Petegh wrote:
Yeah, I'm a bit dubious about testing telephoto lenses for microcontrast at infinity: the risk of negative atmospherics is too high. When Lloyd compared using short-midfield planar subjects, the difference between pre and post correction on the 85 DGDN was clearly visible; although Lloyd has the advantage of shooting in rarefied atmospheres, I spose.
Time of day and elevation are critical. Also choosing the right distance to count as ♾ . (Not too close so focus is an issue, not too far for atmospheric reasons, obviously that varies with FL)
I’m more dubious about non infinity tests outside a laser calibrated lab.
But the crucial thing is that anything which is hard to see peeping is irrelevant at any viewable whole image size. So whenever I see people arguing about which life size crop is best, it’s time to move on.
bjornthun wrote:
No, the Sigma 90/2.8 has excellent contrast whether I view the image as a whole or zoom in on details.
---------------------------------------------
This still doesn’t preclude the net result from being better than what could be avhieved with opticall distortion correction. It is not a given that optical correction is lossless. Think of what happens when a teleconverter stretches an image in order to do its job.
No one is suggesting that the Sigma doesn't have good contrast, just that it doesn't have great microcontrast. Take a picture with the Zeiss Otus 100mm f1.4, and compare the same image with the Sigma, and you'll see the Sigma doesn't have great microcontrast, whereas the Otus has it in spades.
Lens software distortion correction is mostly about money. It fits Sigma's recent business model well: that is, to produce lenses as small and light and as cheaply as possible, so they can then sell them relatively cheaply; its all about maximising performance relative to size,weight and retail price.
I've never seen any science that proves optical correction of distortion is a lossy process. I think the best evidence that optical correction is superior to software correction, is that most premium lenses employ optical correction.
Petegh wrote:
No one is suggesting that the Sigma doesn't have good contrast, just that it doesn't have great microcontrast. Take a picture with the Zeiss Otus 100mm f1.4, and compare the same image with the Sigma, and you'll see the Sigma doesn't have great microcontrast, whereas the Otus has it in spades.
Lens software distortion correction is mostly about money. It fits Sigma's recent business model well: that is, to produce lenses as small and light and as cheaply as possible, so they can then sell them relatively cheaply; its all about maximising performance relative to size,weight and retail price.
I've never seen any science that proves optical correction of distortion is a lossy process. I think the best evidence that optical correction is superior to software correction, is that most premium lenses employ optical correction....Show more →
Zeiss Otus 100/1.4 is 6-7 times more costly than a Sigma 90/2.8, so yes it ought to be the better lens of the two.
I’ve not seen any science to prove that optical correction isn’t lossy. Why did DSLR wide angles have moustache distortion? It was because optical correction couldn’t achieve straight lines. Digital correction eventually came to the rescue.
bjornthun wrote:
Zeiss Otus 100/1.4 is 6-7 times more costly than a Sigma 90/2.8, so yes it ought to be the better lens of the two.
I’ve not seen any science to prove that optical correction isn’t lossy. Why did DSLR wide angles have moustache distortion? It was because optical correction couldn’t achieve straight lines. Digital correction eventually came to the rescue.
The price difference is irrelevant. When you judge a lenses microcontrast, it has to be in the context of all other lenses on the market, and the Otus is at the top.
DSLR wide angles have to be of retrofocus design, because their focal length is less than the distance of the mount from the sensor. This was a major hassle for lens designers, and yes, as you say, made optical distortion correction more difficult. Mirrorless, and modern aspherics etc have largely solved that. Remember, distortion is nothing more than a variation in focal length across the frame.
Petegh wrote:
The price difference is irrelevant. When you judge a lenses microcontrast, it has to be in the context of all other lenses on the market, and the Otus is at the top.
DSLR wide angles have to be of retrofocus design, because their focal length is less than the distance of the mount from the sensor. This was a major hassle for lens designers, and yes, as you say, made optical distortion correction more difficult. Mirrorless, and modern aspherics etc have largely solved that. Remember, distortion is nothing more than a variation in focal length across the frame.
Price does matter.
Most or all modern wide angles are retrofocus, including rangefinder and mirrorless wides.
This is off topic, so I’ll say no more about this here.
bjornthun wrote:
Price does matter.
Most or all modern wide angles are retrofocus, including rangefinder and mirrorless wides.
This is off topic, so I’ll say no more about this here.
Ok, I understand now, you were actually arguing that the Sigma's microcontrast is good relative to its price. My bad.
Petegh wrote:
The price difference is irrelevant. When you judge a lenses microcontrast, it has to be in the context of all other lenses on the market, and the Otus is at the top.
DSLR wide angles have to be of retrofocus design, because their focal length is less than the distance of the mount from the sensor. This was a major hassle for lens designers, and yes, as you say, made optical distortion correction more difficult. Mirrorless, and modern aspherics etc have largely solved that. Remember, distortion is nothing more than a variation in focal length across the frame.
Sorry, judging micro contrast is an objective measurement and doesn't have to be done in the context of all other lenses on the market. Sure, the Otus has wonderful micro contrast, but it is a lens that weighs 1.5 kg and is nothing like any of the lenses we are discussing here. One would be unlikely to buy the Voigtlander 90 f/2.8 APO or the Sigma 90 f/2.8 as an alternative to the Otus. If you are willing to pay over $4,000 for the 1.5kg Otus, then you probably aren't considering the sub 300g less than $800 Voigtlander 90 f/2.8 APO or Sigma 90 f/2.8 (these lenses are literally 5 times lighter and 5 times less money). Because of the difference in size we are talking lenses that aren't really in the same category. So comparing them, to me anyway, makes little sense. I think Fred's tests and Rich's tests provide useful information. I would love to see MTF charts as well, but so far Cosina hasn't proved them (they do for the APO Lanthars, however, so they might; Sigma typically does not provide measured MTFs). What we really want to know for landscapes is how well a lens renders fine detail at a distance. Fred and Rich tests are not perfect, but they do help us answer that question. The question of whether it resolves such fine detail as well as the 100 Otus is not really relevant (of course they don't) but comparing them might be interesting to see how close they get. Maybe you or someone else would like to do that test.
Fred Miranda wrote:
Perhaps aside from the slower f/2.8 aperture, Cosina decided to name it APO-Skopar instead of APO-Lanthar, because they knew is was not a reference 85-90mm lens like the other APO-Lanthars. I have tried 3 loan copies of the CV 90/2.8 APO and that's how it performs.
I’m also suspicious why they did not publish MTFs for it, unlike the other APOs.
After receiving the comments regarding his Nikon version of the Apo-Skopar from Vivek, my results, and some suggestions to contact Stephen at Cameraquest from Ron, I called Stephen. He indicated that since I was not happy to send my copy back (I will send it back tomorrow in the same box that it was shipped to me) tomorrow via USPS Priority Mail. Stephen should be getting additional copies of the Black lenses in early January and my lens should arrive about the same time as he is receiving new copies. He will send me another copy.
naturephoto1 wrote:
After receiving the comments regarding his Nikon version of the Apo-Skopar from Vivek, my results, and some suggestions to contact Stephen at Cameraquest from Ron, I called Stephen. He indicated that since I was not happy to send my copy back (I will send it back tomorrow in the same box that it was shipped to me) tomorrow via USPS Priority Mail. Stephen should be getting additional copies of the Black lenses in early January and my lens should arrive about the same time as he is receiving new copies. He will send me another copy.
Rich
Hi Rich,
From the samples you posted, I don't think there is an issue with your copy but the best way to find out for sure is to do a proper centering test.
As we've seen in this review, the Voigtlander 90/2.8 is not the highest resolving 85-90mm lens. It has many positive attributes but it's not breaking any resolution records. I have shot with the Zeiss 85/4 ZM for many years and it's capable of higher resolution, especially off-axis.
Also, although they share the same optical design perhaps there is a small difference in performance between the Nikon and M mount versions.
Petegh wrote:
No one is suggesting that the Sigma doesn't have good contrast, just that it doesn't have great microcontrast. Take a picture with the Zeiss Otus 100mm f1.4, and compare the same image with the Sigma, and you'll see the Sigma doesn't have great microcontrast, whereas the Otus has it in spades.
Lens software distortion correction is mostly about money. It fits Sigma's recent business model well: that is, to produce lenses as small and light and as cheaply as possible, so they can then sell them relatively cheaply; its all about maximising performance relative to size,weight and retail price.
I've never seen any science that proves optical correction of distortion is a lossy process. I think the best evidence that optical correction is superior to software correction, is that most premium lenses employ optical correction....Show more →
If money is unlimited, and there are few constraints on size, optical correction can be relatively unlossy, though not completely so. Or so a lens designer tells me.
But for a fixed size and to some extent budget if you want optical correction it will trade off against other things, so that you often end up with a better result concentrating on resolution and contrast, and correcting distortion in post.
Thus you’d expect big no holds barred lenses to sometimes be best if corrected optically, and compact premium ones and budget ones to usually be best if correct electronically.
Many reviews already don't mention (for example) simple distortion correction and many manufacturer blurbs lack any such information as well. If they were proud of it as an achievement, they would be trumpeting it from the rooftops - but it's just the opposite.
From the B&H page extolling this recent (11/10!) 90mm lens, we see a 320 word summary that makes no mention of the very significant distortion, or options for its correction.
The maker's own page does better, but gives fair warning of what is to come:
"The lens also utilizes the camera’s optical correction functionality, which is an advantage of mirrorless systems, SIGMA’s optical designers were therefore able to concentrate on *correcting aberrations that can only be corrected by the optical design*, thereby improving rendering performance and reducing the size and weight of the lens." (asterisks are mine)
So the optical engineers will work on the optical design flaws *only* if they have no computing alternative, as carried out by their software engineers. The clear implication is that if all aberrations could be corrected digitally, they would do so and feel justified in doing it.
They offer no alternative medium size lens weighing, say 75g more, but with sound optical correction. It's the top of the slippery dip, and the trip down the slide only gets faster.
Fred Miranda wrote:
Hi Rich,
From the samples you posted, I don't think there is an issue with your copy but the best way to find out for sure is to do a proper centering test.
As we've seen in this review, the Voigtlander 90/2.8 is not the highest resolving 85-90mm lens. It has many positive attributes but it's not breaking any resolution records. I have shot with the Zeiss 85/4 ZM for many years and it's capable of higher resolution, especially off-axis.
Also, although they share the same optical design perhaps there is a small difference in performance between the Nikon and M mount versions....Show more →
Hi Fred,
Since you have brought up the question of the centering of my lens, I will present my photos taken of a wall hanging with my tripod mounted A1 and the lens. These were taken of a wall hanging in our Dining Room. Unfortunately there is a rod that is visible in the lower right corner, but you can still get an idea from these images. The wall hanging was about 7' away from the tripod mounted camera which was leveled using the level on the tripod head and the camera. I can not guarantee how accurately the camera was equi distant from the left side and the right side to the wall hanging. Images are SOC camera with LR defaults and only made into JPEGs from the original RAW Compressed files.
Rich,
It does not look centered from this test but...
It's hard to know if you were squared with the target, so I suggest doing this test at infinity distance so we can remove this variable and get an accurate assessment.
I have tried 3 loan copies and they were all centered. One was perfectly centered even on my A7R IV and the other two were almost as good. The issue with some copies was rangefinder calibration, not centering.
?
