Home · Register · Join Upload & Sell

Moderated by: Fred Miranda

New fredmiranda.com Mobile Site

New Feature: SMS Notification alert

New Feature: Buy & Sell Watchlist

FM Forums | Leica & Alternative Gear | Join Upload & Sell

1 2 3

5 end

Archive 2014 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

Tariq Gibran
Offline
• • • • • •
Upload & Sell: Off

p.4 #1 · p.4 #1 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

Dave McGaughey wrote:
Thanks Sal, this is interesting. I think the crux of the X-Trans "look" relative to Bayer is that when you apply a lot of USM sharpening to a Bayer image, you start getting "dot noise" or "texture." This gives the impression of extra detail. X-Trans reacts much differently - they start smooth and stay smooth even with moderately aggressive sharpening. From a bit of tinkering I've found that the oddly smooth ACR X-Trans images look a little more Bayer-like when you add noise to the image.

I think it's down to just a bit more real resolution in the bayer scheme versus the X-Trans scheme (this fact has been stated a few times by theSuede who is in a position to know why this is so technically). When the optical system is really pushed - as it is in some of the situations posted here with fine, intricate branches and so forth - we see the difference. With closer subject matter, the difference is not as noticeable.

I also believe the X-Trans "smoothness" is at least partially a result of the extra, default noise reduction required by the color filter scheme it uses. That's why we see such terrible moire and other color artifacts when certain types of images are put through the raw converters which attempt to extract more detail (PN and Iridient Developer) by "going around" the type of demosaicing and noise reduction that Fuji originally specced (and why it's no coincidence that the Fuji jpegs look the same as say Adobe converted raws). This is exactly why Adobe's raws from X-Trans look as they do. It's a compromise between noise and detail.

Mar 12, 2014 at 03:46 PM

galenapass
Offline
• • • • •
Upload & Sell: Off

p.4 #2 · p.4 #2 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

Tariq Gibran wrote:
I think it's down to just a bit more real resolution in the bayer scheme versus the X-Trans scheme (this fact has been stated a few times by theSuede who is in a position to know why this is so technically). When the optical system is really pushed - as it is in some of the situations posted here with fine, intricate branches and so forth - we see the difference. With closer subject matter, the difference is not as noticeable.

I also believe the X-Trans "smoothness" is at least partially a result of the extra, default noise reduction required...Show more →

If there is a bit more resolution in the bayer scheme, I should be able to look at 100% crops from an X-E2/X-A1 comparison, from an infinity scene, and visually see more resolution from the X-A1, correct?

The following was shot with the 23mm f/1.4 on both cameras. In full disclosure, this was hand-held but the shutter speed was in ~ 1/800 of a sec. Both raws were converted in PN, with sharpening and detail turned off.

Not sure I see resolution advantage for the X-A1 here.

X-E2 file and X-A1 file

To my eye the differences are so minor - it's a wash.

X-A1

X-E2

Mar 13, 2014 at 01:26 AM

justruss
Offline
• • • • •
Upload & Sell: Off

p.4 #3 · p.4 #3 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

Pretty similar Galena. I actually prefer the X-E1 in this case.

I'll not, on top of Tariq's and Dave's respective points: If you use Aperture, go into the RAW Fine Tuning, give the sharpening there a small bump, and give the detail a small bump, you get a much more Bayer/PN-looking image. Aperture has never had any of the watercolor-funk, but it does hit the infinity-resolution wall with an odd smoothing (it almost looks like a combination of CA and camera shake); when bumping the detail and sharpening in RAW Fine Tuning (as opposed to the separate tools in Aperture that also address these things post-RAW Fine Tuning), it really drives the images towards that pointillism noise/detail side of the continuum you see with a sharpened Bayer sensor.

The problem is that if you go too far in this direction it looks equally bad, and you start to get a pseudo-noise even in lower-ISO images.

The point is this though: (at least in Aperture) DO NOT settle for the default RAW Fine Tuning settings. Experiment by the image type. I've started creating a few presets I can hit, or toggle through, that work for certain types of images. Sometimes the standalone tools do the trick better, and sometimes the Fine Tuning does it.

Mar 13, 2014 at 03:05 AM

Sal Baker
Offline
• • •
Upload & Sell: Off

p.4 #4 · p.4 #4 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

The problem for me is all of these converters mostly rely on USM for sharpening. The appearance of sharpness with USM is actually from the artifacts of blurring comparing differences and then increasing sharpness. Presharpening with USM and then performing final sharpening with USM adds even more artifacts. Yes, it looks sharper, but soft detailed has not been reconstructed to less pixels like R-L deconvolution sharpening in Iridient and several other programs.

Here's an interesting look at R-L deconvolution compared to USM. It's not X-Trans specific but the comparisons are clearly illustrated.

http://www.clarkvision.com/photoinfo/image-restoration1/index.html

I've been comparing deconvolution in Iridient with Bayer (Canon CR2 21mp) and X-TransII (X-E2). The X-Trans raw files seem to respond better than Bayer (with AA filter) re reconstructing edges and actually increasing apparent resolution as discussed in the link above. If you use an X-Trans sensor you owe it to yourself to try a free demo of a program that uses R-L deconvolution. It's not perfect, but I find the sharpened results much more "photographic" in edge structure than the "electric" looking details USM sometimes delivers.

Sal

Mar 13, 2014 at 06:59 AM

Tariq Gibran
Offline
• • • • • •
Upload & Sell: Off

p.4 #5 · p.4 #5 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

galenapass wrote:
If there is a bit more resolution in the bayer scheme, I should be able to look at 100% crops from an X-E2/X-A1 comparison, from an infinity scene, and visually see more resolution from the X-A1, correct?.

I don't think you even need to look at pixel level, 100% detail to see the difference with certain subject matter. In your previous examples posted, the fine twigs and branches at distance create a high frequency situation in the detail, particularly where they overlap one another. I see the difference in the 50% crops I posted (as did my wife) and I also see it in the example crops you posted on the previous page. To my eyes, the X-Trans versions fail at resolving areas quite as good as the bayer version with this sort of intricate, fine, organic detail. Man made structures with less organic shapes and less or no high frequency detail are much easier to resolve - and interpolate - so X-Trans does not run into these issues in that case (and you can use the alternative converters in most cases with good results). But ultimately, if you are happy with what you're shooting and don't see the difference, that's all that matters.

Mar 13, 2014 at 07:16 AM

galenapass
Offline
• • • • •
Upload & Sell: Off

p.4 #6 · p.4 #6 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

I am happy with what I am shooting, just trying to understand it better.

....and probably driving a few people crazy with all this pixel peeping.

Mar 13, 2014 at 12:49 PM

galenapass
Offline
• • • • •
Upload & Sell: Off

p.4 #7 · p.4 #7 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

Sal Baker wrote:
The problem for me is all of these converters mostly rely on USM for sharpening. The appearance of sharpness with USM is actually from the artifacts of blurring comparing differences and then increasing sharpness. Presharpening with USM and then performing final sharpening with USM adds even more artifacts. Yes, it looks sharper, but soft detailed has not been reconstructed to less pixels like R-L deconvolution sharpening in Iridient and several other programs.

Here's an interesting look at R-L deconvolution compared to USM. It's not X-Trans specific but the comparisons are clearly illustrated.

http://www.clarkvision.com/photoinfo/image-restoration1/index.html

I've been comparing deconvolution in Iridient with Bayer...Show more →

Thanks Sal, I'll take a look at this next.

Mar 13, 2014 at 12:52 PM

galenapass
Offline
• • • • •
Upload & Sell: Off

p.4 #8 · p.4 #8 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

Well, I've been through the following raw converters in detail: Lightroom, Capture One, and PhotoNinja. I have spent hours working on raws in these converters and have not really achieved exactly what I want. Today I bought a Mac (for a number of reasons) and tried Iridient. I have to say I think my search is over. Iridient is by far the best raw converter for Fuji files so far. Now bear in mind my primary goal has been to make branches and foliage look as natural as possible because these seem to be problematic areas for the X-trans sensor. Iridient had done the job. As I work with this more I may find new issues but for right now I am done. 100% crops.
X-E2

X-A1

Mar 16, 2014 at 01:29 AM

justruss
Offline
• • • • •
Upload & Sell: Off

p.4 #9 · p.4 #9 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

In the above, the X-A1 looks better to me. More natural.

Mar 16, 2014 at 03:31 AM

galenapass
Offline
• • • • •
Upload & Sell: Off

p.4 #10 · p.4 #10 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

The X-A1 picks up more color in some of the fine branches (orange-brown) but the highlights are a little more blown in the X-E2 shot. The X-E2 shot still looks a little more sharpened as well. This likely has more to do with my complete inability to use Aperture right now (took me a 5 minutes to even figure out how to even get a 100% crop!). But, these are close enough, I think, so that in print one would not notice difference.

Tariq mentioned that he bought the X-A1 for ~$450, and then sold the kit lens for $200 netting him an APS-C sensor for $200. If you are an X-trans shooter, seems like a "no brainer" to pick one up as a backup for situations where the X-trans may struggle.

Russ - these do look similar, correct? I am not just fooling myself here?

Mar 16, 2014 at 08:25 AM

Spyro P.
Offline
• • • •
Upload & Sell: Off

p.4 #11 · p.4 #11 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

The Xe looks like half a stop brighter, or the shadows pushed a bit. Also it looks a bit redder for some reason?

Mar 16, 2014 at 08:32 AM

galenapass
Offline
• • • • •
Upload & Sell: Off

p.4 #12 · p.4 #12 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

Yep, I agree. Not only am I terrible with a Aperture, but I am fumbling around with the Mac interface as well. I think Iridient is close, and even better if/when I can get used to the whole platform and program change. I would be nice to see what a competent Aperture user could do. Besides exposure issues, I am not seeing a water color effect on green foliage and the tree bark does not have a "smoothed" look. That is what I am excited about.

Time to take some pictures and let all of this settle for a while!

Mar 16, 2014 at 08:36 AM

Sal Baker
Offline
• • •
Upload & Sell: Off

p.4 #13 · p.4 #13 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

Be sure to try Iridient using R-L deconvolution sharpening. Hybrid sharpening is on by default. It takes a while to see the relationship between "amount" and "iterations" but once fine-tuned the image will look much better than USM-based sharpening. Also, if you interpolate to larger sizes, try one of the many scaling methods with the image pre-sharpened directly to final size/resolution. The result will need little to no final USM sharpening for prints.

Sal

Mar 16, 2014 at 09:06 AM

justruss
Offline
• • • • •
Upload & Sell: Off

p.4 #14 · p.4 #14 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

galenapass wrote:
The X-A1 picks up more color in some of the fine branches (orange-brown) but the highlights are a little more blown in the X-E2 shot. The X-E2 shot still looks a little more sharpened as well. This likely has more to do with my complete inability to use Aperture right now (took me a 5 minutes to even figure out how to even get a 100% crop!). But, these are close enough, I think, so that in print one would not notice difference.

Tariq mentioned that he bought the X-A1 for ~$450, and then sold the kit lens for $200...Show more →

Yes, they do look similar. And I suspect you are right that if these went to print-- they might be difficult to tell apart (or the XT version might look better).

The question of printing XT in difficult situations hasn't-- as far as I've seen-- been answered yet.

Aperture workflow is amazing once you figure it out. With XT, definitely open the RAW Fine Tuning panel and try adjusting Sharpening and Detail from there.

Mar 16, 2014 at 09:36 AM

Sal Baker
Offline
• • •
Upload & Sell: Off

p.4 #15 · p.4 #15 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

justruss wrote:
Yes, they do look similar. And I suspect you are right that if these went to print-- they might be difficult to tell apart (or the XT version might look better).

The question of printing XT in difficult situations hasn't-- as far as I've seen-- been answered yet.

Aperture workflow is amazing once you figure it out. With XT, definitely open the RAW Fine Tuning panel and try adjusting Sharpening and Detail from there.

What difficult printing situation are you referring to?

Sal

Mar 16, 2014 at 09:43 AM

justruss
Offline
• • • • •
Upload & Sell: Off

p.4 #16 · p.4 #16 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

I wasn't being very articulate.

I meant that in difficult imaging conditions and subjects for XT (infinity, fine detail, leaves/branches, resolution limit) I suspect that XT images hold up better when printed large than when displayed at 100% on screen, particularly when using PN, Iridient, or Aperture with bumped RAW fine tuning sharpness/detail.

All I meant in saying that it "hasn't been answered yet" was that we've now seen many, many different shots with XT vs. Bayer in many scenarios on screen at 100%. But we haven't had the same kind of investigation of the same but when printed. I've printed some large images (24 x 36 inch roughly) from XT, but they were images that were not the sort that trigger the funk, the smoothing, the speckle-artifacts.

All of this is to say that I tend to either display images at relatively small size on screen, or print them at various sizes, often larger (except for work when they show up in newspaper, which is low-res medium, or magazine, which is higher but small image size even at double-truck). So it interests me!

Mar 16, 2014 at 12:56 PM

Dave McGaughey
Offline
• • •
Upload & Sell: Off

p.4 #17 · p.4 #17 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

I know the trend here is against ACR, but I've been tinkering with different sharpening approaches and I find this works much better than anything I can get out of the ACR sharpening panel:

ACR
Turn off sharpening and color noise reduction

Photoshop
Smart Sharpen: 100%, 1.5 radius, Remove: Lens Blur, More Accurate (checked)
Smart Sharpen: 300%, 0.2 radius, Remove: Gaussian Blur, More Accurate (UNchecked)

Mar 16, 2014 at 08:03 PM

kwoodard
Offline
• • • • •
Upload & Sell: On

p.4 #18 · p.4 #18 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

I know this is an oldish thread, but I am going to take a crack at the latest image posted by Galenapass. I will post my results this evening. I seem to be in the minority on what I process XTrans with... I use DCRaw to generate a TIFF then process that in either Photoshop or Lightroom.

May 01, 2014 at 10:57 AM

kwoodard
Offline
• • • • •
Upload & Sell: On

p.4 #19 · p.4 #19 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

Here is what I came up with using a process I came up with using DCRaw and Photoshop, then touched up a little bit in LR to attempt to match the color from Galenapass' crop.

X-E2 Bldg_ by Kevin.Woodard, on Flickr

May 03, 2014 at 03:58 PM

cputeq
Offline
• • • • •
Upload & Sell: Off

p.4 #20 · p.4 #20 · Fujifilm X-Trans Infinity Scene RAW Processing Comparison

^ Not bad.

I'll have to play around some hopefully this week. I thought I liked Irident the best, but it seems to apply a bit of tonal contrast / extra sharpening that I don't like. Out of the box at least, PhotoNinja seems to be giving me slightly better results. Both are usually better than LR 5.4 though.

We're splitting hairs here though - both good converters.

May 05, 2014 at 08:17 AM

1 2 3 4 5 end
FM Forums \| Leica & Alternative Gear \| Join Upload & Sell 1 2 3 4 5 end

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.