Luminous Landscape Forum
Equipment & Techniques => Medium Format / Film / Digital Backs – and Large Sensor Photography => Topic started by: madlantern on July 05, 2017, 08:14:43 am
-
I'm an X1D user, and I'm really interested in how its lenses compare with those of the GFX50. Are there directly comparisons of X1D vs GFX lenses? If no one has made them yet, is a Fujifilm owner interested in doing some comparisons with me? Not sure what'd be a good reference to test both lenses from though...
-
Location ?
Gesendet von iPhone mit Tapatalk
-
Seattle... I don't think we're going to actually be able to shoot the same scene. Perhaps just some test patterns in a controlled environment?
-
Sorry not interested in patterns. ;) I tested all my GFX lenses on my usual location and they perform extremely well.
Perhaps in the summer if I have nothing to do I can get a X1D and shoot the same scene.
Gesendet von iPhone mit Tapatalk
-
I'm an X1D user, and I'm really interested in how its lenses compare with those of the GFX50. Are there directly comparisons of X1D vs GFX lenses? If no one has made them yet, is a Fujifilm owner interested in doing some comparisons with me? Not sure what'd be a good reference to test both lenses from though...
I know of no way to compare the lenses except on an optical bench, since the two sets of lenses won't work on the other manufacturer's cameras. You could do test shots with similar focal lengths (are there any primes that are identical) or the native bodies, but the microlens design of the two cameras is different, and you'd be looking at that, too. In my testing, a significant part of the GFX lens sharpness is the microlens system over the sensor. It's an open question whether this level of sharpness is a good thing or not. I've seen complaints for GFX users about aliasing, but gosh, the sharpness is incredible.
Jim
-
I tested the X1D 30 mm and 45mm prime lenses against the comparable focal lengths of the GFX 32-64 Zoom at various apertures. Ordinarily, this would not be a fair fight, comparing a zoom against primes. However, the Fuji GFX zoom is a very impressive lens. Yes, the X1D prime lenses were somewhat "better" in terms of sharpness, but this was pixel peeping. Even in a big print, you may be hard pressed to see the differences. I decided that the optical quality of the lenses for the two systems was NOT a reason for me to choose one or the other system. There were other differences that I found much more significant.
One other difficulty with comparing lenses is that the GFX files with the Fuji lenses are quite sharp even when you thought you had turned off sharpening completely in LR. I have no idea if there is sharpening baked into the files or it's the microlenses on the Fuji sensor. Whatever it is, it makes it hard to do an apples to apples comparison. What I did was turn the Sharpening in LR down to 0% on both sets of files and use Focus Magic to taste for pre-sharpening.
I don't understand the physics behind the Fuji microlenses, but if they really do produce sharper files, you have to ask why no one has done the same sort of thing in the past. Did Fuji really figure out something that nobody else knew? Or, did Fuji accept certain trade offs in moire and aliasing that would appear in raw files but that Fuji could "correct" in its JPEGs? Fuji does seem to be shockingly obsessed about its JPEGs for the GFX given it is styled as a "professional" camera. Fuji still thinks it is in the film business, and considers JPEGs to be the "film."
-
I tested the X1D 30 mm and 45mm prime lenses against the comparable focal lengths of the GFX 32-64 Zoom at various apertures. Ordinarily, this would not be a fair fight, comparing a zoom against primes. However, the Fuji GFX zoom is a very impressive lens. Yes, the X1D prime lenses were somewhat "better" in terms of sharpness, but this was pixel peeping. Even in a big print, you may be hard pressed to see the differences. I decided that the optical quality of the lenses for the two systems were NOT a reason for me to choose one or the other system. There were other differences that I found much more significant.
One other difficulty with comparing lenses is that the GFX files with the Fuji lenses are quite sharp even when you thought you had turned off sharpening completely in LR. I have no idea if there is sharpening baked into the files or it's the microlenses on the Fuji sensor. Whatever it is, it makes it hard to do an apples to apples comparison. What I did was turn the Sharpening in LR down to 0% on both sets of files and use Focus Magic to taste for pre-sharpening.
I don't understand the physics behind the Fuji microlenses, but if they really do produce sharper files, you have to ask why no one has done the same sort of thing in the past. Did Fuji really figure out something that nobody else knew? Or, did Fuji accept certain trade offs in moire and aliasing that would appear in raw files but that Fuji could "correct" in its JPEGs? Fuji does seem to be shockingly obsessed about its JPEGs for the GFX given it is styled as a "professional" camera. Fuji still thinks it is in the film business, and considers JPEGs to be the "film."
There is no magic here. Making the microlenses smaller means sampling a smaller area per pixel. That makes the files sharper. If also makes them more prone to aliasing.
I'm beginning to get a handle on what's going on, thanks to some modeling proposed and performed by Jack Hogan. He took one of my raw files with the 110/2 at f/2.8 and the razor blade target (razor blades are sharp in more ways than one!) and fit the resulting MTF curve with a set of parameters to his lens and sensor MTF program. He ended up with linear fill-factor of a bit over 0.8. That's a preliminary number at this point, but it's interestingly low. Expressed in conventional fill factor terms, that would be about 65%. Most cameras have been getting close to 100% recently.
Jack is still crunching numbers (and dealing with real life, as opposed to Internet life) so it may be a while before we have a report ready to go, but we are (mostly he is, as all I'm doing is feeding him raw files) making progress.
Jim
-
There is no magic here. Making the microlenses smaller means sampling a smaller area per pixel. That makes the files sharper. If also makes them more prone to aliasing.
Jim
It sounds like there are trade offs here in the use of microlenses that are optimized for sharpness at the expense of issues with aliasing and/or moire. It's no secret, right? Yet, Fuji is apparently the only manufacturer to go down this path with the Sony 50MP sensor. Are they just smarter than everyone else? The other question is whether the apparent "loss" of sharpness with Phase, Hassy and Pentax files using the same sensor but bigger microlenses is compensated for by effective sharpening techniques on the raw files. What's easier to compensate for, "loss" of sharpness or aliasing?
-
It sounds like there are trade offs here in the use of microlenses that are optimized for sharpness at the expense of issues with aliasing and/or moire. It's no secret, right? Yet, Fuji is apparently the only manufacturer to go down this path with the Sony 50MP sensor. Are they just smarter than everyone else? The other question is whether the apparent "loss" of sharpness with Phase, Hassy and Pentax files using the same sensor but bigger microlenses is compensated for by effective sharpening techniques on the raw files. What's easier to compensate for, "loss" of sharpness or aliasing?
Some people will love the trade-off that Fuji made. Others (I'm thinking architectural photographers) may hate it. I did some textile tests against the a7RII and the GFX didn't seem markedly worse:
http://blog.kasson.com/the-last-word/fujifilm-gfx-sony-a7rii-moire/
I myself think that the 33x44 sensor should have AA filters, but nobody listened to me, and no 33x44 sensor camera has them, I think.
Jim
-
What's easier to compensate for, "loss" of sharpness or aliasing?
That's easy. Loss of sharpness. Deconvolution mostly just costs you SNR. Once it's aliased, you can't un-alias it.
However, I'm not sure that what is desired is no aliasing. I think what people want is no visible artifacts. That's more complicated.
Jim
-
Sorry not interested in patterns. ;) I tested all my GFX lenses on my usual location and they perform extremely well.
Perhaps in the summer if I have nothing to do I can get a X1D and shoot the same scene.
Gesendet von iPhone mit Tapatalk
Where are you physically? I was under the impression that you were in Europe, thus making shooting the same scene rather difficult this summer
-
there are trade offs here in the use of microlenses that are optimized for sharpness at the expense of issues with aliasing and/or moire.
Not only moire issues, photon collection becomes lower too. In the end that means SNR is also sacrificed for sharpness.
Regards
Enviado desde mi ALE-L21 mediante Tapatalk
-
Not only moire issues, photon collection becomes lower too. In the end that means SNR is also sacrificed for sharpness.
Regards
Enviado desde mi ALE-L21 mediante Tapatalk
True enough. About a half stop less light, in the case of the GFX.
Jim
-
There is no magic here. Making the microlenses smaller means sampling a smaller area per pixel. That makes the files sharper. If also makes them more prone to aliasing.
I'm beginning to get a handle on what's going on, thanks to some modeling proposed and performed by Jack Hogan. He took one of my raw files with the 110/2 at f/2.8 and the razor blade target (razor blades are sharp in more ways than one!) and fit the resulting MTF curve with a set of parameters to his lens and sensor MTF program. He ended up with linear fill-factor of a bit over 0.8. That's a preliminary number at this point, but it's interestingly low. Expressed in conventional fill factor terms, that would be about 65%. Most cameras have been getting close to 100% recently.
Jack is still crunching numbers (and dealing with real life, as opposed to Internet life) so it may be a while before we have a report ready to go, but we are (mostly he is, as all I'm doing is feeding him raw files) making progress.
Jim
Since it's impossible to isolate lens performance from other aspects of the body, how do you think the X1D+native lenses would compare to the GFX+native lenses at roughly the same focal lengths and apertures?
-
Since it's impossible to isolate lens performance from other aspects of the body, how do you think the X1D+native lenses would compare to the GFX+native lenses at roughly the same focal lengths and apertures?
I can't help you there because i don't have access to an X1D.
Sorry.
Jim
-
Since it's impossible to isolate lens performance from other aspects of the body, how do you think the X1D+native lenses would compare to the GFX+native lenses at roughly the same focal lengths and apertures?
I used to own both and can offer some qualitative opinions. Both lens systems are impressive in terms of contrast at various levels. I found the XCD 30 to be a little bit sharper than the 32-64@32mm, a difference that was still there (though hard to discern) even when stopped down. On the long end the HCD 90 and Fuji 120 were both on par but the XCD had more LoCA and less pleasing bokeh in some circumstances. I found myself more often drawn to the X1D images and found them just a little more pleasing but I suspect that has more to do with how Lightroom renders files from each camera.
-
I can't help you there because i don't have access to an X1D.
Sorry.
Jim
Is there anything I'd be able to do remotely to help?
-
Is there anything I'd be able to do remotely to help?
Send me a camera and some lenses?
If you look at my GFX lens tests on my blog, you can see the kind of things that I do to test a lens.
Question: is is easy to get these from LensRentals? I'm kinda up to my eyeballs in GFX testing plus the new Sony 12-24 and the batis 18, but when I work through that I could rent an X1D is there's' enough interest. There's also some interest in the P1 IQ3 100 MP, but I personally have no interest in that camera because of size, weight, and expected plummeting asset value (I've got no interest in the X1D, either, but it seems like a lot of people do).
jim
-
Send me a camera and some lenses?
If you look at my GFX lens tests on my blog, you can see the kind of things that I do to test a lens.
Question: is is easy to get these from LensRentals? I'm kinda up to my eyeballs in GFX testing plus the new Sony 12-24 and the batis 18, but when I work through that I could rent an X1D is there's' enough interest. There's also some interest in the P1 IQ3 100 MP, but I personally have no interest in that camera because of size, weight, and expected plummeting asset value (I've got no interest in the X1D, either, but it seems like a lot of people do).
jim
Id be very interested in that comparison. It'd be fantastic if you could do it