rvh23
Offline
• • •
Upload & Sell: Off
|
 p.16 #4 · Sony FE 12-24mm f/2.8 GM Review | |
inksandpaper wrote:
A focus shift from 2 to 4 meters for 12mm focal length is 36 microns of extension on-axis. For 24mm focal length it is 146 microns. The difference is quite huge. It implies your shim could be as thick as 300(!!) microns. It's certainly not ~5-10 microns worth on-axis (you say you think your shim thickness cannot be more than 20 mics, so you also are saying it cannot be more than 10 mics on-axis). You can use my lens extensions calculator to work it out for yourself: https://www.dropbox.com/s/h63qai5me3zt3u6/Lens%20Extension%20Calculator.xlsx?dl=0
You might have seen Roger Cicala's flange distance articles where he measured thousands of cameras, and 487 Sony E-mount cameras. The variation of Sony E-mount flange distance, excluding far outliers, spans about 100 microns. Alpha 7 series cameras alone were about 60-70 microns in variation on-axis. In fact, it is remarkable that any lens, without calibration to a user's specific camera, could guarantee accurate focus distance readouts unless you got lucky. I wouldn't consider this a fault since it is easily adjusted by a service center. It is not so good to depend on this readout for focus stacking, I've never found it to be precise enough. I would recommend to use focus peaking indication. Hopefully one day Sony will program automatic focus stacking into their cameras.
https://www.lensrentals.com/blog/2020/06/the-great-flange-to-sensor-distance-article-part-ii-photo-cameras/
Maybe there was an observational error? Do you have the full resolution before and after test shots that demonstrate this severe field curvature that was induced by your shimming?
Anyway, don't take my word for it. Here were Brandon Dube's comments on field curvature vs spacing across various comments on the Lensrentals blog, from which I understood that spacing and alignment does not affect field curvature:
The mechanism of misalignment in a lens determines how its MTF departs from the average or nominal value. A spacing error will cause a relatively uniform error across the field of view because it changes only spherical aberration and axial color, aberrations that are constant over the field. There is a balance between aberrations so it is not precisely uniform over the field, but for small changes the aberrations are reasonably orthogonal.
Also:
Surfaces are not equal for different purposes. If you want to image on a flat surface, you absolutely must have both positive and negative powered elements for the Petzval sum to be anywhere near zero. If you curve the image, you don’t need negative lenses anymore because the image plane can curve as the positive lenses want it to. This means you can delete the negative lenses, which probably injected aberrations into your design anyway.
Or, in the more general case, it gives you a knob that is essentially a 1:1 control over field curvature. Where you may have had 5, maybe 10 elements all fighting each other to get near zero field curvature otherwise.
This makes the curved image surface as valuable as several other elements, not just parameters, in the design.
And:
The reason the image is curved is first and foremost because of "petzval curvature," which has a design type named after it, since that is the only primary aberration the design does not correct for. Petzval curvature comes simply from the positive focal length of the elements and is, as a result, not very sensitive to alignment.
Next from petzval you have "astigmatism," which affects the tangential ray fan three times more than the sagittal because of geometry. It's obvious, then, that the design should immediately manifest any perturbation of astigmatism three times more strongly in T than S. Then you have the designer's efforts, where they introduce "tangential astigmatism" or "sagittal astigmatism" to flatten the field. They need three times as much of this pill to bring T on top of S, and generally the more aberration you put in, the more sensitivity you get, too.
And:
Astigmatism and field curvature are invariant with aperture. They are reduced in apparent magnitude when the aperture is closed, because the depth of field is increased. If a lens has focus shift from spherical aberration, the plane selected in the MTFvFvF plots may move to one that is more astigmatic (e.g. if the T and S fields move away at the edge, which is what astigmatism is).
The separation of the T and S lines in an MTF vs Field plot is not directly indicative of astigmatism; there can be several causes for that (astigmatism being one of them.)
...Show more →
Thanks very much for all this additional information and taking the time.
I don't have a good explanation for the significant distance readout disruption, but there is no way the shimming material is a hundred, let alone several hundred microns thick. The only thing I can think of is that when I punched a hole through it for the screws to pass through it left a burr that acted to increase the effective thickness. But I would expect a shim that is several hundred microns thick and placed in one corner to be way too much adjustment given the lens was not severely asymmetric in the first place.
In relation to the induced FC, no I unfortunately don't have images I can show. It was just an anecdotal observation through the EVF, because by that stage I had already decided I wasn't going to be keeping the lens. But it seemed quite obvious even at 24mm where the lens unshimmed is essentially flat field.
When you say the spacing and allignment shouldn't affect FC, I don't at all doubt your expertise, and even less so the folks at lensrenatals. But my recollection of my adapter shimming days, and also discussion here on FM, is that FC often was affected. But it's possible I'm misremembering the details.
|
