Fuji X-Trans files and multi-image stitching - Some conclusions after the ACR update


Just as a final piece on the improvements to the Fuji X-Trans sensor with the new version of Adobe Camera Raw, I shot some multi-image files on the X-Pro 1 when I had it, but never felt inclined with the software available to me at the time to stitch them together. Stitching files, because of the warping etc. thats involved to get them to fit together, usually results in a softening of the overall look. Taking into account the problems with Photoshop at the time, the inadaquacies of the Fuji / Silkypix software and the difficulties of getting several RPP 64 files to look the same, I just didn't think it was worth it. However this morning going through the files I decided to see what I could come up with, using the updated ACR and photomerge in Photoshop CS6.



Somewhat to my surprise, they turned out very well indeed. I haven't got any explanation as to why this might be, but they certainly seemed to loose a lot less sharpness than I'm used to. What normally happens is that I usually have to reduce the file size of the stiched file. When these files are put together they are often pretty big - 100MB+ and my usual practice is to take this down to around 60-70MB, which makes them sharpen up nicely. However I felt no need to do this with the Fuji files, which seemed fine at the large size. The full-size high-res file of the top of the page shot of the garden is in fact just over 120MB in size and it looks great. Pretty much the same as what a native one off file from a 42MP sensor would produce. 

As I indicated I can see no reason for this, unless the sensor array is creating files that respond differently to stitching. I have upsized a few files from the Fuji, and while they look OK, they are really no better than anything else I use and certainly inferior to when I do this with my Sigma DP Merrills. So that gives me no particular clue as to why the files stitch together so successfully. I've done five so far and got the same thing with each, so its not just a one-off. 

CONCLUSIONS

So finally, after lots of messing about by Fuji or Adobe or both, I think I can finally come to some conclusion as to what I think about the X-Trans sensor. The simple answer to that is that its very good indeed. But that does need some explanation. When it was first announced, I and I suspect others were expecting something that looked similar to Leica M9 files and those from other cameras that had no AA filter. I was also I think expecting a bit of moire here and there as is usual with these sensors. However the Fuji files don't look like that, though it is possible to get something similar to a Leica M digital sensor file using RPP 64. In the main however they look more like super-clean bayer sensor files. There's no eye-popping colour either, again its a very clean natural look. The really impressive thing though is how far up the ISO scale this continues. This has always been a superb high ISO sensor, even with the software limitations, but now I can see just how good it is. 

As I mentioned in a previous post this opens up the possibility of using higher ISO settings, 320, 400, 640 and 800 on a regular basis and not having the files suffer as a consequence. Yes its very good at ISO 3200 and even ISO 6400, but those I use very rarely, so, if I had a Fuji camera I might be inclined to use higher settings than I usually do. 

Dynamic range is OK, Fuji claim a lot for it but its good not great in terms of the raw files. For me however with ACR and Photoshop its no problem to keep within usable limits. The look of the files is different however. They are not high-saturation and high contrast but something a bit subtler. Now I can get some decent results I'm starting to warm to that look however and there does seem a good depth to the images.

In terms of comparison with what else I use, these are clearly the best files from any of my mirrorless cameras, other than my DP Merrill files at ISO 100 of course. My m4/3 cameras are capable of great results but close inspection does reveal that grainy luminance noise which is always present even with the latest sensors. I also prefer what I'm getting now from these Fuji files to those that I get from my NEX cameras. They just look sharper and less digital. 

When I first got my X-Pro 1 I was fulsome in my praise for the output. Calling it the best IQ I'd seen. Since that time that statement has been hard to justify sometimes, but I always stuck by it in the belief that if a satisfactory mainstream software solution ever appeared then the system could have great potential. And I think we are finally there. From what I've used and seen, there is simply nothing better out there at high ISO's. Also as has been well expressed, the jpgs. are top quality and now we have raw processing that can take that on as well. 

So what of the system? Since I sold my X-Pro 1 I gather the AF has speeded up via firmware, which is encouraging. I have always thought Fuji menus very confusing and that needs to be addressed. In terms of the actual hardware Fuji, who after all made some pretty decent medium format film cameras, seem to be on the right track. I'm not totally convinced by the X-Pro 1 body and the optical viewfinder strikes me as a retro styled gimmick, but one area that they do seem to be getting right is the lenses. The three initial primes, all of which I bought, were very good indeed. The 18-55mm zoom is getting some great reviews and there are more lenses already announced. Unlike Sony, Fuji seem to be trying to get these out pretty quickly which is good to see. 

I wish they would leave the 'Pro' out of their camera names, but thats just me. Its both misleading and pointless. They also seem to be the usual closeted in their boardrooms, paranoid suits that seems to be the norm for camera manufacturers these days. But one thing is for sure. They have some great engineers. The sensor and the lenses testify to that. 

I'm much more inclined to view the system with some optimism after yesterdays software release, since the X-Trans sensor has now become a mainstream product rather than a niche one. Some might have time to fiddle about with different software programmes but many of us don't. I want quick and efficient raw processing and I've now got that. Why this wasn't done months ago is still a mystery and it has, I am convinced, affected how people view Fuji's products and indeed sales of their cameras. Hopefully this can now change and the sensor they have produced will get the plaudits it deserves. 

So overall I believe this is now a system with great potential. Its one with interchangeable lenses which makes it much more useful than a fixed-lens camera, certainly to me, and the lenses are seemingly being made to a high quality. Its a system which sees the virtues of whats gone before and sees no problems with viewfinders, dials, knobs etc. and offers the ability to change things manually. It taps into the retro chic, rangefinderesque, lookaleica aesthetic which is what the market likes currently and the products are reasonably, if not yet attractively priced. 

I thought I had something special when I bought my X-Pro 1 and looked at the files, though I never had the slightest idea that it would take nearly a year for me to see the proof of that on my screen. However, I can see it now and from the generally favourable response to the ACR update, so do lots of other people. Lets just hope that Fuji don't make the same kind of mistake again. If they avoid that then I believe that the X system could have a very bright future.


Join the Soundimageplus Blog Readers Group at Google+

For comment and discussion - join me over at Google+