 |
| This is the full test image as made with the Fujifilm X-T1, 18-55mm lens |
I love the camera, how it feels in my hands, the traditional dial control sets, etc. I love how the camera reproduces colors better than my E-M1 and my Nikon D810. I love the "potential" of the image files it produces. I love the quality of the lenses, even the "kit" lenses. Okay, then what is it don't I love?
I don't love how Adobe Camera Raw and Lightroom's Develop Module demosaic the image files. As I said, that is not Fujifilm's fault. I believe Adobe created those two raw converters for Bayer sensors alone. After all, Bayer sensors were really the only game in town when these two programs were created. When Fujifilm introduced the X-Trans sensor, I'm sure the Adobe engineers started scratching their heads and asking themselves, "what are we going to do with this?"
I only have two basic issues with these raw converters. First, fine green foliage is mushy and lacks detail. Second, fine detail, such as in the image I am using for an illustration here, seems to acquire patterns and a sort of outline that doesn't really exist. As far as urban scapes, portraits and any other subject matter, the files are beautiful.
I will give Adobe credit in that they have improved the raw conversions with each update of ACR and Lightroom, but they are not where I want them to be. I still find these two issues regularly in my landscape and nature images.
There are other raw converters out there that do a better job. From what I hear and read, Iridient may be the best, but it is for Mac only. The majority of the world don't use Apple computers so I am a bit confounded as to why the developer doesn't develop a cross-platform version of that software as Adobe has for Photoshop and Lightroom. I'm hoping one day it is coming. I've tried and kind of liked Photo Ninja, but it has its drawbacks as well. It does a better job of converting the files, especially fine detail and green foliage detail, but it is not perfect. One thing that I don't like is that it doesn't use the Fujifilm lens correction profiles as does ACR and Lightroom to automatically correct distortion, vignetting and chromatic aberrations. With some lenses, the distortion is not simple and very difficult to correct with the tools available. I like the idea that the raw converter developers have solved that complex problem and apply the algorithms automatically.
Now, onto the reason for this post...
As I continue to ponder which of the three camera systems I need to sell (three really is too many), and before I make a final decision, I thought I would give one more of the other raw converters, of which I've read good things, a try to see how it works with the X-Trans sensor. That converter is Phase One's Capture One Pro.
I downloaded a free trial of the software over the weekend and have been playing with it as well as learning it. I would recommend starting out with the free Phase One video tutorials as this software is every bit as full featured, even more so, than many other raw converters. It will take me a few weeks to get comfortable with it and become somewhat proficient.
 |
| This is a crop of the image above converted and edited in Lightroom CC 2015 (click to enlarge) Notice the pebbles. It looks as though they have been painted in. Also look at the patterns in the tree bark and root. |
 |
| This is the same file converted in Capture One Pro. Disregard the color differences as I did not try to match color. (click to enlarge) Notice the pebbles look natural as does the bark and root. |
The downside? The going is slow as I try to learn this software. I suspect, when I do get comfortable with it, that I could edit as fast as I do now in Lightroom. The second downside is that I would have to incorporate many extra steps into my workflow by using a second image editing software instead of just using the industry default, ACR and Lightroom.
A couple of additional thoughts have already come to mind. Converting RAW images with Capture One Pro and then importing them into my Lightroom database, would result in a greater necessity for more file space as 16-bit TIFF files take up a lot more room than RAW files. Second, to cut down on extra time required in Capture One Pro, one could just determine some default settings for the software, convert the files and not do any additional editing, then import the files into Lightroom, using Lightroom's Develop Module to then accomplish all of the necessary editing. That versus spending a lot of time with each image in Capture One Pro editing each image fully. I'm sure I'll have some additional thoughts with time.
I'm always open to beneficial change, so if I find a better way to edit my image files, I'm not opposed to it.
I plan on working with this over the 30-day free trial period, then decide whether or not to keep it and incorporate it into my workflow, let the trial expire and go back to my current workflow or maybe even divest myself of the X-T1 and lenses and just work with two systems—or even just one system.
Currently, I use a full frame system, the Nikon D810 and some great lenses, as well as a M4/3 system, the Olympus E-M1 and some equally great lenses as well as the APS-C format Fujifilm X-T1 and its wonderful lenses. Do I really need three systems? Keep only the full frame and it will do everything at a cost of size and weight? Keep a full frame and a M4/3 for the best of both worlds? Keep only the APS-C format camera which would be a great compromise situation of size, weight and quality? No answers yet, but I'm working toward figuring this out for my future.
Thanks for looking. Enjoy!
Dennis A. Mook
All content on this blog is © 2013-2015 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.
Dennis,
ReplyDeleteDid you use the Peter Bridgwood settings for sharpening in LR6.1.1, for example, Amount to 30, Radius 1, Detail 100 (the deconvolution algorithims kick in at that setting) and Masking adjusted to taste (Alt+slider).
Since setting Detail to 100 in all my .RAF processing I have not seen any 'smearing' in foliage, certainly not any that can't be put down to DoF or wind effects.
Regards,
Hector P
Hector, thank you for your comment.
DeleteYes, I've been using Bridgewood's recommendation. Actually, in my early experimentation with the camera, I independently came to the same conclusion of using the detail slider at 100, but I wasn't aware of the technology behind it until I found Bridgewood's explanation. For my taste, I like to up the Amount slider a little higher for most of my images. If you go back to some of my posts in which I highlighted my tests, you will still see it with the settings as you outlined. In fact, the image above was used with those exact settings.
I agree. I own similar systems, i.e. Olympus OMD 5, Canon 5D MkII, etc. I like the Fuji X system for the same reasons you state, but I believe the time for Adobe to accommodate the xtrans sensor has passed. Because of my investment in the Fuji system, I am inclined to look at an alternative to the Lightroom / ACR software. However, recently I have spent some time with my Canon 5D Mk II and I must say I am more inclined to keep the full frame sensor over the crop sensors. For what I have invested in Fuji and Olympus systems, I could have funded a fine assortment of Canon lenses. As you point out, all software products do a fine job supporting the Bayer sensor.
ReplyDeleteBen, thank you for your comment.
DeleteEverytime I think I'm ready to sell my Nikon D810, I look at those gorgeous files and realize that anything smaller really can't compete, on a dynamic range and fine detail basis. Then I remind myself that I don't need 36mp very often and I'm normally very happy with my E-M1 and Olympus PRO lenses.
Something will have to go before the end of the year....
Your blog post just confirms what I have been seeing when comparing LR to C1P. I agree 100%. For some reason, Adobe seems to add "stuff" to my RAFs. It has taken a little while to adapt to Capture One Pro, but, now I am used to it. I use it exclusively for my Nikon D750 NEFs as well. It does a splendid job with those too.
ReplyDeleteI've done a bunch of side by side comparisons and always choose the Capture One version over LR. LR is quick and easy, but, by no means the best editing software out there, in terms of quality.
Rick, thank you for your comment.
DeleteI didn't even think about trying Capture One Pro with my D810 files. Thanks for raising that possibility as now I will try it and see what potential improvement I may see.
I'm hoping Adobe is not yet finished with their work in bettering X-Trans file and ACR. I have been using Lightroom as my digital database and editing software since it was introduced and I really don't want to "start all over," so to speak, from the beginning again. One gets used to doing things in a certain way and becomes proficient and comfortable. But then, once in a while it may be good to get out of that comfort zone...
Hallo Dennis,
ReplyDeleteyou wrote that the problem of rendering fine detail is not Fuji's fault. This is very polite, but in my opinion, it's mainly Fuji's fault! The basic problem is the X-Trans-Sensor with it's very special layout. This layout violates some fundamental physical laws. Apart from that problem (which is a very serious one) I agree with you and many other users that the Fuji X-T1 is a great cam and does a very good job in many circumstances. But not at all and this unreliability is caused by the X-Trans-Sensor and the raw converters have to work hard to compensate.
I will try to explain the fundamental misconception of the X-Trans-Sensor:
All sensors (Bayer, X-Trans, ...) are analog/digital converters. The (analog) image is sampled point to point and converted into a corresponding sequence of numeric values. This process is subject to the so called Nyquist-Shannon sampling theorem.
This theorem claims:
1) the maximal bandwidth of analog signal (=optical resolution) must be equal or less 1/2 of the sampling rate (= pixel distance)
2) the sampling points should be equidistant.
The X-Trans-Sensor layout violates the Nyquist-Shannon sampling theorem in both terms:
1) the low amount of red and blue pixels may cause under-sampling errors.
2) the distance between the pixels of same color varies heavily in all directions and is different for each type of color.
A violation of the Nyquist-Shannon sampling theorem requirements vitiates the digital sample values in a way that additional information (= aliasing errors) are added to the digital image raw data which were not present in the origin analog image. There is no direct way to detect and remove this false data fraction. For the X-Sensor this means in particular that very fine image details are most critical ... and we could see it in several images.
The trouble is, that the artefacts in Fuji's images are not the result of the raw convert process. The raw data itself enclose the artefacts. The raw process has to filter this artefacts with varying success.
Fuji claims that using the X-Trans-Sensor with its unregular pattern arrangement makes an anti-aliasing filter unnecessary. Ironically this unregular pattern arrangement produces new aliasing errors itself. In Germany we have a phrase for this: "cast out the devil by Beelzebub"
With the X-Trans-Sensor Fuji is caught in their own trap! They had pushed this type of sensor into their cams. A return to the 'normal' Bayer pattern would be a slap in the face for many faithful customers and Fuji is afraid of loss of market share.
We see that a serious company as Adobe didn't invest much manpower to support this X-Trans-Sensor layout. The other raw converter vendors have noticed this gap left by Adobe and tried to jump into with their products to gain market share. The improve in image processing quality varies from vendor to vendor and image content. Reliability looks different!
Last summer I was short before buying a Fuji X-T1 when I heard about the green foliage problem mainly via your blog. So I switched to Olympus M1 and I am very happy with this cam. I own a small Fuji X10. I love this cam too. The advantage is, it has no X-Trans-Sensor. I would never buy it's successor X20 or X30 because of the X-Sensor.
My recommendation is:
stop your investigations with Fuji. Spare this time by using other good cams. The noise of mFT could be handled much easier than the unpredictable artefacts of X-Sensor. And if you need best quality, take your Nikon gear (withBayer sensor).
May be in future Fuji returns (I suppose silent and abashed) to a more physical compliant sensor type, whatever it may be.
Best Regards,
Peter H.
For the record, the Fuji X Trans sensor has the same ratio of red, green and blue pixels for any given megapixel total as a Bayer sensor, ie 1:2:1. Also the X-Trans sensor has all three colours represented in any horizontal or vertical line of six pixels whereas a Bayer sensor has only two out of three colours on each line, which may explain the better colour resolution we get from X-Trans than equivalent sized Bayer sensors.
DeleteWhere the X-Trans CFA may have a problem is that the green pixels are clumped together in groups of four, so it will be interesting to see if the rumoured X Pro 2 has a 24 megapixel APS-C sensor and whether the greater pixel density will mitigate what some users perceive to be a smearing problem with foliage.
Yours in hope,
Hector P
Couldn’t be said any better! I have tried it two times, first with a X-E1, then with an X-Pro1, but always the same, photos look a bit like rendered in 3D software, hence I sold both cameras shortly after buying them (yes, I also tried Iridient Developer with the RAWs, they somehow look good, but still a bit to clean, for what I consider real).
ReplyDeleteI’m keeping my X100 though, this one still has a classic Bayer pattern sensor. I also had an Canon 5D Mark II, love my little Ricoh GR, Olympus Pen E-P2, and E-M10 with good prime lenses.
Hi Dennis
ReplyDeleteI've been using a Fuji X-E1 with Capture One for some time now, and I've found that to get the best results, you need to tweak the default Capture One sharpening settings. I've found that this makes a huge difference. I've posted the settings that I use and the workflow that I use when workflow with lightroom on my blog, which might be of interest to you:
http://blog.thomasfitzgeraldphotography.com/blog/2015/8/my-workflow-and-settings-for-processing-fuji-x-trans-files-in-capture-one
I hope this is of some help
All the best
Thomas
Thomas, good to hear from you. Thank you for taking time to comment. I am an admirer of your work and blog. I have read your blog and have tried your settings with good success. I continued my experiments with C1P and X-Trans files to go a bit beyond the sharpening settings. I have written three additional posts on C1P and X-Trans subsequent to this one.
DeleteAgain, I'm flattered you have read my blog. Thank you again for your comments.
Dennis