yet again for your research and results! I am using the final (non-beta) version of X-Transformer for my "portfolio" images, and very pleased. Like you, I have both Fuji and Olympus, so don't want to have 2 "Lightroom-type" programs and this allows me to retain LR as a basic workflow/DAM. I had noted in the recent onlinephotographer post discussion that it's hard to understand how a one-man shop, like Iridient, can deal properly with X-trans but huge Adobe still cannot after many years. Ah well, .....
    Rick

    ReplyDelete
    Replies
    1. Thanks, Rick. I looked at this post on my current version iPad and I’m sorry to say, the differences between the two comparison images isn’t as apparent as on a large monitor. If anyone wants to really see the differences that I saw, I suggest downloading the free trial of the software and converting a few Fujifilm RAW images with it. The differences are quite apparent.

      Delete
    2. Dennis,
      I've been infatuated with the Fujis since they came out for their old-school direct access knobs and great build quality. Having rented an X-T1 some years and being put off by the 16 MP RAW files, i put tabled the consideration. But with all the on-going love buzz for the Fujis, and how they supposedly have addressed much (but not all of those concerns) I rented an xt-100F just to try out the newer sensor. Well, just like you i saw those issues in the fine details and especially the green colored areas of vegetation. But here's a twist, to my eye, its challenge seems to be with the edges and lines of naturally curvy organic shaped objects whereas hard and straight edged objects seem to have fewer of the false lines and artifacts. I'm glad you shared your findings. And given how immediate they seemed to me, i'm a bit mystified why more don't report it to be of more concern.

      What I haven't done is to see how much of what we see on the computer screen translates to the printed iteration of the image. I'd be very interested in hearing your and other readers' impressions of how well these raw files (from Lightroom) print...do the artifacts convey or are they somehow processed out as it is converted to a printed image rather than digitally viewed?

      Craig

      Delete
    3. Dennis- this post, seen on my laptop, was perhaps the first time I actually saw a difference in side-by-side images, the difference was marked. Most of the time (perhaps because I am on a smaller device), there is no discernible difference to me IMHO :)
      Rick

      Delete
    4. Craig, if you have the need to print a couple of files which show an anomalies on your computer screen, I would appreciate you sharing your findings as to whether or not you can see them in the enlargement themselves.

      Delete
    5. Rick, your comment brings me back to one of my original points. How do most people view images on the web? Is it on a large monitor or possibly a mobile device of some sort. If one cannot discern minute differences in detail or sharpness when looking at your work, then our quest for extreme detail and sharpness is a fool's errand if not only for our own satisfaction. Again, does extracting maximum quality make our images aesthetically better?

      Delete
  2. Thanks for your great work!
    It seems to me that the M43 with a little bit less pixels is giving the same details as the Fuji APS-C sensor?

    ReplyDelete