The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

GFX microlens discussion at Kasson

P. Chong

Well-known member
I was wondering what you guys think of the thoughts expressed in the following article. I believe Lloyd Chambers also wrote something similar.


https://blog.kasson.com/gfx-50s/on-microlens-size-in-the-gfx-100-and-gfx-50r-s/

https://diglloyd.com/blog/2019/20190818_0900-FujifilmGFX100-vs-FujifilmGFX50S.html

If as Kasson suggested, the issue with the GFX50 is due to the microlenses on the sensor, is the same argument also applicable to the X1D sensor, since they are essentially the same. I haven't shot with GFX50, but I don't think I am seeing the crunchiness or oversharpening that is being discussed. Thoughts, anyone?
 

hcubell

Well-known member
I was wondering what you guys think of the thoughts expressed in the following article. I believe Lloyd Chambers also wrote something similar.


https://blog.kasson.com/gfx-50s/on-microlens-size-in-the-gfx-100-and-gfx-50r-s/

https://diglloyd.com/blog/2019/20190818_0900-FujifilmGFX100-vs-FujifilmGFX50S.html

If as Kasson suggested, the issue with the GFX50 is due to the microlenses on the sensor, is the same argument also applicable to the X1D sensor, since they are essentially the same. I haven't shot with GFX50, but I don't think I am seeing the crunchiness or oversharpening that is being discussed. Thoughts, anyone?
Only Fuji customized the size of the microlenses on the Sony 50 mp sensor used in the GFX 50S and 50R. The Sony sensor used in the X1D is different in that respect. Fuji’s marketing campaign made a point of emphasizing how it had customized the microlenses, leading to the appearance of sharper images. When I tested the GFX 50 S side by side with the X1D back in early 2017, I did notice that the Fuji files using LR default sharpening settings appeared “sharper” than the X1D files with LR defaults. However, with sophisticated capture sharpening, the X1D files improved very significantly and looked to me more natural. I recall questioning Mr. Kasson at the time about the potential downsides of how Fuji had reduced the size of the microlenses. After all, Hasselblad and Phase had used the same sensor and had never reduced the size of the microlenses. There is generally no free lunch in such matters. Mr. Kasson at the time seemed to have fully bought into Fuji’s secret sauce, and I think he said there was a theoretical possibility of artifacts in some corner cases but it was unlikely to be an issue in practice. He stated that sophisticated capture sharpening could level the playing field somewhat, but it was always better to start with a raw file that was “inherently “ sharper. Now, 2 and 1/2 years later, we are hearing second thoughts from two sources.
IMO, this is not all that big a deal. Most of the world is in love with oversaturated and oversharpened images.
 

P. Chong

Well-known member
Ah, I see...thanks for the reply. Makes sense. I always thought that the X1D images were better looking than the GFX, though I am not able to describe why. Now, I know.


Only Fuji customized the size of the microlenses on the Sony 50 mp sensor used in the GFX 50S and 50R. The Sony sensor used in the X1D is different in that respect. Fuji’s marketing campaign made a point of emphasizing how it had customized the microlenses, leading to the appearance of sharper images. When I tested the GFX 50 S side by side with the X1D back in early 2017, I did notice that the Fuji files using LR default sharpening settings appeared “sharper” than the X1D files with LR defaults. However, with sophisticated capture sharpening, the X1D files improved very significantly and looked to me more natural. I recall questioning Mr. Kasson at the time about the potential downsides of how Fuji had reduced the size of the microlenses. After all, Hasselblad and Phase had used the same sensor and had never reduced the size of the microlenses. There is generally no free lunch in such matters. Mr. Kasson at the time seemed to have fully bought into Fuji’s secret sauce, and I think he said there was a theoretical possibility of artifacts in some corner cases but it was unlikely to be an issue in practice. He stated that sophisticated capture sharpening could level the playing field somewhat, but it was always better to start with a raw file that was “inherently “ sharper. Now, 2 and 1/2 years later, we are hearing second thoughts from two sources.
IMO, this is not all that big a deal. Most of the world is in love with oversaturated and oversharpened images.
 

JimKasson

Well-known member
Only Fuji customized the size of the microlenses on the Sony 50 mp sensor used in the GFX 50S and 50R. The Sony sensor used in the X1D is different in that respect. Fuji’s marketing campaign made a point of emphasizing how it had customized the microlenses, leading to the appearance of sharper images. When I tested the GFX 50 S side by side with the X1D back in early 2017, I did notice that the Fuji files using LR default sharpening settings appeared “sharper” than the X1D files with LR defaults. However, with sophisticated capture sharpening, the X1D files improved very significantly and looked to me more natural. I recall questioning Mr. Kasson at the time about the potential downsides of how Fuji had reduced the size of the microlenses. After all, Hasselblad and Phase had used the same sensor and had never reduced the size of the microlenses. There is generally no free lunch in such matters. Mr. Kasson at the time seemed to have fully bought into Fuji’s secret sauce, and I think he said there was a theoretical possibility of artifacts in some corner cases but it was unlikely to be an issue in practice. He stated that sophisticated capture sharpening could level the playing field somewhat, but it was always better to start with a raw file that was “inherently “ sharper. Now, 2 and 1/2 years later, we are hearing second thoughts from two sources.
IMO, this is not all that big a deal. Most of the world is in love with oversaturated and oversharpened images.
What you say about the microlenses in the X1D vs the GFX 50x is true, as far as I know. I believe what you have said about my thoughts about the GFX 50S at the time of its introduction is a mischaracterization. I was always aware of the downsides of the small microlenses, and made at least one post in my blog illustrating somewhat increased false color and moire.

https://blog.kasson.com/the-last-word/fujifilm-gfx-sony-a7rii-moire/

In that post, I said: "We have seen that the GFX sensor produces much sharper images than the a7RII sensor, with both the Fuji native lenses and some selected Zeiss full frame lenses. It is even sharper sometimes on a cycles/pixel basis, not just cycles/picture height. This is probably because of Fuji’s smaller-than-normal microlenses. But there’s a potential downside to all that sharpness: aliasing. If the aliasing occurs with a subject with strong spatial frequency components within a factor of 10 or 20 of the sampling frequency, the result is difference frequency artifacts in the captured image."

In that post, I also said: "This is an image where the difference is striking. Possibly some focus distance differences were involved, but I could consistently get moire in this area of the striped shirt with the GFX, and hardly any with the a7RII."

And also: "Netting it out, the GFX is more susceptible to moire, as the earlier testing predicted. But the difference is certainly not night and day."

If you can document my "fully [buying] into Fuji's secret sauce", please do so. If you can provide a link to where I said it's *always* better to start with a raw file that is inherently sharper, please do that.

Otherwise, please stop paraphrasing what you think I said on this point.

It is true that I have become more negative about the GFX 50S small microlenses since getting the GFX 100, but that's not about the absolute microlens size in isolation from the pitch -- the GFX 100 microlenses are only slightly larger than the GFX 50x ones.

Jim
 

JimKasson

Well-known member
I was wondering what you guys think of the thoughts expressed in the following article. I believe Lloyd Chambers also wrote something similar.


https://blog.kasson.com/gfx-50s/on-microlens-size-in-the-gfx-100-and-gfx-50r-s/

https://diglloyd.com/blog/2019/20190818_0900-FujifilmGFX100-vs-FujifilmGFX50S.html

If as Kasson suggested, the issue with the GFX50 is due to the microlenses on the sensor, is the same argument also applicable to the X1D sensor, since they are essentially the same. I haven't shot with GFX50, but I don't think I am seeing the crunchiness or oversharpening that is being discussed. Thoughts, anyone?
Lloyd talks about the GFX 100 "oversampling", but in my tests, it doesn't do that with even the worst G-mount native lens at its worst focal length. There is still aliasing visible there, which is prima facie evidence of the lens delivering a projected image to the sensor that contains energy at spatial frequencies in excess of the Nyquist frequency. With the best G-mount lenses, at anywhere near their sharpest apertures, with demanding subjects, there is gobs of aliasing with the GFX 100.



Jim
 
Last edited:

P. Chong

Well-known member
Thanks Jim for your explanation.

In your blog article you showed both the pattern with heavy moire in the center taken by both GFX50R and GFX100. Certainly the effect on the 50R is worse. But the GFX100 is not great either, or in your opinion, this is normal.

Am I reading correctly that the GFX100 uses normal micro lenses, and not smaller ones in the GFX50RS? Or as the GFX100 being BSI, this comparison is meaningless?What about compared to the X1D? I understand pixel pitch is totally different, so does it make sense to make such a comparison?
 

JimKasson

Well-known member
Thanks Jim for your explanation.

In your blog article you showed both the pattern with heavy moire in the center taken by both GFX50R and GFX100. Certainly the effect on the 50R is worse. But the GFX100 is not great either, or in your opinion, this is normal.
Yes. It is normal practice to make the microlenses cover about the full pitch. The GFX 50S and GFX 50R were outliers. The GFX 100 returns to conventional practice. We will need cameras with about 1000 MP before we are completely free from aliasing with the lenses that we have today. Note that, with the GFX 100, you can make almost all the aliasing go away if you stop down to f/11, and it completely goes away at f/16.

Am I reading correctly that the GFX100 uses normal micro lenses, and not smaller ones in the GFX50RS?
Relative to the pixel pitch, yes. In absolute terms, the GFX 100 microlenses are slightly smaller than the GFX 50x ones.

Or as the GFX100 being BSI, this comparison is meaningless?
Whether the sensor is BSI or not is a second-order effect here. Probably best ignored.

What about compared to the X1D? I understand pixel pitch is totally different, so does it make sense to make such a comparison?
I have not tested the X1D, but from all the reports I've read it has about 100% effective fill factor, which is the standard practice these days. That makes the X1D microlenses the same size as the GFX 100 relative to the pitch, and about 1.4 times as large linearly (2x the area) in absolute terms.

Jim
 

vjbelle

Well-known member
Jim..... your analysis, which I have great confidence in, explains to some extent the softness of my GFX 100 compared to my 50S. I was very used to seeing files that were sharp and was surprised and disappointed at the softness of the files from the 100.

However, FWIW, the files from my 3100 and the files from my 4150 are similar in sharpness to what I am seeing with my 50s. Of course none of what I am doing approaches the technical approach you are taking - I'm just using my eyes. So what I was seeing from the 100 as compared to 'both' my Phase DBs and my 50s caused me to have some 'real' concern about what was going on under the hood.

If you ever want files from a 4150 let me know.

Victor
 

JimKasson

Well-known member
Jim..... your analysis, which I have great confidence in, explains to some extent the softness of my GFX 100 compared to my 50S. I was very used to seeing files that were sharp and was surprised and disappointed at the softness of the files from the 100.

However, FWIW, the files from my 3100 and the files from my 4150 are similar in sharpness to what I am seeing with my 50s. Of course none of what I am doing approaches the technical approach you are taking - I'm just using my eyes. So what I was seeing from the 100 as compared to 'both' my Phase DBs and my 50s caused me to have some 'real' concern about what was going on under the hood.

If you ever want files from a 4150 let me know.

Victor
I think the GFX 100 files are just fine wrt sharpness. I am beginning to form a theory: if the images look sharp at the pixel level, there is aliasing there.

Thanks for the offer of the 4150 files. I expect a central 33x44 mm section from those files will look the same as the GFX 100 files, but it would be interesting to test. If you're interested, shoot a Siemens Star with a good lens at its optimum aperture from far enough away that you can see aliasing about a quarter to a third of the way out from the center to the rim. The Siemens Star that Zeiss makes is fine. Then get me some raw files.

Thanks.

Jim
 

hcubell

Well-known member
What you say about the microlenses in the X1D vs the GFX 50x is true, as far as I know. I believe what you have said about my thoughts about the GFX 50S at the time of its introduction is a mischaracterization. I was always aware of the downsides of the small microlenses, and made at least one post in my blog illustrating somewhat increased false color and moire.

https://blog.kasson.com/the-last-word/fujifilm-gfx-sony-a7rii-moire/

In that post, I said: "We have seen that the GFX sensor produces much sharper images than the a7RII sensor, with both the Fuji native lenses and some selected Zeiss full frame lenses. It is even sharper sometimes on a cycles/pixel basis, not just cycles/picture height. This is probably because of FujiÂ’s smaller-than-normal microlenses. But thereÂ’s a potential downside to all that sharpness: aliasing. If the aliasing occurs with a subject with strong spatial frequency components within a factor of 10 or 20 of the sampling frequency, the result is difference frequency artifacts in the captured image."

In that post, I also said: "This is an image where the difference is striking. Possibly some focus distance differences were involved, but I could consistently get moire in this area of the striped shirt with the GFX, and hardly any with the a7RII."

And also: "Netting it out, the GFX is more susceptible to moire, as the earlier testing predicted. But the difference is certainly not night and day."

If you can document my "fully [buying] into Fuji's secret sauce", please do so. If you can provide a link to where I said it's *always* better to start with a raw file that is inherently sharper, please do that.

Otherwise, please stop paraphrasing what you think I said on this point.

It is true that I have become more negative about the GFX 50S small microlenses since getting the GFX 100, but that's not about the absolute microlens size in isolation from the pitch -- the GFX 100 microlenses are only slightly larger than the GFX 50x ones.

Jim
I am sure that you told ME personally that it was better to start with a GFX file that needed less capture sharpening because of the smaller microlenses than to try to "compensate" to achieve that level of sharpness by using sophisticated capture sharpening techniques. I distinctly remember the discussion. I am not going to spend a lot of time trying to find it, as I can't recall whether it was an email or a PM on a Forum. It doesn't really matter, because the issue of sharpness is overrated by people who obsess over it.

However, with all of the work you have done on the GFX over 2 and 1/2 years, I don't ever recall you questioning whether the smaller microlenses was problematic because it would tend to produce oversharpened images. It is strange that this issue is now coming up. It's not like the "secret sauce" is new. And, it's not like there haven't been many GFX files from that sensor for people to examine. I think you said you were surprised by the proliferation of oversharpened images from the GFX that you were seeing online, because GFX photographers were more sophisticated about sharpening than FF shooters. You pointed to the microlenses as the (or perhaps "an") explanation. I have no idea as to whether it is the microlenses, but I think your basic premise is problematic. Do you have any idea as to how many buyers of the GFX 50S and R are former APS-C and FF shooters? Do you have any idea what percentage of them shoot JPEGs and leave it to Fuji to do the sharpening for them? I assume you understand how images that are otherwise properly capture sharpened in their native file size are susceptible to looking oversharpened when they are exported into small JPEGs for the web.
 

JimKasson

Well-known member
I am sure that you told ME personally that it was better to start with a GFX file that needed less capture sharpening because of the smaller microlenses than to try to "compensate" to achieve that level of sharpness by using smaller microlenses. I distinctly remember the discussion. I am not going to spend a lot of time trying to find it, as I can't recall whether it was an email or a PM on a Forum.
Then please stop putting words in my mouth.


It doesn't really matter, because the issue of sharpness is overrated by people who obsess over it.
This is more a discussion of artifacts than of sharpness.

However, with all of the work you have done on the GFX over 2 and 1/2 years, I don't ever recall you questioning whether the smaller microlenses was problematic because it would tend to produce oversharpened images.
I am not saying that the small microlenses are a problem because they encourage oversharpened images. I'm saying that they are a problem because of aliasing. That's a different thing entirely.

The oversharpening occurs, I believe, because Lr and C1 did not adjust their default sharpening in order to take into account the small microlenses, leading to oversharpened images at the default sharpness settings. You can't blame that on the camera.

It is strange that this issue is now coming up. It's not like the "secret sauce" is new. And, it's not like there haven't been many GFX files from that sensor for people to examine.
It's not just coming up now. I have been on a minor tear about oversharpening in general for the last couple of years, and the GFX 50x images are no exception.

I think you said you were surprised by the proliferation of oversharpened images from the GFX that you were seeing online, because GFX photographers were more sophisticated about sharpening than FF shooters. You pointed to the microlenses as the (or perhaps "an") explanation. I have no idea as to whether it is the microlenses, but I think your basic premise is problematic. Do you have any idea as to how many buyers of the GFX 50S and R are former APS-C and FF shooters? Do you have any idea what percentage of them shoot JPEGs and leave it to Fuji to do the sharpening for them? I assume you understand how images that are otherwise properly capture sharpened in their native file size are susceptible to looking oversharpened when they are exported into small JPEGs for the web.
I'll never be able to test the reason for the proliferation of oversharpened files in general and GFX 50x files in particular. I can only guess. Many of the GFX 50x images that I think are oversharpened are otherwise high-quality images made by working commercial photographers.

Jim
 

hcubell

Well-known member
Then please stop putting words in my mouth.




This is more a discussion of artifacts than of sharpness.



I am not saying that the small microlenses are a problem because they encourage oversharpened images. I'm saying that they are a problem because of aliasing. That's a different thing entirely.


Jim
Oh, please. Just go back and reread your blog post. You announce that the small microlenses "made the captured images sharper---astonishingly so, in measurements that I made--- and also made the camera more prone to aliasing." You go on to say that "many, if not most," of the GFX 50S files displayed on the web are "oversharpened," and you link that with the smaller microlenses. Did I also put those words in your mouth? Does "also" not mean "also"?

Though it's not at the level of the Catholic Church changing its view on birth control, I do find this incipient reassessment of the smaller microlenses mildly amusing.
 

JimKasson

Well-known member
Oh, please. Just go back and reread your blog post. You announce that the small microlenses "made the captured images sharper---astonishingly so, in measurements that I made--- and also made the camera more prone to aliasing." You go on to say that "many, if not most," of the GFX 50S files displayed on the web are "oversharpened," and you link that with the smaller microlenses. Did I also put those words in your mouth? Does "also" not mean "also"?

Though it's not at the level of the Catholic Church changing its view on birth control, I do find this incipient reassessment of the smaller microlenses mildly amusing.
It's perfectly fine to discuss things I said in my blog posts. That's different from saying I said something which I don't remember saying, that seems at odds with my other writings at the time, and for which you have no documentation that I said it.

Jim
 

vjbelle

Well-known member
Victor,
If you can't find one let me know. I can do it this afternoon and upload the images.

Dave
Dave..... go ahead. I'll have my chart in a couple of days and shoot it with the Rody 90 and then Jim can have a couple of charts to look at.

Victor
 

dchew

Well-known member
Thanks. I wouldn't mind having two sets of images...
Here you go, Jim. IQ4150, sk60xl @ f/8. Don't hesitate to tell me if you need something different, either magnification or anything else. Might take my not-so-zoomy DSL a few minutes to get around to finishing the upload.
https://www.dropbox.com/sh/oq7j503zyd8hxgn/AADXpmhLkhrryC7180ZqK-dwa?dl=0

Dave

Note: The metadata says f/11 but that is just the lose nut behind the camera who forgot to change it to f/8.
 

dchew

Well-known member
Dave..... go ahead. I'll have my chart in a couple of days and shoot it with the Rody 90 and then Jim can have a couple of charts to look at.

Victor
Ok, I added the sk90 too. Jim, do me a favor: When you find out Victor's 90hrsw is sharper than my sk90 apo-digitar, you are welcome to tell him but please don't mention it to me.

Dave
 

JimKasson

Well-known member
Last edited:
Top