p.134 #2 · Official: Sony A7 and A7R Fullframe Mirrorless
uhoh7 wrote:
I care
I think you can open these in tabs to see them easily. Interested to know what you think. The crops are 15mm @ f/8 on A7 vs. 10mm @ f/5.6 on 5N, tried to hit the sweet spot on both sensors for sharpness. No processing on the files, just the RAW deal. I know there's a big white dust bunny in the A7 shot.
p.134 #3 · Official: Sony A7 and A7R Fullframe Mirrorless
mco_970 wrote:
I think you can open these in tabs to see them easily. Interested to know what you think. The crops are 15mm @ f/8 on A7 vs. 10mm @ f/5.6 on 5N, tried to hit the sweet spot on both sensors for sharpness. No processing on the files, just the RAW deal. I know there's a big white dust bunny in the A7 shot.
The A7 seems OK in this shot--within reason--but not as good as the APS-C, however what would really be interesting is to do the same thing, when you have a chance, outdoors, at infinity with details in the distance. multiple stops would be cool, but at least one f/8 with careful focus.
That's when smearing is most likely and easiest to see. The rear element, as sebboh taught me, is then closest to the sensor.
Distortion is also much more evident in the A7: one wonders how the lens will do in crop mode.
p.134 #4 · Official: Sony A7 and A7R Fullframe Mirrorless
uhoh7 wrote:
The A7 seems fine in this shot--within reason--however what would really be interesting is to do the same thing, when you have a chance, outdoors, at infinity with details in the distance.
That's when smearing is most likely and easiest to see. The rear element, as sebboh taught me, is them closest to the sensor.
Any idea how far out 'infinity' is for this lens? Problem right now is atmospheric distortion. You know about the super-sunny cold waviness...
p.134 #6 · Official: Sony A7 and A7R Fullframe Mirrorless
Thanks. I double checked on the Samyang 15 and it shows infinity at about 4 yards, so 10-40 should be OK.
Yeah, I can probably grab a set either later today or tomorrow. I tend to use UWA's with a close in subject, or indoors, so that is definitely of interest to me personally.
p.134 #8 · Official: Sony A7 and A7R Fullframe Mirrorless
I know! Why do you think I was gear-lusting for one? The NEX thread is a bad, bad place to hang out. There have been quite a few great 10-18 pictures lately.
On the A7, I did go through and process both pics. Even after correcting A7's fairly strong vignette, the corners look pretty decent on it. At least on my copy of the gear, I'd give the nod to A7 + 10-18 over the 5N.
The lens itself might just be optically stronger at 15mm, though, and that could easily account for all differences. So it all means nothing anyway.
p.134 #9 · Official: Sony A7 and A7R Fullframe Mirrorless
This is weird the variability of the 10-18, tested by different users. There was a test done by someone in China or Korea (before any of us got our cameras) that showed the 10-18 severely cutoff in FF mode on the A7(r), but in my test it was not nearly as bad unless you move towards 10mm. Your tests shows the 10mm not shading nearly as bad as what I was seeing. In my 10mm shot, it was like looking through a cylindrical tube. Are you sure you didn't have auto crop mode enabled? I did all my tests wide open. I wonder if that could be the explanation? I'll try to find some time to take my tripod out today and shoot at all aperture. In auto crop mode, this lens is fine on the A7R.
p.134 #10 · Official: Sony A7 and A7R Fullframe Mirrorless
theSuede wrote:
Actually it does, at least with / at F1.8 or larger apertures. In the center, at F5.6 - no real difference.
Optical flats in the ray path change the spherical aberration behavior. It's the same type of aberration change as corner blur - the main problem is angle, or actually angle difference. Like this; article originally meant for microscopy, but it's the same principle. http://www.nikon.com/products/instruments/resources/tech/info/microscope_tech/coverglass/index_02.htm
The thicker the glass, the more it changes the SA and comatization effect.
And additionally, if you introduce more surfaces you also get more reflection/scatter/internal scatter. I've seen some recent cameras that actually uses four separate layers above the sensor surface proper (one AA sensor plate cover, one AA/phase plate lamination, one hot-mirror/absorption color filter, one dustoff filter (static charge+vibration). Total madness according to me, but there must be some reason. Probably price, or prioritizing the efficiency of the dust-off function over the optical performance.
Another thing with angle is that reflection/refraction for different polarizations diverge more and more as you increase angle. And reflection increases in total effect. When you have optical flats close to each other like this you see that as a lowering of global/LF contrast. If you take two surfaces with a 4% reflection, you get 0.04*0.04 = 0.16% = -9Ev contamination, increase that reflection to 10% at 35º angle and you get 0.1*0.1 = 1% = -6.6Ev contamination. ANY bright patch close to darker areas will then contribute almost as much veiling flare to the dark detail as the detail itself - effectively halving the MTF....Show more →
Would you say there are any significant disadvantages if one swops to a thinner UV/IR filter plus getting rid of AA + antidust? I've heard stories about M9 sensor filters that crack, which perhaps is due to it being more sensitive to quick (local) changes in temperature because of lesser thickness.
It seems a bit odd that Sony chose a thick filter stack for this kind of camera, since it obviously limits the design possibilities for smallish lenses. On the other hand, perhaps the sensor guys never got that feedback from the optics department. Or no one even thought about it.
p.134 #11 · Official: Sony A7 and A7R Fullframe Mirrorless
sflxn wrote:
This is weird the variability of the 10-18, tested by different users. There was a test done by someone in China or Korea (before any of us got our cameras) that showed the 10-18 severely cutoff in FF mode on the A7(r), but in my test it was not nearly as bad unless you move towards 10mm. Your tests shows the 10mm not shading nearly as bad as what I was seeing. In my 10mm shot, it was like looking through a cylindrical tube. Are you sure you didn't have auto crop mode enabled? I did all my tests wide open. I wonder if that could be the explanation? I'll try to find some time to take my tripod out today and shoot at all aperture. In auto crop mode, this lens is fine on the A7R....Show more →
That's because Michelle's 10mm shots were from the 5N ... ie comparing similar FOV to 15mm on the FF A7
" 10mm @ f/5.6 on 5N"
p.134 #12 · Official: Sony A7 and A7R Fullframe Mirrorless
Makten wrote:
Would you say there are any significant disadvantages if one swops to a thinner UV/IR filter plus getting rid of AA + antidust? I've heard stories about M9 sensor filters that crack, which perhaps is due to it being more sensitive to quick (local) changes in temperature because of lesser thickness.
It seems a bit odd that Sony chose a thick filter stack for this kind of camera, since it obviously limits the design possibilities for smallish lenses. On the other hand, perhaps the sensor guys never got that feedback from the optics department. Or no one even thought about it....Show more →
I might get back to you regarding practical examples of what happens in reality with some samples after X-mas, but at the moment we only have one of each. I'm not going to tear up the only samples we have.
A thinner color filter plate is no problem, as long as you match the filtering properties. You could sacrifice some near-IR and n-UV and chose slightly more narrow passband dielectrics surfaces.
The three large manufacturers have very little concerns for WA large aperture lenses - the main money producer is the high-end F2.8 zoom offerings.
That Leica's filters had a problem is a material quality problem, and some bad mechanical constructions combined. I'd say that it was a low-probability combination of parameters that caused the problem, just bad luck for them.
For the A7 (not -R) I think I'll replace the dust-off plate with a thin UV-filter plate, and leave the absorption filter as it is. Basically remove the three birefringents then.
No-one seen a tear-down of the A7 yet? that surprises me, but I haven't really looked around for a while, been busy with other stuff.
p.134 #13 · Official: Sony A7 and A7R Fullframe Mirrorless
theSuede wrote:
Actually it does, at least with / at F1.8 or larger apertures. In the center, at F5.6 - no real difference.
Optical flats in the ray path change the spherical aberration behavior. It's the same type of aberration change as corner blur - the main problem is angle, or actually angle difference. Like this; article originally meant for microscopy, but it's the same principle. http://www.nikon.com/products/instruments/resources/tech/info/microscope_tech/coverglass/index_02.htm
The thicker the glass, the more it changes the SA and comatization effect.
And additionally, if you introduce more surfaces you also get more reflection/scatter/internal scatter. I've seen some recent cameras that actually uses four separate layers above the sensor surface proper (one AA sensor plate cover, one AA/phase plate lamination, one hot-mirror/absorption color filter, one dustoff filter (static charge+vibration). Total madness according to me, but there must be some reason. Probably price, or prioritizing the efficiency of the dust-off function over the optical performance.
Another thing with angle is that reflection/refraction for different polarizations diverge more and more as you increase angle. And reflection increases in total effect. When you have optical flats close to each other like this you see that as a lowering of global/LF contrast. If you take two surfaces with a 4% reflection, you get 0.04*0.04 = 0.16% = -9Ev contamination, increase that reflection to 10% at 35º angle and you get 0.1*0.1 = 1% = -6.6Ev contamination. ANY bright patch close to darker areas will then contribute almost as much veiling flare to the dark detail as the detail itself - effectively halving the MTF....Show more →
This makes perfect sense, but if you have time I have a question. Off-axis effects on adapted lenses seem to be more significant than with the native lenses. Is that because the Sony lens is designed with the cover glass thickness taken into account, because of in-camera raw modification as some have claimed, or because of some other difference I've overlooked?
p.134 #16 · Official: Sony A7 and A7R Fullframe Mirrorless
It seems very likely that Sony are doing plenty of signal processing (their term) with the new cameras. No one knows much about it but they have talked about careful anti-diffraction measures, noted here by Roger Slavens:
'BIONZ X also brings two new features. Diffraction-reducing technology combats the effects of diffraction limiting, improving detail at smaller apertures. Interestingly, the effect applied is both lens and aperture specific, and said to work even with Alpha-mount lenses shot through an adapter. It also has what Sony bills as "Detail reproduction technology", which tries to draw out finer details without creating halos in the process.'
and
'Sony seems to have finally gotten a handle on their in-camera JPEG processing, to the extent that they now lead the field, in our estimation. Thanks to their new BIONZ X processor and the more extensive image processing it enables, they've managed to suppress noise while maintaining great subject detail. Equally significantly, they've really refined their sharpening algorithms.'
I still read how Sony jpegs are poor from these cameras however.
Perhaps this is being done to some extent in RAW files as well as jpegs?
Some users are reporting very good results at ISO 50 as well...
p.134 #17 · Official: Sony A7 and A7R Fullframe Mirrorless
philip_pj wrote:
'Sony seems to have finally gotten a handle on their in-camera JPEG processing, to the extent that they now lead the field, in our estimation. Thanks to their new BIONZ X processor and the more extensive image processing it enables, they've managed to suppress noise while maintaining great subject detail. Equally significantly, they've really refined their sharpening algorithms.'
I still read how Sony jpegs are poor from these cameras however.
Perhaps this is being done to some extent in RAW files as well as jpegs?
.
The problem is there is a lot of hype and people that grift a little money from every camera sold and then there are those of us that pay for less than what we should get.
I will give up fast FPS, I will give up wifi - I will give up scene modes - if I can just capture a scene other than an Imatest chart, a b&w photo or a detailed and complex composition aka trees with a wide angle and get a decent night shot, teleshot - even tonal transitions.
The camera is capable - the compression is getting in the way.
p.134 #18 · Official: Sony A7 and A7R Fullframe Mirrorless
RCicala wrote:
This makes perfect sense, but if you have time I have a question. Off-axis effects on adapted lenses seem to be more significant than with the native lenses. Is that because the Sony lens is designed with the cover glass thickness taken into account, because of in-camera raw modification as some have claimed, or because of some other difference I've overlooked?
There's more to the "designed for digital" moniker than what many skeptics would like you to believe.
I know for a fact that they optimize for the change in field curvature, but regarding astigmatism I don't know. It's a complex system. And the effects are varying with exit pupil distance and aperture - making the effects hard to predict.
Often you design systems to have "working windows" of operation, specifying a set of interdependent parameters like for instance [exit pupil distance] = [D+D/N] * [+/-15%] where D is a set distance and N is the maximum aperture. Knowing this gives you a window of operation of field curvature, a field curvature that the the sensor will receive as "flat".
Given the mechanical layout of the mount "well", you can also include limiters for mechanical shading and so on. But the basic formula is set to enable a good interaction between the sensor system and the lenses.
On "longer" lenses, not that I personally would call 85mm "long", the difference between optimized and non-optimized lenses get smaller. I get very little difference except at F1.2-1.8 between an on-camera test for a 85mm+ lens and a bench test of the same lens. The difference with short symmetricals on the other hand is often huge (but field curvature is of course also dependent on FL, shorter FL's need more fc correction).
The problem is larger with AA-filtered cameras, since the introduce yet another effect - aberration is modulated by angle + polarization in stead of just angle.
Normally you don't get polarization dependent blur in sensors without AA filters, at least until you get very shallow angles, only pol.dependent reflection/scatter.
None of this includes any sensor surface proper induced effects, like microlens effects, cell bleed (color vignette) and so on.
the edges are just not close to as sharp as the center.
it could be the lens, I suppose, but I'm more inclined to believe it's Edward's sensor toppings, much as I hate to admit it
It's bugged me enough on a bunch of lenses that I'm trading my r in for the plain A7.
Of course Sony may have the native solution down to eventually turn the r into a landscape machine, but I'm looking for something that will play nice with a lot of glass.
p.134 #20 · Official: Sony A7 and A7R Fullframe Mirrorless
The SEL1018 samples do not look good at all but to be fair, this lens was not optimized for full frame.
My experience so far is with adapted TS-E Canon lenses like the 17mm and 24mm. What I'm getting at the corners is an effect that resembles 'diffraction'.
In that case, the explanation of glass thickness in front of the sensor makes a lot of sense.
The NEX thread is a bad, bad place to hang out. There have been quite a few great 10-18 pictures lately.
