GFX100RF for Fine Art, tripod work - first look

[vonalpen]

I have been exploring the "rainbow effect" which you see in these images here when shooting directly into the sun for years now.

They appear on all sensors from any manufacturer, more or less obviously.

When I aquire a new camera and / or a new wideangle lens, I'm always interested to find out how the results are shooting straight into the sun and how well defined the sunstars are.

Mind you, these examples are from a worst case scenario!
The front lenses were all clean.

To be able to produce this collection quickly, I opened all four images in LR Classic, although I use other RAW converters as well.
So please don't start comparing the colors of theses four images here, as I have not taken any attempt in equalizing them.
For me this comparison is purely about the rendering of the sunstars and the "rainbow effect", which some describe as PDAF-banding.
But for me that "rainbow effect" got nothing to do with the PDAF array on these sensors, as they also appear on sensors without PDAF, like Phase One backs and Leica M cameras.

From left to right (all at f11):
Hasselblad XCD 25mm V, XCD 38mm V (both shot on the X2DII), Fuji GFX100RF (35mm), and Tamron 25-200mm (at 30mm, shot on a Sony A7CR).
Sorry for also including a full frame camera with such a cheap lens

Best regards

Jost

 

[Ben730]

But for me that "rainbow effect" got nothing to do with the PDAF array on these sensors, as they also appear on sensors without PDAF, like Phase One backs and Leica M cameras.

Jost, I've never had this effect with a Phase One back, only with my Nikons and Fujis.
I am annoyed by these rainbow patterns, but I know that PDAF has a price, but also many, many advantages.
Are you sure you had that with a Phase One back too?

 

[darr]

PDAF aberrations typically look like:

• Subtle striping
• Grid or dotted patterns
• Low-contrast texture irregularities
• Often visible only after heavy shadow lifting
• Usually fixed to the sensor geometry, not radiating from a light source

Lens Flare:

• Bright, colorful, organic shapes
• Radial alignment with the sun
• Strong color saturation
• Visible straight out of camera
• Varies in size and position

It looks like lens flare to me.

 

[wattsy]

For me this comparison is purely about the rendering of the sunstars and the "rainbow effect", which some describe as PDAF-banding.
But for me that "rainbow effect" got nothing to do with the PDAF array on these sensors, as they also appear on sensors without PDAF, like Phase One backs and Leica M cameras.

I don’t know whether it is caused by PDAF or not but I haven’t seen that kind of effect from any of the digital cameras I have used (which I don’t think includes any with a PDAF array). I shoot into the sun very frequently and encounter all kinds of flare, good and bad.

Your image examples are quite interesting. The two X2D photos do show a regular pattern in the flare/sun rays but the GFX100RF example is noticeably more severe. I wonder whether this effect is exaggerated by the design of this particular lens?

 

[wattsy]

PDAF aberrations typically look like:

• Subtle striping
• Grid or dotted patterns
• Low-contrast texture irregularities
• Often visible only after heavy shadow lifting
• Usually fixed to the sensor geometry, not radiating from a light source

Lens Flare:

• Bright, colorful, organic shapes
• Radial alignment with the sun
• Strong color saturation
• Visible straight out of camera
• Varies in size and position

It looks like lens flare to me.

Well, there is obviously some flare but we are talking about the regular pattern of red and green ‘stripes’. That cannot be a product of the lens alone but is presumably either a sensor artefact or, more likely, resulting from the lens and sensor combination. It’s not an effect I remember seeing with film.

 

[darr]

Well, there is obviously some flare but we are talking about the regular pattern of red and green ‘stripes’. That cannot be a product of the lens alone but is presumably either a sensor artefact or, more likely, resulting from the lens and sensor combination. It’s not an effect I remember seeing with film.

Hasselblad 907x + 28P

 

[vonalpen]

Darr: Thank you for your image examples and comments on PDAF aberrations and lens flare.
The XCD 28mm actually shows very little lens flare.
And the Hasselblad 907X has no PDAF pattern!

After Ben's question about the rainbow effect, I searched my archives: In fact, the rainbow effect only seems to occur with newer cameras and with sensors with PDAF arrays.
So not with Phase One backs or Leica M cameras.
I have crossed out my comment above in my post accordingly.

And this thread should mainly be about the Fuji GFX100RF anyway....

best regards

Jost

 

[wattsy]

Hasselblad 907x + 28P

I’m not sure what your point is, Darr (or why you seem to be starting an argument)? Those photos show what I consider to be perfectly reasonable flare artefacts for a digital camera pointing into the sun. Jost’s example earlier has a much more noticeable grid pattern of red and green.

 

[darr]

I’m not sure what your point is, Darr (or why you seem to be starting an argument)? Those photos show what I consider to be perfectly reasonable flare artefacts for a digital camera pointing into the sun. Jost’s example earlier has a much more noticeable grid pattern of red and green.

I want to clarify my intent here, because I think it may have come across differently than I meant.

I wasn’t trying to argue against anyone’s interpretation or suggest there’s only one possible explanation. I was simply explaining why I personally read the artifacts in that image as optical flare rather than PDAF.

The way I approached it was by asking myself whether the artifacts would move or change if the sun were repositioned in the frame. In this case, they appear tied to the light source — bright, saturated, and radially aligned — which matches what I’ve consistently seen from lens flare and internal reflections in digital systems.

When I’ve encountered PDAF artifacts, they’ve tended to show a fixed, repeating structure tied to sensor geometry, which is why I mentally separate the two.

That’s all I was trying to contribute — not to pick an argument, just to explain the reasoning behind my conclusion based on my experience.

 

[vieri]

On the colour issue, I agree completely with @rdeloe. Any camera comes with an idea of what colours should look like baked into their RAW files; and, any RAW file can then be made to look any way one likes, given a bit of proficiency with one's RAW development tool.

On top of that, there is personal preference involved: someone feels "closer" to the idea of colour coming from cameras made by brand X, someone else feels closer to colours coming from brand Y; and, since we all love to take the lath of least resistance when being creative, and since nobody wants to do more work than it's necessary to get from A to B, then each of us has a brand that looks "better" to them. There is no brand which colours are objectively "better". since there is no such thing as "objectively" when it comes to artistic expression.

Back to the GFX100RF, I shoot the camera in black & white, therefore the whole discussion is moot to me, since I haven't even tried to process a single colour image (and I am currently not interested in doing that).

That said, and more in general, n case anyone were interested in my personal preferences about colours, I always preferred cameras that do not "nudge" me towards any particular post-processing direction; the "blander" the RAW, so to speak, the better for me.

Over the 20 years of my digital career, I used for my professional work Leica (SL, S(Typ 007) and any colour M up to the M10); Pentax (645D and 645z, and K3); Sigma SD1; Hasselblad (X1D and X1D II); Nikon (D70, D800E and any single digit D up to the D3/D3x); Phase One (P45+, P65+, IQ4 150); and, Fujifilm GFX100S II to digitise my negatives. Plus, for my personal work and pleasure, I used the Nikon P7000, the Sony Nex7, the original Fujifilm X100 and the original Fujifilm Xpro. For those interested in my views on them, many of these cameras have been reviewed on my blog, here: https://www.vieribottazzini.com/blog

I can easily say that, except for Canon digital cameras, I worked with different models made by pretty much any brand, and I have a pretty good idea about each brand's colour science.

As someone looking for files that do not push me towards a given "look", and leave me freer to go for what I want in the easiest way, to me the easiest files to work with were the Hasselblad & Pentax files; then the Leica ones; then Nikon and Fujifilm; then Phase One; last, Sigma and Sony. That has nothing to do with the end result: I do have portfolio work created with all these cameras, which I equally love. It's just something to do with the ease of getting there for me personally, and for the time it took me to "get to grips" with a new camera's files; once I did that, it was just as easy and fast to work with any of the above mentioned cameras.

Best regards,

Vieri

 

[drevil]

those green and red sun flares, have something to do with the sensor or sensor glass, and they were part of the reason why i left the Fuji system again.
As Paul mentioned "deal breaker"

 

[SrMphoto]

Back to the GFX100RF, I shoot the camera in black & white, therefore the whole discussion is moot to me, since I haven't even tried to process a single colour image (and I am currently not interested in doing that).

If you are shooting B&W with RF, reidreviews.com may be worth subscribing as Sean Reid discusses new Iridient tool features that help with B&W on RF (better demosaicing).

 

[vonalpen]

Sean Reid confirms that the GFX100RF has great potential for B&W as well.
And I wonder if Vieri will come to the same conclusion.
As I'm on Windows, I can't comment on the possibilities with the Iridient tools, as they are Mac only.
Here are a few examples which I took with the GFX100RF last autumn.
Prints look nice, too.

Jost

 

[miriquidi]

View attachment 225835

That's not a PDAF-artifact.
PDAF works in letting groups of pixels just see light from the left or right side of the lens (or the upper or lower side). So pair of two adjacent pixels works like the classical phase AF of a DSLR. Not that precise, but you have many thousands of these groups on a PDAF image sensor.

There are different implementations of PDAF. The most common is, to physically block the light, that comes from one side of lens. You later interpolate and recalculate the image so, that you don't see that pattern for the "normal light path". That digital compensation doesn't work, if the light shines steep from the side to the sensor. Than you can see PDAF-artifacts. Here is a good sample picture, search for "PDAF artifact" in that page.

https://www.dpreview.com/reviews/olympus-om-d-e-m5-iii-review/4

The better way to implement PDAF is to use a double pixel structure behind each sensor picture element (Canon, especially the R1 is a brilliant implementation) or use Quad-pixel AF (OM System). So you don't block light, you can just read how much is coming from the left and the right. And for the image taking you sum up the two (or four) sub-pixel-elements.
Then you don't see PDAF-artifacts, just the normal lens colour cast as with a Phase One body. The drawback is, that you need two times the pixel (or four times in case of OM System)


What you see in the image above is the effect of periodic stray light. An image sensor is regular pattern of small elements (a few micrometer in period). Regular patterns interact with light and create something, that is called diffraction. If you use your blank image sensor to mirror a bright light source in a dark room, you can see that rainbow pattern around the mirror light source with your eye. Little rainbows (in case of a LED light). As it's a 2D pixel-structure, you see two directions of diffraction.
In that picture you have a brutal bright light source, hitting the sensor. A part of that incomming light is getting reflected back to the lens, another part gets diffracted back to the lens. And depending on the lens design, the coatings and so on, a part of that back reflected and back diffracted light is hitting the sensor again. To find the light paths, is very complicated.
You can omit this effect by using a well blacked pinhole lens.

With film instead of CCDs or CMOS you don't see such patterns, as film grain is not periodic structured but randomly immersed. Random structures create just a diffuse back reflection, but no rainbow.

The only solution I know (which works in this case) is to take two frames. One with the sun and one with blocking the sun.

I see a lot of GFX100RF bodies for sale on MPB so wonder if there is an obvious cause of regret with this camera? Personally, I usually live in single focal length worlds for certain types of photography so the limitation of the 35mm/F4 isn't a deal-breaker for me.

That's not a camera I would desire or would ever afford, but I played a few minutes with it. For the shooting style (fixed focal wide angle) it's quite large and heavy. And I also found it quite slow. (Also the GFX 100s with the 80 mm F1.7 is soo slow in focussing, and I use small sensor stuff one decade old). On the other side you get that huge resolution from the GFXs.
That speed factor doesn't bother for landscape and so on, but for socials I would miss something. I think there are a few people, who underestimated that.

 

[vonalpen]

Hello miriquidi,

Thank you for the detailed information.
It was always clear to me that the rainbow pattern in my extreme examples had nothing to do with the PDAF pixels.

The GFX100RF is by far the smallest medium format camera on the market. When you consider that it has a built-in 35mm lens, it's an amazingly compact combination.
The AF is much faster than, for example, a GFX100 with the GF 1.7/80mm, but of course clearly slower than any current full-frame camera.
For landscape photography, the AF is absolutely adequate and the camera also responds quickly.

Best regards,

Jost

 

[vieri]

If you are shooting B&W with RF, reidreviews.com may be worth subscribing as Sean Reid discusses new Iridient tool features that help with B&W on RF (better demosaicing).

Thank you for the tip, I have been a subscriber of Sean's for ages, but stopped a few years ago. About Iridient, and more in general about this or that software that people recommends online, I am a believer that marketing hype aside, or personal reviewer's preferences aside, differences in between tools are quite small in practice between Adobe and anything else these days. I also believe that mastering one's tools pretty much always results in better results than using a tool that gives marginally better results but which you either don't know how to make the most of, or need to spend a lot of time to make the most of (time which I don't have). I have been using Camera RAW for ages, and I am happy with what I see; the only plugin I use is Nik Silver Efex for B&W, and even that I have been using less and less recently.

Sean Reid confirms that the GFX100RF has great potential for B&W as well.
And I wonder if Vieri will come to the same conclusion.
As I'm on Windows, I can't comment on the possibilities with the Iridient tools, as they are Mac only.
Here are a few examples which I took with the GFX100RF last autumn.
Prints look nice, too.

Jost

Indeed, I have been loving what I saw coming out of the GFX100RF so far - I will have the camera until I came back home in late March, and after that I will prepare a detailed write up for my blog, with photographs of course, about my experience with it. Even if it will be limited to a very small user case, that of more "formal" landscape work on a tripod, I think it might be of interest.

Best regards,

Vieri

 

[wattsy]

those green and red sun flares, have something to do with the sensor or sensor glass, and they were part of the reason why i left the Fuji system again.
As Paul mentioned "deal breaker"

I must admit it has put me off. I photograph towards the light often (admittedly mostly towards more diffuse and shielded lighting so sunstars are not a regular feature) and would not welcome this red/green grid pattern in the flare. Jost's examples in post #21 suggest that it is also an unwelcome feature of similar photos taken with the X2D. Whether it is a byproduct of PDAF sensors or not (I note miriquidi's comments above) it is not something I have seen using my non-PDAF Hasselblad, Leica and Nikon sensors over many years. I made a point of looking back through some of my 907x 50c II photos over the last three years and have plenty of sunstar flare examples but none with this grid pattern. I think I'll stick with what I already have.

 

[rdeloe]

...I am a believer that marketing hype aside, or personal reviewer's preferences aside, differences in between tools are quite small in practice between Adobe and anything else these days. I also believe that mastering one's tools pretty much always results in better results than using a tool that gives marginally better results but which you either don't know how to make the most of, or need to spend a lot of time to make the most of ...

I could not agree more. Whether software for processing images, or the cameras themselves, these are complicated tools. Learning how to use them well and understanding deeply how they work takes a long time. Constantly searching for the next new thing uses up time that could be spent learning how to use what you have.

I don't know anyone who isn't using Microsoft Word at work. However, back in the misty dawn of time, Word had competitors. People would actually wrangle over whether Word, Word Perfect, Word Pro, etc. was the best. As a known computer nerd, colleagues would ask my opinion. When I was in a mood to be blunt, I would say, "It doesn't matter because you're only going to use 5% of the capabilities anyway. Just pick one and learn it."

 

[mristuccia]

I don't know anyone who isn't using Microsoft Word at work.

I'm using LaTeX.
A bit of joking of course, but not too much. The HBComposer PDF manual is truly done with LaTeX. Same for all my invoices and all printed letters I send here and there.
Speaking of small details... I still find the way LaTeX places & kerns all characters together the best in town.

 

[vieri]

I could not agree more. Whether software for processing images, or the cameras themselves, these are complicated tools. Learning how to use them well and understanding deeply how they work takes a long time. Constantly searching for the next new thing uses up time that could be spent learning how to use what you have.

I don't know anyone who isn't using Microsoft Word at work. However, back in the misty dawn of time, Word had competitors. People would actually wrangle over whether Word, Word Perfect, Word Pro, etc. was the best. As a known computer nerd, colleagues would ask my opinion. When I was in a mood to be blunt, I would say, "It doesn't matter because you're only going to use 5% of the capabilities anyway. Just pick one and learn it."

Indeed!

I might add, re: RAW editors and the like, that it makes sense - software capabilities being equal, or thereabouts - to invest our learning time in a software that is likely to stick around, to avoid wasting learning time for nothing. Hence my preference for Camera RAW, which is competent enough and it's likelier to be still around long after I am gone than other options.

No offence intended to the valiant developers of other alternatives, of course, but e.g. as a Nikon user I do remember dedicating time to learn Nikon Capture NX, now defunct; that taught me not to trust camera softwares ever again, camera makers should make cameras. The Nik plugin suite was another cautionary tale, happily resurrected by DxO but not developed for a very long time (now it's again alive). And so on...

Of course, there is the argument that if nobody supports smaller developers, there won't be any, and we will be at the whim of Adobe even more than we already are; I completely see that, of course, as I see how smaller developer can be more creative in new features and so on, or better, because they need to compete with the behemoths and they can only do so by been innovative and unique, whereas Adobe could just sit on their laurels. But, they don't look like they are doing that, and they offer me all I need (until someone will offer something that I didn't know I needed, that is ), so for now I am very happy to dedicate my time to master Camera RAW and Photoshop rather than disperse it around.

Best regards,

Vieri