I've been shooting with an A7RII (previously A7R) & 17mm TSE for a while now and never experienced any of these problems. I just acquired a 24mm TSE and shot with it today only slight shifted. Files look perfectly normal to me. Will test some more soon.
Ken Rockwell did a test demonstrating issues with the 11-24mm on the A7RII, ie. showing that it was very soft away from the center. Tony Northrup later repeated this experiment. I've been considering getting the 11-24, but I've been concerned about this issue. I wonder if it's something that potentially affects any adapted wide angle lens on Sony.
I measured the register in the a7r2 and the a7r and found a difference of .290mm over an average of 10 measurements on each camera.
I used a digital caliper and measured against a stainless steel rule placed across the flange in approximately the center of the sensor. I have no bench so the variations are probably due to differences in angle front to back and less from side to side if at all.
My measurements were as follows:
a7r2
15.76
15.84
15.78
15.80
15.78
15.76
15.81
15.74
15.77
15.73
--------
157.77/10 = 15.78 rounded up median 15.77
For the A7r
15.47
15.41
15.48
15.44
15.55
15.56
15.41
15.56
15.48
15.58
-------
154.94/10 = 15.49 rounded down median= 15.48
Trying to measure around the a7r2 sensor resulted in much more widely varying values even though stabilization was turned off and the camera was turned on to "stabilize" the sensor. I really wonder how well the sensor is aligned when stabilization is turned off.
My personal opinion is that there is a difference in register. I'd like to know where the a6300 is with regards to these two values.
I regret not having followed this up when I first found it but now that others are finding problems quite possibly originating from this same difference it might be time to do carefully controlled tests. Perhaps Roger Cicala could be enticed into having a closer look. I suspect that with his optical bench checking register on the two cameras would be both simpler and a lot more accurate.
engardeknave wrote:
I've been shooting with an A7RII (previously A7R) & 17mm TSE for a while now and never experienced any of these problems. I just acquired a 24mm TSE and shot with it today only slight shifted. Files look perfectly normal to me. Will test some more soon.
Ken Rockwell did a test demonstrating issues with the 11-24mm on the A7RII, ie. showing that it was very soft away from the center. Tony Northrup later repeated this experiment. I've been considering getting the 11-24, but I've been concerned about this issue. I wonder if it's something that potentially affects any adapted wide angle lens on Sony....Show more →
I suspect the potential for problems is there for lenses with floating elements and for lenses with hard stops at infinity. I don't think lenses that focus past infinity and then are retracted back to infinity will have a problem unless infinity falls short of the focal length difference.
Which canon, minolta af or sony lenses have floating elements?
That is what I suspect myself as well. (IBIS only causes problems)
Edit: Well well...check out Tony Northrups video about Sigma adapter. He tested it against MB4 with canon 16-35/4 and results are...drumroll...wild. Sigma screws up image corners. This just suggests adapters are really the culprit. And Sigma tolerances are not any better it seems.
Schlotkins wrote:
OK, here is my 24 TSE at f8. shifted extreme corner, 100% crop, no sharping. First is no shift, then +6, +10, and +12. +12 is not great but even +10 is pretty good.
EDIT: I didn't change exposure and color balance was auto so that probably explains a lot of the color shift. I also don't have a 5DsR to compare to or I would. I am probably going to pick one up soon for my trip to Iceland in June.
Thanks for that! I'll post mine up too tomorrow. Before I' did the shimming my extreme edges were far worse than yours, f11 or f13 should give you even better results.
TakenWild wrote:
Thanks for that! I'll post mine up too tomorrow. Before I' did the shimming my extreme edges were far worse than yours, f11 or f13 should give you even better results.
My pleasure. I'm going to do my 17TSE today at f11 to see what happens there and will post up and compare to my 16-35 f4 IS
First I think the a7rII will show worse corners obviously because it's more MPX. For general lenses, I think things are OK. You can see here my canon - even adapted - was generally better than the sony 16-35 in the corner:
But I think the bottom line is when you get into these "extreme" lens situations - be it these wide TSEs shifted a lot or probably the 11-14 portion of the 11-24 - you are going to have problems. No matter what the adapters do, they are not going to replicate how the light hits the sensor on a canon mount versus a sony mount.
But I don't know if the A7rII is worse than the A7r with adapted lenses. Really my only issue with the A7r is the shutter shock with lenses longer than 70. If that wasn't there, I wouldn't think about upgrading really.
What I couldn't understand was how my 24 TSE II was terrible unshifted and my canon 16-35 f/4 and rokinon 24 were fine from the same setup with the same adaptor. All I changed was the lenses and took the same test shots with all 3 lenses. The rok was the sharpest, next was the 16-35 (still sharp just not as sharp as the rok) and then the tse 24 ii was a complete mess and looked pixelated at 100%. If it was an adaptor issue surely they would all be bad? (Bear in mind I was unshifted)
yea that is weird. I think we can all agree my 24TSE is sharp in the corner. I'm gonna do my 17TSE as well as my 16-35 @ 17 and 24 and my 24-70 @ 24 as a comparison and will post later.
I would love to try my adapter on an A7rII. I am debating if I should add a 5dsR or a7RII to my kit. My guess is usually I wouldn't notice shutter shock on my landscape shots but this guys shows shutter shock with a 35mm lens on a tripod:
I don't think it can be the sensor glass: Canon and Sony have pretty similar sensor glass thicknesses (and Canon's vary a bit from model to model according to Lensrentals - at lest Sony is consistent)
I don't see how it can be the adapter if it's in spec: someone talked about "replicating how the light hits the sensor in a Canon". But it's a tube: you replicate how the light hits the sensor by holding the lens away from the sensor! And the baffles and so on should only cause vignetting if they get in the way (I suppose if the baffle are a *tiny* bit too small, light might be refracted around the edges of them. Easy to compare by using an empty tube adapter, and you might get more flare but less smearing if that's the culprit. I doubt it.
Registration distance should not matter if the lens can cover the full range. But as others have noted, if there's a floating element then exact registration will matter more.
Parallellness of the mounts: cumulative between camera - adapter and adapter - lens - can certainly be a problem with wides. I would have thought though that if there's a problem that would be an asymmetrical problem.
It seems some people are just lucky with their combos, schlotkins for example, some people are finding zero details in the shifted corners (me) and others are having problems with the combo even unshifted.
I've got my shifted corners to look as good (maybe a tiny bit better) than Schlotkins on my A7R2 and 24mm TSE by adding 0.05mm shims. I've now got a Metabones 4 adapter on the way and I'll compare that as, before and after shims were added.
The shifted extreme corners aren't ever tack sharp, but I only ever really wanted usable, not a smeared mess I was getting.
There have been reports of some not being able to reach infinity so the range being adequate is not a given but a variable and leaves a change in register, already demonstrated by my measurements, in play.
I'm curious as to the serial numbers of the cameras. I wonder if there was a manufacturing fault initially that was fixed. If so then cameras that seem to be ok should have a higher serial number than those that don't.
I think the easiest and probably best solution is to remove the camera flange and mill the bottom surface to insure infinity focus. I've been scanning threads about the sony a7r flange replacements from fotodiox and shoten. Many of the problems encountered sound very much like those being reported here. In at least one post I saw a report that fotodiox's last signature version in bronze did appear to work well for at least one person. It's available on amazon and should be plug and play but I'm not wholly convinced it is thin enough.
The further back in the path between the sensor and the image the fewer changes that will have to be made on intermediate parts; that's why milling the camera flange seems to me to be the best solution. If it turns out that newer cameras don't have the problem then just getting a new flange from sony may be a viable option as well.
DavidBM wrote:
It's a bit of a mystery.
I don't think it can be the sensor glass: Canon and Sony have pretty similar sensor glass thicknesses (and Canon's vary a bit from model to model according to Lensrentals - at lest Sony is consistent)
I don't see how it can be the adapter if it's in spec: someone talked about "replicating how the light hits the sensor in a Canon". But it's a tube: you replicate how the light hits the sensor by holding the lens away from the sensor! And the baffles and so on should only cause vignetting if they get in the way (I suppose if the baffle are a *tiny* bit too small, light might be refracted around the edges of them. Easy to compare by using an empty tube adapter, and you might get more flare but less smearing if that's the culprit. I doubt it.
Registration distance should not matter if the lens can cover the full range. But as others have noted, if there's a floating element then exact registration will matter more.
Parallellness of the mounts: cumulative between camera - adapter and adapter - lens - can certainly be a problem with wides. I would have thought though that if there's a problem that would be an asymmetrical problem. ...Show more →
Here are a few more examples. I compared my TSE to the 24-70 f4 IS and the 16-35 IS I have. The crops are in that order and the sharpness seems to be in that order, which is what I would expect. All look pretty darn good though.
And here is my 17TSE. The first crop is from my 16-35 f4 IS @ 17mm. I must have misfocused it a bit as usually it's pretty close to my 17TSE, although I do that test at f8. These are all at f11, extreme corner, 100% crop no sharpening. Again, I think this looks pretty darn good. Order is 16-35, 17 no shift, +6, +10, +12.
EDIT: This is probably not the best test since I'm tilting the camera down a ton to get the tower in the upper left corner. This cases the element to be bigger which causes the shifts to look worse by default. But in any case, you can see the shift doesn't really kill the IQ until 12mm.
Fred's experimenting would suggest that Keith should have used at least f/10-11 instead of f/8.
Not sure if it would make a difference in this shot, granted, but Fred has noted a higher f-stop is needed with significant shifting.
dgdg wrote:
Fred's experimenting would suggest that Keith should have used at least f/10-11 instead of f/8.
Not sure if it would make a difference in this shot, granted, but Fred has noted a higher f-stop is needed with significant shifting.
David
My shots of the 24mm were at f8 so they are directly comparable. Either I have a really good copy of the TSE or he has a really bad copy.
Here are two images of the extreme edges fully shifted (12mm) of the Canon 24mm TSEII on my a7r2 after adapter modification. The edges of the frame are in the middle.
This modified Metabones 3 adapter has been in a fall in the past, which might explain why the right side is a little weaker (only a problem in this extreme corner), I've got a Metabones 4 coming in the mail, so I'll test that too. I have another Metabones 3 adapter, but I have modified that for my 17mm TSE.
Caleb
Screenshot of lightroom 100% crop of 24mm TSE on A7R2 at 12mm of shift
This dropped adapter was a good test bunny. Especially since I decided to try opening up the rear of the adapter. My other metabones is far better centered.
