p.3 #1 · The Glass in the Path: Sensor Stacks and Adapted Lenses
Tariq Gibran wrote:
Their entire, fully native arsenal of FE FF lenses? Perhaps there is a reason why Sony is not investing heavily in the development of native glass at the moment (and will let Zeiss do that for them). The optical adapter is an interesting idea.
Looks like Sony is testing the water first with the A7 series before developing too much of fitting E-mount glass. The lens assembly is done by Sony (since they include an AF system which Zeiss is not doing).
p.3 #3 · The Glass in the Path: Sensor Stacks and Adapted Lenses
retrofocus wrote:
Looks like Sony is testing the water first with the A7 series before developing too much of fitting E-mount glass. The lens assembly is done by Sony (since they include an AF system which Zeiss is not doing).
Yes, I know about the current AF Sony Zeiss lenses. I was referring to the upcoming, direct from Zeiss, manual focus lenses that are to be launched at Photokina in the Fall.
p.3 #4 · The Glass in the Path: Sensor Stacks and Adapted Lenses
It's much easier to make the cover plates slightly convex than to make a truly curved (concave) sensor surface. Except for chromatic aberrations, they do the same thing.
The current FE lens models are what you might call "undercorrected" for their current filter thickness. They get slightly better results with thinner cover glass, something that you can easily see on an optical bench (as Roger's post correctly had as a bottom line...)
AFAIK they plan the optical model after the amount of color filtering required, without the AA filter included.
If you on the other hand do as the µFT consortium and specify the exact thickness and refractive indexes of the cover glass all µFT cameras disregarding brand should include, you've committed to a certain compensation (which is quite high...).
I once built an electronic bench for lens testing for a video surveillance client, and at first we thought about using a 16MP Pana µFT sensor, but found that it had the color correction filter laminated in the middle of the stack... :/
In the end a stripped Sony RX100 did the job even better.
You have to remember that this problem increases with aperture, as it increases with about angle squared. Addition of square increase gives an aberration total increase of about square root of that, so back to unity.
>>> An f/1.4 lens will get about twice as high impact (on PSF additional spread) from increased cover glass thickness as an f/2.8 lens.
(that ratio is not true for corners, where changes - impact - in field curvature, astigmatism and coma is much larger - especially for short exit pupil distances...)
p.3 #5 · The Glass in the Path: Sensor Stacks and Adapted Lenses
theSuede wrote:
The current FE lens models are what you might call "undercorrected" for their current filter thickness. They get slightly better results with thinner cover glass, something that you can easily see on an optical bench (as Roger's post correctly had as a bottom line...)
That is something I have already suspected by scrutinizing the peripheral areas of the photos I took with these lenses. Thank you for confirming my doubts.
If anything, that would be a good sign that Sony might reduce the filter stack thickness in future models.
p.3 #8 · The Glass in the Path: Sensor Stacks and Adapted Lenses
Will Zeiss perhaps be tweaking the optical design for the A7r on their new lenses? I've been looking at raws shot with the 25mm f2 on an A7r and the edges are mush. I'd kill for WA solution with decent corners and not the size of the Nikon 14-24mm but the A7r seems to mushify everything wide that was OK or decent on a DSLR, nevermind a rangefinder.
p.3 #9 · The Glass in the Path: Sensor Stacks and Adapted Lenses
If you're referring to the performance at f/2 maybe you're expecting too much .A 36mp. sensor is very demanding for any wideangle if you look at the corners/edges at full aperture. My guess ( I may be wrong, of course) is that the Zeiss 25 at f/2 wouldn't be stellar in the corners even if the sensor had no toppings at all.
It would be interesting if somebody did a rigorous test of a Nikkor 14-24, or a Distagon ZF 21, on a D800 vs the same lens on a Sony A7r.
p.3 #10 · The Glass in the Path: Sensor Stacks and Adapted Lenses
My experience thus far with 28 and 35mm SLR lenses on the A7r is that it's very picky even with these as regards border/ corner sharpness (seemingly more-so than on DSLR's). Perhaps some of that does have to do with the cover glass thickness + lens design compatibility. Oddly, some of the least expensive and older film lenses seem to perform the best (Pentax K 28/3.5 and a Zeiss Distagon Rollei 35/2.8 that outperforms the later C/Y 35/2.8 in the corners).
p.3 #11 · The Glass in the Path: Sensor Stacks and Adapted Lenses
artur5 wrote:
If you're referring to the performance at f/2 maybe you're expecting too much .A 36mp. sensor is very demanding for any wideangle if you look at the corners/edges at full aperture. My guess ( I may be wrong, of course) is that the Zeiss 25 at f/2 wouldn't be stellar in the corners even if the sensor had no toppings at all.
It would be interesting if somebody did a rigorous test of a Nikkor 14-24, or a Distagon ZF 21, on a D800 vs the same lens on a Sony A7r.
Not quite what you've asked for, but I did do a quick side-by side test of my 16/3.5 AI fisheye, 20/2.8D, and 45/2.8 AI-P Nikkors on my D800 and A7r (and the A7).
The A7 showed some not nice ghosting that the A7r and D800 did not.
The A7r and D800 showed just about identical CA and sharpness with these three lenses in the center and off into the far corners. The 16/3.5 was excellent off into the corners, the 20/2.8D was mush unless stopped down to f/8 (and a lot of CA that cleaned up nicely), and the 45/2.8P was the usual excellent center, poor corners until stopped down to f/5.6 or so. Seems the sensor and cover glass are either nearly the same in the A7r and D800 or have no effect with these lenses.
Now that I actually own an A7r as of yesterday (which I chose due to the ghosting issues of the A7) I will try some of my other wides on it to see how they do.
p.3 #12 · The Glass in the Path: Sensor Stacks and Adapted Lenses
artur5 wrote:
If you're referring to the performance at f/2 maybe you're expecting too much .A 36mp. sensor is very demanding for any wideangle if you look at the corners/edges at full aperture. My guess ( I may be wrong, of course) is that the Zeiss 25 at f/2 wouldn't be stellar in the corners even if the sensor had no toppings at all.
It would be interesting if somebody did a rigorous test of a Nikkor 14-24, or a Distagon ZF 21, on a D800 vs the same lens on a Sony A7r.
Performance at f8 and f11. I was rather surprised too.
p.3 #13 · The Glass in the Path: Sensor Stacks and Adapted Lenses
I think the Zeiss 2.8 version has better border/ corner performance than the faster F2 version (the MTF as best I can tell confirms this as well). I was pretty happy with the corners of the 25/2.8 when I owned it (though I never used it on the A7r).
p.3 #14 · The Glass in the Path: Sensor Stacks and Adapted Lenses
I had seen RAW's of the f2 version on the D800 and it was considerably better. The A7r is really tough on wides.
Jun 10, 2014 at 12:35 PM
Steve Spencer Online • • • • • • • Upload & Sell: On
p.3 #15 · The Glass in the Path: Sensor Stacks and Adapted Lenses
Beni wrote:
I had seen RAW's of the f2 version on the D800 and it was considerably better. The A7r is really tough on wides.
Keep in mind it might be adapter issues as well. On the D800 it would be shot without an adapter, but would of course need an adapter on the A7r. So it may be tough to tell if it is an adapter issue or a filter stack issue. Importantly for this thread it also depends on the exit pupil distance. Longer exit pupil distances should be better. SLR lenses by having big adapters will have longer exit pupil distances, but some lenses still might benefit from a longer distance.
p.3 #19 · The Glass in the Path: Sensor Stacks and Adapted Lenses
Very interesting articles and research. Does this mean that it would theoretically be possible to get better performance on the Sony A7R by replacing the sensor glass with a thinner equivalent? Has anyone tried this ?
p.3 #20 · The Glass in the Path: Sensor Stacks and Adapted Lenses
nicoimages wrote:
Very interesting articles and research. Does this mean that it would theoretically be possible to get better performance on the Sony A7R by replacing the sensor glass with a thinner equivalent? Has anyone tried this ?
Read the thread? There's someone on page 2 who has done it and says he gets better performance with M mount lenses.
Perhaps there is a reason why Sony is not investing heavily in the development of native glass at the moment (and will let Zeiss do that for them). The optical adapter is an interesting idea. 