If I have learned anything on this forum, it's that MTF are just an indication. The real test is in how the lens performs in real life, as shown in Steve's comparison with the 24-70. Theoretically the 24-70 is better at 24mm, but in practice the prime's samples are outstanding. I'm sure we will have a better understanding of this lens when more samples are available online.
This said, I must confess that Sony's pricing is still a mystery to me. How can they sell this lens for 1250$ (B&H pricing) while the ZF.2 25/2.8 is selling for 1000$. Logically the ZA should cost much more, with 1 extra stop and AF. Same for the 85/1.4, the ZA is only about 80$ more than the ZE/ZF.2.
edwardkaraa wrote:
If I have learned anything on this forum, it's that MTF are just an indication. The real test is in how the lens performs in real life, as shown in Steve's comparison with the 24-70. Theoretically the 24-70 is better at 24mm, but in practice the prime's samples are outstanding. I'm sure we will have a better understanding of this lens when more samples are available online.
What you CAN see in the MTF's is that they relied a bit too much on the aspherics when simplifying the design. An aspheric can in a best case scenario replace up to three lens/air surfaces (two spherical lenses gives four sph. air/lens surfaces to model the light beam > one asph give one sph. and one asph. air/lens surface).
It can also be used to fight distortion, but at the cost of highly irregular sagital/meridonal performance (the "waviness" in the meridonal/tangential MTF line). Keeping the S/M lines as close and similar as possible retains plasticity better in the scene.
edwardkaraa wrote:
This said, I must confess that Sony's pricing is still a mystery to me. How can they sell this lens for 1250$ (B&H pricing) while the ZF.2 25/2.8 is selling for 1000$. Logically the ZA should cost much more, with 1 extra stop and AF. Same for the 85/1.4, the ZA is only about 80$ more than the ZE/ZF.2.
I think there are 2 factors at work. Sony/Minolta/Tamron/whoever-else probably is using more automated, mass production techniques than Cosina which is more traditional batch construction that probably involves more hand assembly. Further, while the Sony/Zeiss lenses are not poorly made by any measure, the Zeiss ZF/ZE lenses are an all-metal construction which generally costs more.
Lotusm50 wrote:
I think there are 2 factors at work. Sony/Minolta/Tamron/whoever-else probably is using more automated, mass production techniques than Cosina which is more traditional batch construction that probably involves more hand assembly. Further, while the Sony/Zeiss lenses are not poorly made by any measure, the Zeiss ZF/ZE lenses are an all-metal construction which generally costs more.
I also think the glass used in the lens design is a factor. The ZF/ZE/ZS/ZK lenses generally don't use aspherics, but instead rely on more exotic glasses and spreading out the power and correction over a greater number of elements (design relaxation). So if more glass and more "exotic" glass types are used, this will increase the cost of the raw materials going into the lens.
e_dawg wrote:
I also think the glass used in the lens design is a factor. The ZF/ZE/ZS/ZK lenses generally don't use aspherics, but instead rely on more exotic glasses and spreading out the power and correction over a greater number of elements (design relaxation). So if more glass and more "exotic" glass types are used, this will increase the cost of the raw materials going into the lens.
What is your source for the use of more expensive, "exotic" glass in the ZF/ZE/ZS/ZK lenses?
theSuede wrote:
What you CAN see in the MTF's is that they relied a bit too much on the aspherics when simplifying the design. An aspheric can in a best case scenario replace up to three lens/air surfaces (two spherical lenses gives four sph. air/lens surfaces to model the light beam > one asph give one sph. and one asph. air/lens surface).
It can also be used to fight distortion, but at the cost of highly irregular sagital/meridonal performance (the "waviness" in the meridonal/tangential MTF line). Keeping the S/M lines as close and similar as possible retains plasticity better in the scene.
So is this responsible for the edgy bokeh that some associate with aspheric lens designs? People often say that S/M divergence on the MTF graph is indicative of rough bokeh. I always thought that it was due to astigmatism and lateral colour, but i haven't been able to make the connection between astigmatism and edgy bokeh. Can you help me understand?
I'm going crazy trying to find a link to a factory tour video showing various Sony ZA and G lenses being assembled, but I can't find it. For those interested, here are a few short videos showing Sony lenses being manufactured.
douglasf13 wrote:
I'm going crazy trying to find a link to a factory tour video showing various Sony ZA and G lenses being assembled, but I can't find it. For those interested, here are a few short videos showing Sony lenses being manufactured.
S/M divergence does not have to indicate rough bokeh - but the kind of waviness that the ZA lens shows is often visible as large differences in character depending on exactly where on the image circle you look. If the M lines take a steep dive towards the outer edge you get compressed bokeh, not like the cat-eye effect you get by hard clipping (mechanical vignetting), but an uneven spread that looks like UFO's circling the center of the image. It is also very often a sign of coma.
The delta (first derivative) of the lines are as important as the lines themselves. A smooth, slightly lower 40lp/mm line might be preferable to a wavy, slightly higher 40lp/mm line.
An aspheric introduces more and more non-symmetrical aberrations the further away you place it from the aperture. But everything depends on the nature of the asphere surface, where the largest deviation from a pure spherical surface is, if it's a positive or negative deviation and so on.
Tariq Gibran wrote:
Jeez, that looks extremely labour intensive to me.
Yeah, I wish I could find the couple of other videos that I've seen. They included some ZA lenses on the line, and, if I remember correctly, they were being assembled and tested right alongside the Sony G lenses in Japan.
theSuede wrote:
S/M divergence does not have to indicate rough bokeh
So in this case, what does the S/M divergence indicate (not the waviness at 40 lp/mm at f/8, specifically, but the general divergence as is common in most lenses)? In the f/2 graph, my take is that the drop in sagittal MTF implies increasing astigmatism across the frame and field curvature that really kicks in around 16-18 mm. The drop in meridional MTF towards the edges indicates coma, but how would we know for sure since the field curvature dominates?
And in the f/8 graph, i would normally say that the drop in meridional MTF at 40 lp/mm indicates uncorrected lateral colour along with some astigmatism, but if the test was done with monochromatic light (which it probably was to get higher results), how would you account for such a significant drop in the meridional? At the edges, the field curvature is clearly apparent at 18 mm with the massive drop in sagittal MTF, and yet the meridional MTF drops at the edges too. But it can't be coma at f/8, can it?
If the M lines take a steep dive towards the outer edge you get compressed bokeh, not like the cat-eye effect you get by hard clipping (mechanical vignetting), but an uneven spread that looks like UFO's circling the center of the image. It is also very often a sign of coma.
Is this what is commonly referred to as sagittal coma flare? Or is that when the S lines drop at the edge?
An aspheric introduces more and more non-symmetrical aberrations the further away you place it from the aperture.
Interesting... I often see lenses designed with an aspheric in the rear group as the last or second last element. I can't count the number of times i've seen that, acutally. And yet you couldn't get much further away from the aperture.
But everything depends on the nature of the asphere surface, where the largest deviation from a pure spherical surface is, if it's a positive or negative deviation and so on.
What do you think of double-sided aspheric elements Olympus and Voigtlander use to make the lens as small as possible?
Lotusm50 wrote:
Just thinking out loud here, but maybe the "redesign" of the ZF 25mm f2.8 is due to Sony not accepting Zeiss' design for the ZA 24mm f1.4. Maybe the "redesign" is just an outlet for their 24/1.4 design Rejected by Sony as too expensive (given their "understanding" of their "target market"), Zeiss turns around and adapts it to their ZF/ZE lines of lenses.
OK. Maybe this isn't "thinking", but more "dreaming". ;-)
To follow my line of thought (or dream) here, the press release for the new Zx 35/1.4 says, "By introducing the Distagon T* 1,4/35, Carl Zeiss is complementing the Planar T*1,4/50 and T*1,4/85 lenses with a wide-angle lens that shares the same high speed." Now wouldn't it be nice if they complemented the line further with a 25/1.4? Then the line-up of f1.4 lenses would be complete! I mean, they have to be thinking about it. Many years ago they had a prototype 25/1.4 for Contax...
Aspherics has to be used with some discretion - they CAN do a lot of good, but more often than not they're just used to simplify/shrink a construction. As always, the result is all up to the designer (within a certain budget! :-) ).
Imagine the full monty of the original Zeiss design, giving the designer the choice to aspherize one or two surfaces just to increase picture quality, without simplifying the construction... :-)
Sagital coma flare indicates how large the comatic spread is, how far from the chief ray the projection of the light from the outer rim of the aperture deviates. A large sagital coma flare puts all of the light energy from the largest aperture well outside the F/8 point spread, which is why I guess they chose the denomination "flare" for this aberration. In a normal construction without aspheres, coma increases with radial distance.
Pure field curvature will cause BOTH of the S/M curves to drop - with the same relative percentage! If just one of the curves drops this indicates an assymetrical aberration. This might be ANY kind of assymetrical aberration.
That there will be some difference between the S/M lines is quite natural - think about it. If you send a bundle of rays towards a spherical surface - at an angle - the spherical surface is no longer symmetrical. The relative (to the ray) curvature will be stronger in the S dimension than in the M dimension of the polar coordinate system if the surface curvature is positive, and weaker if the curvature is negative...
Some of these things can be deduced from exactly HOW the lines differ, and if one curves but not the other and so on - but it's mostly guesswork. Optical design is so complex that many types of aberrations can give the same effect on a MTF curve.
Lotusm50 wrote:
To follow my line of thought (or dream) here, the press release for the new Zx 35/1.4 says, "By introducing the Distagon T* 1,4/35, Carl Zeiss is complementing the Planar T*1,4/50 and T*1,4/85 lenses with a wide-angle lens that shares the same high speed." Now wouldn't it be nice if they complemented the line further with a 25/1.4? Then the line-up of f1.4 lenses would be complete! I mean, they have to be thinking about it. Many years ago they had a prototype 25/1.4 for Contax...
I'd be much happier if they complete their f/2 macro line with a 25/2 macro......
