Fuji explains why GFX improves on others using 50mp MF sensor.

[Iktinos]

In this article, GFX Technologies #2 | X Stories | FUJIFILM X (2 pages) Fuji explains why their design will perform better that the rest based on the same Sony origin sensor.

It is explained why longer flange distance is used, why there will be less cross-talking between pixels, why there will be less color casts (tech camera users pay attention) and why the result will be sharper. It seems that (long) 26.7mm flange distance was a choice among other things that are very well (and scientificantly) explained in a simple but perfectly backed up with reasoning way.

 

[iiiNelson]

The proof is in the pudding. A Cambo rep and others have chimed in with some specific solutions and challenges to overcome which aren't possible yet if ever.

I'm not saying that it won't perform better but microlenses are generally designed for use of the correction of optical errors for native lenses. Many cameras have microlenses but may not function flawlessly with T/S lenses or adapters without heavy software correction.

 

[ErikKaffehr]

Hi,

Microlenses do not correct for optical errors of lenses. Their intention is to direct the incoming light on the active area of the sensor, and also to keep the incoming light away from the neighbouring pixels.

Obviously, the microlenses are designed for some outlet pupil distance.

Best regards
Erik

The proof is in the pudding. A Cambo rep and others have chimed in with some specific solutions and challenges to overcome which aren't possible yet if ever.

I'm not saying that it won't perform better but microlenses are generally designed for use of the correction of optical errors for native lenses. Many cameras have microlenses but may not function flawlessly with T/S lenses or adapters without heavy software correction.

 

[daf]

Hi,

Microlenses do not correct for optical errors of lenses. Their intention is to direct the incoming light on the active area of the sensor, and also to keep the incoming light away from the neighbouring pixels.

Obviously, the microlenses are designed for some outlet pupil distance.

Best regards
Erik

And until now, with mfdb and symetrical lens : more microlens = more lens cast

 

[iiiNelson]

Hi,

Microlenses do not correct for optical errors of lenses. Their intention is to direct the incoming light on the active area of the sensor, and also to keep the incoming light away from the neighbouring pixels.

Obviously, the microlenses are designed for some outlet pupil distance.

Best regards
Erik

understand that and I meant the optical effects of the combination of lens design in combination with the sensor. For instance to straighten and direct the light of an otherwise non-optimum design for a variety of reasons.

 

[Iktinos]

The color casts that require LCC to be applied, are not due to "lens optical errors", if they were, the problem would be evident with older CCD MFDBs (with larger pixels and no microlenses present) or on film too, however they are not. The phenomenon is a consequence of increased cross talking between pixels on modern sensors and therefore the shape and size of microlenses directly affects (magnifies or reduces) the phenomenon.

On the GFX, special care has been given (according to Fuji's article) to the sensor's microlenses shape and size (reduced) which should reduce cross talking and therefore color casts.

 

[daf]

The color casts that require LCC to be applied, are not due to "lens optical errors", if they were, the problem would be evident with older CCD MFDBs (with larger pixels and no microlenses present) or on film too, however they are not. The phenomenon is a consequence of increased cross talking between pixels on modern sensors and therefore the shape and size of microlenses directly affects (magnifies or reduces) the phenomenon.

On the GFX, special care has been given (according to Fuji's article) to the sensor's microlenses shape and size (reduced) which should reduce cross talking and therefore color casts.

Fuji has probably done the same kind of job, as leica has on their M.
But keep in mind "shift" is another story

 

[iiiNelson]

Fuji has probably done the same kind of job, as leica has on their M.
But keep in mind "shift" is another story

...and Sony for that matter. case being that microlenses are primarily designed for sensors to work with and correct flaws of native lenses. If/when a particular lens works with a particular sensor and lens combo, then that's a bonus versus the design.

I just recommend that people manage their expectations before testing is completed.

 

[2WK]

Fuji was also very good at explaining how Xtrans sensor tech was the best thing since sliced bread...

So surely they must have used Xtrans in the GFX right? Or have they dropped it like I did after not being impressed...

In this article, GFX Technologies #2 | X Stories | FUJIFILM X (2 pages) Fuji explains why their design will perform better that the rest based on the same Sony origin sensor.

It is explained why longer flange distance is used, why there will be less cross-talking between pixels, why there will be less color casts (tech camera users pay attention) and why the result will be sharper. It seems that (long) 26.7mm flange distance was a choice among other things that are very well (and scientificantly) explained in a simple but perfectly backed up with reasoning way.

 

[Puma Cat]

Fuji was also very good at explaining how Xtrans sensor tech was the best thing since sliced bread...

So surely they must have used Xtrans in the GFX right? Or have they dropped it like I did after not being impressed...

Fuji is on record with stating that the reason they did not use X-trans on the GFX sensor is that it would require too much processing power. This could have an impact on several different performance attributes as well as battery life.

 

[Paul2660]

I am not a big fan of X-trans having used it since Fuji's old days with the camera based on the Nikon N-90 and all the way through the X-T2. Personally I was very happy to see Fuji NOT use X-trans on the GFX.

On the X-trans side, Fuji is now taking about x-Trans on the X-T2 follow on camera so they are still going to use it.

Paul Caldwell

 

[dnercesian]

In my opinion, Fuji is still using the X-Trans CFA in the small sensor cameras to save face. I have used and owned most of what they have produced and have never favored the X-Trans raw files. Their stated purpose of using it in the beginning was to eliminate or minimize moire since they were not using a AA filter. The CFA did not accomplish that. But it didn't matter. In cult like fashion, many users started claiming the X-Trans CFA was responsible for a number of qualities that even Fuji wasn't going to touch with a ten foot pole. I have always believed and still believe that this is a gimmick, and dropping it would be admitting something that the company simply does not wish to admit.

They didn't put it on the GFX for a number of reasons, and they stated that it was because of processing power. I am sure that is a correct assessment, but I am also sure that pros and others spending that much money on a system do not want to do the constant raw converter dance that the APSC Fuji community has been doing for years. Personally, I don't want to play that game. I need to work, and I need my gear to just flat out work. I gave them several tries because there are other aspects of their cameras that I REALLY like, but every time I did, I ended up reversing the decision for this reason.

So here they are again claiming some special sauce on the GFX and I am skeptical. Having shot the IQ150 and the Pentax 645Z, both with amazing performance, I think this Fuji will be right on par. I don't know why Fuji feels the need to market like this. It is almost as if they are insecure about the fact that they use someone else's sensor, which is unnecessary because I believe they actually do a lot of things very right when they make cameras. Let those things set you apart. Just my 2 cents...

 

[Paul2660]

Iridient Transforner for x gives me what I have been looking for. Great batch setup and files are ready for LR.

Paul Caldwell

 

[dnercesian]

Forgot to mention for the record. Although I didn't care for X Trans raw files, Capture One actually does a great job with them. Best I have seen/used.

 

[Iktinos]

Apparently, according to imaging resource, the Fuji files out of the GFX look to clearly beat the Pentax 645z https://www.imaging-resource.com/ne...irst-shots-the-best-lab-images-weve-ever-seen

Imaging resource claims: "the best lab images we've ever seen". It looks like it may be the case that the modifications Fuji applied on the Sony sensor, may actually work quite well.

 

[Tmuussoni]

Apparently, according to imaging resource, the Fuji files out of the GFX look to clearly beat the Pentax 645z https://www.imaging-resource.com/ne...irst-shots-the-best-lab-images-weve-ever-seen

Imaging resource claims: "the best lab images we've ever seen". It looks like it may be the case that the modifications Fuji applied on the Sony sensor, may actually work quite well.

I looked at the samples. Fuji definitely has small edge at high ISO over Pentax. But at least I am glad Fuji does not seem to overstating ISO values this time, they look about same in terms of brightness. So is it a "property" of X-trans sensors to have overstated ISO values? I certainly am glad that Fuji did not go the X-trans route, I have yet to seen any evidence of X-trans benefits (apart for less Moire at low MP sensors, like 12 MP). While there certainly are some very well known disadvantages with X-Trans... And Fuji 63/2.8 seems to be better lens than Pentax 55/2.8.

Edit: on second look it does seem like the Pentax lens is sharper on the background (studio scene with the napkins), so perhaps a slight difference in where the focus point is.