Fuji shooters, grab Iridient X-Transformer now!

[taemo]

Found out about this on FredMiranda this morning.
http://www.iridientdigital.com/products/xtransformer.html
Basically you buy and install this program and convert your Fuji RAF files to DNG before importing and processing them into Lightroom or whichever photo editing program you use.
I can see a big improvement on how LR renders the converted DNG file vs RAF.

For comparison here are 2 snapshots I took of DNG(left) vs RAF(right)
DNGvsRAF by Earl Dieta, on Flickr
DNGvsRAF2 by Earl Dieta, on Flickr

 

[taemo]

bump, here are comparisons between DNG and JPG (I tried to match the PP that Acros Y did, didnt apply any sharpness)

DNGvsJPG by Earl Dieta, on Flickr

DNGvsJPG2 by Earl Dieta, on Flickr

 

[krötenblender]

There is currently a similar thread on a German forum and for both, there and here, I fail to see a very big difference. For your DNG vs. RAF this may be because of the small format of the images (1:1 crops of the problematic areas may help).

In the second comparison I clearly prefer the jpgs out of camera.

 

[taemo]

All print screens I posted were 1:1 crops except the 2nd one which was a 1:2 crop
For the JPG, is it the tonality of the B&W that you prefer or the sharpness?
As for me, I definitely find the DNG versus the RAF and SOOC JPG to be sharper.

note, the JPG comparisons also weren't on the center of the image but on the side/corner.

 

[user237428934]

In you first post the snow photo looks a bit better with Iridient, more details in the snow. To me the coast photo is unusable in both version because of the blurry green trees. No Iridient can rescue this mess.

 

[______]

Reading the blurb on the website, it says there are two algorithms, one "detail" and the other "smooth". In addition, it says that the program applies sharpening and noise reduction. I'd prefer to have the raw RAW file and do my own sharpening/noise reduction rather than having the program do it for me.

 

[willie_901]

... didnt apply any sharpness)

Capture sharpening's purpose is to compensate for the fact that demosaiking algorithms are not able to precisely represent a continuous phenomenon using incomplete (i.e. discrete) data. Capture sharpening is a form of spatial mathematical filtering.

So if these two demosaiking algorithms use fundamentally different mathematical models, why wouldn't their optimum sharpening parameters be different? The need for post-rendering sharpening does not necessarily mean a demosaiking algorithm is inferior, similar or superior to one that does not.

Do we know if Iridient applies some form of 'sharpening' during the DNG transformation? Data filtering is typically part of the Demosaiking process.[1].

It is difficult to rigorously compare two completely different rendering platforms. There are numerous variables that affect our perception of images rendered by different platforms. It is likely the parameters that affect these variables behave differently when the demosaiking algorithms are different.

[1]Technical Link (sharpening section starts at slide #24)

 

[______]

This is the Iridient site says:

The key RAW processing stages that can be applied by Iridient X-Transformer during conversion to DNG include:

Interpolation or demosaic processing. This transforms the camera sensor's X-Trans or Bayer color filter information to produce a full color RGB image. This processing is always applied by Iridient X-Transformer and cannot be disabled or undone. There are two interpolation options, one for a more detailed rendering and a second smoother, lower artifact version. The algorithms are identical to the latest "detail+" and "smooth" options in Iridient Developer 3.1.2.

Sharpening. Iridient X-Transformer features advanced sharpening based on the very highly regarded "Iridient Reveal" sharpening algorithms featured in Iridient Developer. Sharpening is optional and can be adjusted (low/med/high) or disabled.

Automatic lens corrections for distortion, chromatic aberration and vignetting. Corrections are based on native Fujifilm lens information specified in their RAF metadata. The lens correction stage is optional and lens correction information can also be passed on through DNG opcode metadata and left to later processing stages or ignored altogether. The lens correction processing in Iridient X-Transformer uses the same high quality resampling algorithms as Iridient Developer.

Color and luminance noise reduction. The noise reduction processing in Iridient X-Transformer shares much in common with the latest noise reduction available in Iridient Developer and is specifically optimized for RAW image processing. Noise reduction is optional and can be adjusted (low/med/high) or disabled.

So it looks like it does demosaicing sharpening, and then additional sharpening and noise reduction.

 

[kai.zorki]

You can turn that all off, the process is very customisable. I find that it gives slightly better results, but it is still a beta and my testing wasn't extensive, so it should improve even further.

 

[shawn]

v3 of the beta just came out and this also includes a MacOS version. If you have a current Iridient Developer license you can get Iridient Transformer for 40% off. I just bought a copy and with the discount it is $19.11.

Already has support for x100f, XT20 and GFX 50S.

Shawn

 

[MaxElmar]

Unfortunately, the already large RAF files get even larger as DNG! I guess if you are dedicated to using LR and X-Trans it's ok. Still evaluating. They do look good. X-Trans 2 (16mp) RAF files go from ~34mb to ~46mb with lossless compression. Compression off they jump to ~86mb. Iridient Developer can output compressed TIFFS ~76mb. These figures are all 16 bits per channel files. You can chop off bits for smaller files, too - there is a 10 bit option for DNG - this gives you ~30mb files. I have not yet evaluated the impact of loosing those 6 bits.... my Mac does NOT like the 10 bit DNGs...

The main advantage seems to be in the everyday file handling - everything seems to render much faster as DNG than RAF. (Like 2.5s vs. 9sec in preview!)