XT2 Zoomed in Image Quality

[Foto Jay]

Hi! I have owned an Xt2 for a little over a year coming from a full frame Canon system. I have been super happy except one thing. Maybe I am going nuts but when I zoom in on an image it falls apart quality wise. Especially wider shots. I always view 1 to 1 in lightroom on a 27" I Mac and sometimes I zoom in a tad more to fix blemishes etc. The edges (not the edges of the picture, but the edges of the people or objects in them) just are more jagged, blotchy, very pixelated etc... More than I remember on my 5DIII. Is that just part of the APSC game? Once again, I am happy user but this has been bothering me. It would be kind of someone to post a picture or two for me to compare would be greatly appreciated. Of course the issue arrises more with the 18-55 kit lens then my others, but the issue is there with all my lens's, so I can rule out optics problems. Thanks in advance for any insights. I would like to put this to bed if I can. Like I said, I am super happy, but this has been on my mind the whole time from the very first shoot I did with it.

PS.... For what it is worth, I have tried all the raw converters to rule out lightroom messing with me I do use photoshop to sharpen, as I do notice lightroom sharpening does amplify the problem slightly.

 

[Bobo Hodls]

I know what you mean. I'll develop at 100% magnification, but feel the best size for viewing 'large' is a step down from 100% (be it in LR, or on Flickr).

I also find that compared to most all Bayer sensors when I sharpen in LR masking needs to be drastically strong. Note when holding Ctrl/Alt down while pushing the masking slider what the white portion of the mask will be sharpened.

 

[JonPB]

...jagged, blotchy, very pixelated etc...

This is not at all how I'd describe my experience of viewing Fuji raw files at 1:1 on a Retina display. If anything, fine details are softer and blander than what I've seen from non-XTrans cameras, at least until sharpening is applied.

I'd be curious to see a screenshot (default on my Mac is CMD-Shift-4 then drag over the area you want to capture) of what you're seeing.

Here's from a test shot from an 18-55, processed at default settings in Capture One 9 and viewed at 100%. Download into Preview and use CMD-0 (zero) to zoom to 100% and see what I'm seeing, which I'd describe as slightly harsh but not jagged, blotchy or pixelated.

 

[Foto Jay]

Thanks JonPB. I have taken a quick screen shot of what I see on the 18-55 at 100 percent via cmd 0 as you suggested. Granted it was a 1200 iso shot if I am not mistaken but I figured image quality should be a tad better than this. Any thoughts?

 

[ediblestarfish]

In my experience with FF the larger sensor provides both less noise and greater potential resolving power, as well as increased dynamic range given similar vintage over a variety of lenses and situations.

I used to shoot m4/3 and took a leap to FF. Later settled down on APS-C, and yes, the potential image quality in those three aspects does change noticeably with the sensor.

It may or may not make much difference depending on how you process your files, and the final output--small screen, or giant print.

 

[Jerry-astro]

Thanks JonPB. I have taken a quick screen shot of what I see on the 18-55 at 100 percent via cmd 0 as you suggested. Granted it was a 1200 iso shot if I am not mistaken but I figured image quality should be a tad better than this. Any thoughts?

Oh, where to start...

For starters, a screen shot is a terrible way to compare or assess image quality. Do your best shot at post processing (or work from a JPG... your choice) and then post a 100% crop for people to assess if you want to compare at that level. There are so many variables here in terms of your initial settings, how you processed the image, sharpening, etc. etc. that without some insight into how you got there, it's hard to really give you any meaningful guidance.

I can say that the 18-55 has been my "go to" lens for much of my photography and it has excelled in terms of sharpness, contrast, color, and build quality both on my original X-T1 and the X-T2 that replaced it. If what you're showing here is truly representative of the best you can get with proper processing, then I'd strongly suggest that you look into a repair or replacement. But before doing that, please make sure you're not doing something that is impairing the quality of your output. If you're not familiar with using post processing tools (such as Lightroom or the many other alternatives out there), then perhaps try shooting in JPG with appropriate settings to get a decent finished result and post it here at full resolution for folks to look at. Then, perhaps it might be possible to help you get better results.

--
Jerry-Astro
Fujifilm X Forum Co-Mod

 

[aarongarybrooks]

Sadly, this seems to be part of the Fuji X experience (which is otherwise great as you say). It's been there on Xtrans II and Xtrans III sensors and indeed worse with the 18-55 than other lenses but present across the board. Considerably worse with Lightroom but present across all processors I've tried.

It stopped me buying the X-T2 and I'm tempted to go Nikon FF and use Fuji as backup, but I'm trying to hold off on until the X-T3, Nikon mirrorless and D750 replacement are here.

It's a shame because using Fuji X cameras is a pure joy but I get the same nagging feeling when processing images.

I haven't tried the X-A3 to see what Fuji X with a Bayer sensor looks like but, despite my curiosity, it would solve nothing for me unless Bayer options were there for flagship models.

This doesn't seems like it will happen anytime soon.

I haven't ruled out staying with Fuji yet as I'm constantly travelling and lites to how much gear I want to carry - one of the many Fuji X benefits. However, I'm not going to invest further until something changes or I've made the decision to stick or twist.

 

[Foto Jay]

Thanks Jerry-astro. I appreciate the input. I definitely understand that many variables are in play and I am very familiar with post processing. I really don't think it is the camera or the lens. If anything it is me. With that said, I see the same problem (not as bad) with my 50-140 also. and I have also processed thousands upon thousands of images (Canon based though) throughout the years without issue. I am a Lightroom user along with Photo raw user. I have tried sharpening in Photoshop and bring back in as tiff etc...I know all the tricks I have tried irrident etc...Unless I am missing one or two of course. This was an early example and more or less a little more severe of a problem for a posting reference. I have no clue where my thought is going, but the bottom line is that when you zoom in, I see quite a lot of pixelation (lack of a better term)Is that normal? Is there a base standard for viewing Fuji images compared to what I am used to? I always thought 1 to 1 was a good representation of the final image. Should I post a different image straight out of camera that represents what I see? I always shoot Raw + Jpg. so finding one should be easy

Thanks,

Jay

 

[Foto Jay]

@Yetanotherusername. That is not good news. I love using my Fuji. I feel guilty that I am pixel peeping of course here but I was one of those things that keep popping up when I process my images. I notice it more in wider shots vs closer shots. I also notice it more with higher ISO images. I just feel I should be getting better out of the XT2. I see fantastic examples posted all over the forums. I just need to know if this is normal, and I guess you were agreeing that this is normal behavior of the files. Bummer.

 

[JonPB]

Hm. Good news/bad news: while I think I see the problems you're talking about, it doesn't look like there's anything wrong with that crop that can be blamed on malfunctioning hardware/software. Nor is it an APS-C issue. Rather ... welcome to X-Trans!

The overall gritty overlay of noise is (probably, I think) due to Fuji's heavy-handed chroma noise reduction. This nearly eliminates nasty color noise. But it is baked into the raw file. Which means that raw processors can't ignore it as color noise when finding edges to sharpen.

The pixelation around the candles and hard reflections is due to over-sharpening, while other details are softer than you might expect due to under-sharpening. The problem is that processing X-Trans images the same as Bayer images won't yield the same result. Personally, I've become contented with the relative lack of fine detail -- after shooting a monochrome camera, nothing else ever seems good at the pixel level anyhow -- because it isn't a problem for my final output (and there is so much else that is good with this system). Others have dedicated a lot of effort to squeezing the most pixel-level detail out of X-Trans files, so you might be well served by looking for those discussions.

Hope this helps, even if it is unlikely to be the answer you wanted.

 

[Foto Jay]

@ediblestarfish. 90% doesn't matter what format I shoot for my work. Like you said, most images are smallish and not many people will notice. I do shoot higher ISO's and wider these day. I was hoping to find a trick or two I missed with the transition to Fuji. The Fuji is sharper than my Canon was Even after micro adjusting all the lens's. so thats a benefit at least.

 

[Foto Jay]

JonPB. That was not the answer I was looking for. I had a strong suspicion that was the case, but wanted to be sure it was not me. This was a lightroom sharpened image (before I knew how to properly sharpen the Xtrans in Photoshop) and bring back into Lightroom and then sharpen the tiff. I appreciate the detailed response of the baked in data. Never new that about the Xtrans. All I know is that they are a pain to sharpen without the worms!

 

[JonPB]

If it helps, I've found that X-Trans truly does a good job at avoiding moire. Going back to my Bayer files, I'm surprised at how much I downplayed the impact of moire, even when it was visible in mere thumbnails. X-Trans is a win some, lose some proposition -- but I feel like I win more than I lose.

 

[Father Bouvier]

I'm inclined to agree that it's a combination of a strong noise reduction (false color suppression) and a strong sharpening. And the softer the lens the more sharpening required. But I'm not sure if that is baked-in in the raw file, it's more likely an unavoidable part of demosaicing. If that were already in the raw file then Bill Claff would have noticed that, and his measurements look normal.

 

[JonPB]

I'm not sure if that is baked-in in the raw file, it's more likely an unavoidable part of demosaicing. If that were already in the raw file then Bill Claff would have noticed that, and his measurements look normal.

You're entirely right. I had never seen reasonable levels of chroma noise in a high ISO X-Trans image, even with noise reduction nominally zeroed. I just opened up an ISO 12,800 shot in Rawtherapee, which is the most "raw" processor I have installed, and there it was. Lots of color noise -- just not shown in Capture One, which I guess I rely on too much for this sort of thing.

Thanks for helping me advance my understanding! Now I have to chew on what that means...

 

[rwbaron]

The only thing I can think of (making the assumption there's nothing wrong with your equipment or processing) is you're seeing the difference between FF and APS C.

I've shot Canon digital since 2004 and always APS C starting with the 10D, then XTi, 40D, 50D and finally the 7D. I'm also a LR user and converted to LR from Canon's Digital Photo Professional a few years prior to experimenting with Fuji. My interest in Fuji was spurred by forum mod Jerry Astro who also was a Canon shooter and DPR Canon forum contributor. Jerry hosted a thread on his experience with a rented X-T1 and 18-55 and after reading and participating my interest was peaked. At my request Jerry was kind enough to take some comparison photos with his 7D and the X-T1 and transferred the raw files to me to play with.

Frankly, playing with the raw files in LR that Jerry provided sold me on the X-T1. The images were markedly better than the Canon 7D in sharpness, color and especially in shadow noise and that has been my experience with my own equipment. After further research I bought an X-T1, 18-55 and 55-200 to try with the small possibility I may end up returning the equipment. I spent a week shooting the X-T1 along side my 7D with comparable lenses on a tripod and scrutinizing the images at 100 & 200% in LR and I never found that I preferred the Canon raw files in any way shape or form. All my Canon gear has sat idle for the past 3 years because I prefer the look and quality of the Fuji raw files hands down and I'm a habitual pixel peeper. I just spent a weekend organizing and cleaning all my Canon gear as well as researching the value and I'm preparing all for sale.

Again, my only thought is you're seeing a FF vs APS C difference.

Bob

--
http://www.pbase.com/rwbaron

 

[rwbaron]

Sadly, this seems to be part of the Fuji X experience (which is otherwise great as you say). It's been there on Xtrans II and Xtrans III sensors and indeed worse with the 18-55 than other lenses but present across the board. Considerably worse with Lightroom but present across all processors I've tried.

It stopped me buying the X-T2 and I'm tempted to go Nikon FF and use Fuji as backup, but I'm trying to hold off on until the X-T3, Nikon mirrorless and D750 replacement are here.

It's a shame because using Fuji X cameras is a pure joy but I get the same nagging feeling when processing images.

I haven't tried the X-A3 to see what Fuji X with a Bayer sensor looks like but, despite my curiosity, it would solve nothing for me unless Bayer options were there for flagship models.

This doesn't seems like it will happen anytime soon.

I haven't ruled out staying with Fuji yet as I'm constantly travelling and lites to how much gear I want to carry - one of the many Fuji X benefits. However, I'm not going to invest further until something changes or I've made the decision to stick or twist.

Sorry and I don't mean any disrespect but from my experience what you posted above is pure rubbish. I'm a former Canon APS C shooter logging tens of thousands of Canon raw frames on a Canon 7D and work with those Canon raws and my X-T1 and X-T2 raws interchangeably in LR and there simply isn't a problem with X-Trans and LR.

The list of notable professional photographers moving to Fuji X-Trans is growing all the time. These people are shooting wedding, portrait, fashion, landscape, street, whatever and coming from Canon and Nikon FF. Many if not most are using LR to process their X-Trans raw files and I doubt they'd be doing so if they saw any inherent problem with X-Trans as there livelihood depends on their output.

I really wish people would give this a rest.

 

[Foto Jay]

This is evidently a hotter topic then I anticipated or I even realized. I know the raw converter one is though! Thats for another day. I just needed to confirm that what I am seeing is there and is apparent in Fuji XT2 files. I also wanted to know that is normal behavior or some technique I am missing etc...When push comes to shove, it is a small price to pay for convenience I assume. Just shooting full frame for years and going to the Fuji files was a different experience. Some look great and sometimes better than what I was used to and some didn't live up to my expectations. I guess its give or take at this point. I can always go to Sony and have it all I guess. Kidding of course! I do indeed love shooting the XT-2 and will look forward to what 2018 brings to the table. In the meantime, I will be shooting an Image for a billboard this week. No joke I will see how that goes with Fuji. No pressure of course.

 

[aarongarybrooks]

Good for you. I'm not the first to document this "pure rubbish". You're not the first to get overemotional about the topic.

Hope you continue to enjoy your setup.

 

[nikkor35f2]

fuji excels in near field selected focus w a good distant blur. but if subjects are flat(as in too close to a wall etc) at infinity, the results has always been more disappointing. doesnt matter which lens you are taking about.

i believe its sensor issue. if you shoot 18-55 w a subject at 2-3 meters away the subj will be amongst the best in any system. but is the subj is 20m away against a wall...everything just becomes plain flat. and nothing is truly sharp nor blur...same w 18-135