[butangmucat]

So I am trying to put up a personal review of my 49mm thread version of the Super-Takumar 35mm f2 lens, and this arrived in my mail, which is a publication that my photo shot with this lens and a Sony a5000 body. And I am trying to decide if this lens can be used for actual work today.

Here's the page which the image got used (the original was scanned with an Epson Perfection V750 scanner at 360dpi, and this was magazine-grade printing on heavy coated paper), scaled down to have the long side to be 1042 pixels.



Here's a 100% crop of an area where I can identify CA (original 360dpi image).



Here's the original image I submitted. Which was shot at either f5.6 or f8.



My question: is this CA acceptable for publication purpose, and if is acceptable, to which degree? Another thing is that the user review section seem have not had a lot of problems with CA, and I am getting quite of them, is this because of my copy, or is that this lens is really not well-corrected?

Sincerely

[D1N0]

The CA is very orange and blue but quite thin, you could try to remove it but it will also affect the orange and blue stripe. Desaturate it with the adjustment brush in Lightroom if you are worried about it. Old glass is never as well corrected as modern glass, especially when it is fast. Whether it is acceptable or not? That really depends on you and your customer.

[BigMackCam]

I'd say that's a pretty minimal level of CA. Whether it would be deemed acceptable or not, I honestly don't know. For me, the photo is strong enough in all other ways that the minor CA doesn't detract from it.

EDIT: I just loaded your photo into Lightroom to see if I could remove the CA. By using an adjustment brush to paint the affected high-contrast areas with a Defringe adjustment (see first image), I was able to get rid of most of it. This approach allows you to defringe without affecting the orange and blue livery mentioned by @D1N0 above. The second image shows the end result - not perfect, but a definite improvement.

[butangmucat]

I'd say that's a pretty minimal level of CA. Whether it would be deemed acceptable or not, I honestly don't know. For me, the photo is strong enough in all other ways that the minor CA doesn't detract from it.

EDIT: I just loaded your photo into Lightroom to see if I could remove the CA. By using an adjustment brush to paint the affected high-contrast areas with a Defringe adjustment (see first image), I was able to get rid of most of it. This approach allows you to defringe without affecting the orange and blue livery mentioned by @D1N0 above. The second image shows the end result - not perfect, but a definite improvement.

Because I have moved away from Lr, I actually used the defringe tool in ACR when I was preparing this image for submission. I did some global CA removal but can't remove all of them as then other color artifacts will be introduced. Is this approach similar to use a layer mask with ACR filter?

---------- Post added 12-11-16 at 06:47 AM ----------

The CA is very orange and blue but quite thin, you could try to remove it but it will also affect the orange and blue stripe. Desaturate it with the adjustment brush in Lightroom if you are worried about it. Old glass is never as well corrected as modern glass, especially when it is fast. Whether it is acceptable or not? That really depends on you and your customer.

That publication was not that professional and I did get paid except a free copy, only a small batch made by railfans back in China. They basically told me at midnight (US Eastern Time) that they wanted me to prepare 4 images for press in less than 12 hours, including one needing complex perspective correction. Time was not on my side so I had to rush through the process and any finer adjustment using masks were omitted as getting it ready to press by the deadline was more important. So I only did a global defringe in ACR and applied some basic prepress post based on a best guess on how the image might look once printed on coated paper.

---------- Post added 12-11-16 at 06:56 AM ----------

The CA is very orange and blue but quite thin, you could try to remove it but it will also affect the orange and blue stripe. Desaturate it with the adjustment brush in Lightroom if you are worried about it. Old glass is never as well corrected as modern glass, especially when it is fast. Whether it is acceptable or not? That really depends on you and your customer.

Does orange fringing imply that the lens is not well-corrected for green light? And is there a way to tell whether the CA is axial or lateral? I did stop the lens down to f5.6 or f8 when taking the shot.

---

Last edited by butangmucat; 12-11-2016 at 06:50 AM.

[BigMackCam]

Because I have moved away from Lr, I actually used the defringe tool in ACR when I was preparing this image for submission. I did some global CA removal but can't remove all of them as then other color artifacts will be introduced. Is this approach similar to use a layer mask with ACR filter?

Global adjustments for defringing can, as you've found, affect other parts of the image with the same colours as the fringing.

Unfortunately, I'm not familiar with the workflow you describe, but use whatever means you have at your disposal to locally adjust only the affected areas. If that's via some sort of layer mask, then so be it.

[D1N0]

Does orange fringing imply that the lens is not well-corrected for green light? And is there a way to tell whether the CA is axial or lateral? I did stop the lens down to f5.6 or f8 when taking the shot.

Lenses are made to be sharp in the green part of the spectrum as that is de most important part. The orange fringing is probably caused by IR light registration by the sensor. When you look at the tree behind the train and the rails in front of it, you see that blue fringing is especially bad it also affects sharpness. In a more modern lens it would be purple. This lens wasn't designed for colour photography (everybody was still shooting black and white in the day) so these aberrations were less important. Also UV and IR aren't so much of an issue, as film didn't register it, but sensors do (some more than others, it depends on the strength of the IR filter and the lens.) Using and UV filter or a polariser could reduce the fringing on the blue/purple side.
Another option is the super (multi coated) Takumar 35mm F3.5 It's aberrations are much less as it is a smaller lens.

[butangmucat]

Lenses are made to be sharp in the green part of the spectrum as that is de most important part. The orange fringing is probably caused by IR light registration by the sensor. When you look at the tree behind the train and the rails in front of it, you see that blue fringing is especially bad it also affects sharpness. In a more modern lens it would be purple. This lens wasn't designed for colour photography (everybody was still shooting black and white in the day) so these aberrations were less important. Also UV and IR aren't so much of an issue, as film didn't register it, but sensors do (some more than others, it depends on the strength of the IR filter and the lens.) Using and UV filter or a polariser could reduce the fringing on the blue/purple side.
Another option is the super (multi coated) Takumar 35mm F3.5 It's aberrations are much less as it is a smaller lens.

So it's basically the IR/UV lights causing the artifacts, which did not matter in the film days? Could it also be that the lens failed to bring red and blue to the same focal plane? I am trying to put up an review so am trying to gather these information. I did use a Tiffen UV filter, but not sure whether it was the protector variant or Haze-1 variant. I have standardized on Haze-1 filters now.

I am keeping this lens as it has personal values to me, and I do have a Mir-1A and a Ricoh XR 35/2.8 and both of them seem to have much better CA control.

通过我的 Nexus 5X 上的 Tapatalk发言

[D1N0]

It's probably a bit of both types of aberration. This lens will probably serve you better in more close up situations where the borders are out of focus anyway.

[butangmucat]

It's probably a bit of both types of aberration. This lens will probably serve you better in more close up situations where the borders are out of focus anyway.

I actually wonder if a telecompressor (i.e. Mitakon Lens Turbo) would improve CA control? Math tells me so...

[D1N0]

It compresses the full frame image circle to aps-c, so any ca would also be compressed. The fringing is only going to be worse on the edges when you use the entire image circle, because it increases towards the edges. Whether the compression supersedes the deterioration I couldn't say. I don't think a speedboosters would automagically get the different wavelengths one the same focal plane again (as that would screw up use with well corrected lenses).

[butangmucat]

It compresses the full frame image circle to aps-c, so any ca would also be compressed. The fringing is only going to be worse on the edges when you use the entire image circle, because it increases towards the edges. Whether the compression supersedes the deterioration I couldn't say. I don't think a speedboosters would automagically get the different wavelengths one the same focal plane again (as that would screw up use with well corrected lenses).

What I've looked up is that a telecompressor refocuses the light, and in certain cases can improve MTF. So I guess that a properly-designed telecompressor might be able to correct CA to an extent.

[Digitalis]

My question: is this CA acceptable for publication purpose, and if is acceptable, to which degree?

This depends upon the output size that the image is intended for, for an 8X10 print the IQ is ok, not perfect. For prints larger than 8X10 some correction will have to be made as the aberration will become more apparent.

Lenses are made to be sharp in the green part of the spectrum as that is the most important part.

Correct, Green light is commonly where lens MTF is tested ,Green light is coincidentally where modern DSLRs have twice the sensitivity: the Bayer CFA has more green elements than in either the red/blue.

The orange fringing is probably caused by IR light registration by the sensor.

Incorrect, this is a clear cut case of first order lateral chromatic aberration, which is very common with an achromatic lens. IR has nothing to do with it. UV filters also do nothing to reduce purple fringing.

UV and IR aren't so much of an issue, as film didn't register it

Incorrect, manufacturers went to great lengths to ensure both film and digital sensors are as unresponsive to UV/IR as possible. Silver halides raw sensitivity goes all the way all the way up to gamma rays. But is very limited in IR. The Sony camera the OP is using is known to have an unusual UV/IR filter that could potentially have a narrower or wider bandpass than Pentax cameras have. But from what I have seen so far ,IR contamination is not present anywhere in this image.

I guess that a properly-designed telecompressor might be able to correct CA to an extent.

Unlikely,It would have to be made to exacting specifications to correct for the CA inherent in the optical design of the lens itself. And as we all know, every lens is different regarding control of optical aberrations.

[butangmucat]

The Sony camera the OP is using is known to have an unusual UV/IR filter that could potentially have a narrower or wider bandpass than Pentax cameras have.

I have moved away from a5000, and now uses a NEX-6 as a smaller camera when I don't feel like carrying my Pentax around. Does the NEX-6 have a similar optical filter?