TaoTeJared, that example of the lights in the Sigma's bokeh is the one thing that I dislake about the lens... and from what I thought I saw on a review of the new Sigma 35mm f/1.4 lens it does the same thing.

Dxo resolution test for Sigma on D7000 below. Same story.

and the Bokeh link given from Mansurov's Sigma 50mm f1.4 test also mentions how bad the sharpness is at apertures wider than f2.8.

I don't know how and why we started talking about lenses being "sharpest" wide open. What happened to our general knowledge on best results coming on around 1.5 to 2 stops above ?

Most wide aperture pro lenses are "very sharp" wide open but that doesn't mean they are "sharpest" wide open.

@ TaoTeJared...thanks... I knew my 17 year old 50mm f/1.4D Nikkor was fairly sharp wide open.

Interesting Paperman - Photozone correct? I love how different sites show different results. Maybe others have seen a few others and we can get a reference for each of the lenses.

Resolution... changes with each camera. Tends to be a moving bullet - speaking of which, so is Sigma's sample variation.

Sigma 50mm f1.4 giving best resolution at f1.4 - clearly not the case ... I wouldn't expect it from any lens.

@Corelli - I know what you are saying but I'm not sure you are expressing it correctly. Resolution would be max at maximum aperture due to diffraction laws , correct . But it is still "resolution" that is better at smaller apertures ; not correct to say "overall performance" as that includes lens faults, distortions, CA etc ... Those factors when put together determine the resolution so it IS the resolution that is better or worse.

Thought this was interesting - ran across it looking for something else. Bokeh examples from all the AF 50s
http://photographylife.com/reviews/sigma-50mm-f1-4

TTJ and Correlli, thanks for explaining, I feel like I'm back on track now, I was a little lost :-)

TaoTeJared said:
From how I understand it, and some depends on lens design, but you can pull the optimum sharpness towards the widest aperture. That doesn't mean it will be the highest resolution though.

I would like to put this in a different way: you can pull the optimum overall performance towards the widest aperture. But the highest resolution will always be at the widest aperture.

An ideal lens has got its highest resolution at the widest aperture. If you take an image of an object every point of the object is "transformed" by the lens elements to a disk called the Airy disk. The diameter of this "image point" depends only on the wave length of the light and the aperture. This is one of the reasons why telescopes have got such a large diameter and why there is something called the "diffraction limit" if you close the aperture (which makes this disk's diameter larger).

Now that does not mean that a lens has got its best overall performance at the larges aperture. There are quite a number of optical errors in real lenses that need correction (coma, CA etc). Also what we call "sharpness" is really just a mix of resolution (what kind of small details can I see in the image) and contrast (how "clearly" can I see those small details).

In reality this means that a lens designer has to do compromises (e.g. cost, weight etc) when designing a lens. So even if a lens has got its highest resolution at the widest aperture it might be that other influences (e.g. CA, lack of contrast) renders this resolution useless and images taken at a higher aperture simply look better and show a better overall performance.

chris_weinert said:
Hm, I really feel like I'm missing the whole core of what you guys are talking about here... I've never heard of a lens that has its max. resolution at it's largest aperture, the rule-of-thumb with f/5.6 being the highest resolution so far has always worked, both in my personal experience and what I've encountered in lab tests like DxO. Sometimes (like in the Sigma case) it is shifted towards something like f/4. Might be that there are special-purpose lenses or the cinematography glass that is constructed differently, but don't even "NOCT" and the like lenses, which are optimized for wide-open sharpness deliver higher resolution when you stop down?

At least the lenses we're talking about here do, and I'd say, they do quite dramatically, which is normal, to my experience.

From how I understand it, and some depends on lens design, but you can pull the optimum sharpness towards the widest aperture. That doesn't mean it will be the highest resolution though.

What I have experienced is there is just a bit of focus shift in the lenses and most out of the factory lenses are calibrated at F/2.8. I have a Zeiss ZM 50 f2 prime (rangefinder) that I sent in to be set that way at F2 instead of 2.8. They did send a notice that the focus would be off a bit at F5.6+ but at that stop, the DOF should cover it enough for it not to make a difference. I had the option when I bought the lens to I took it otherwise I would have never bothered. I'm not sure exactly what they do, but I have read people sending their lenses in to have them "set" to their widest aperture.

Maybe someone here knows more.

hmmm or we could all just stick with nikon glass and never worry about any of anything any one of you just said :)

TaoTeJared said:
If you do the comparison or look at the scores, DXO places the highest score settings in the title. "Reached for xxxx" is the title for the scores. From DXO that is where the highest resolution was met.

Hm, I really feel like I'm missing the whole core of what you guys are talking about here... I've never heard of a lens that has its max. resolution at it's largest aperture, the rule-of-thumb with f/5.6 being the highest resolution so far has always worked, both in my personal experience and what I've encountered in lab tests like DxO. Sometimes (like in the Sigma case) it is shifted towards something like f/4. Might be that there are special-purpose lenses or the cinematography glass that is constructed differently, but don't even "NOCT" and the like lenses, which are optimized for wide-open sharpness deliver higher resolution when you stop down?

At least the lenses we're talking about here do, and I'd say, they do quite dramatically, which is normal, to my experience.

(What you actually see of all that when you look at the image in small resolution, small prints etc., that's a separate issue of course, and I'm totally with TTJ and others who state that you wouldn't see (much of) a difference. But we're talking about the mere technical perfomance here.)

But as I said, I seem to be misunderstanding what you guys are talking about.

TaoTeJared said:
On the "who makes what for who" I find the statements to be a loaded derogatory slam at some country/region/company that has no basis on actual performance or quality. Any company's manufacturing no matter what country they reside in, or if it is farmed out to another company, can make the best products in the world.

Just to make that clear, I wasn't suggesting that the Zeiss lenses (or any lenses, for that matter) are worse because XYZ makes them in country ZYX. I was just specifying Cosina vs. Contax because I wasn't sure what lens Adamz was talking about, and because Adamz was asking what Cosina had to do with Zeiss.

And I don't think the 50mm f/1.4 Zeiss is bad because Cosina makes it. The 85, 15 and many others seem to be top notch, and they're made by Cosina, too.

It's more the other way around, just because there's Zeiss on it, doesn't mean it's guaranteed excellence. Or, for some of the really cheap Nikkor consumer zooms: Just because it's all plastic and cheap, it doesn't mean it's a bad lens. Didn't Donaldejose even start a thread about that recently?

TaoTeJared said:
BTW from what little I have seen examples, the CP.2 prime 50mm f/1.5 probably is the sharpest 50 created for Nikon. I guess we will see what the new 55mm one is like as well shortly.

That CP.2 stuff, is that cinematography-related? Ah, nevermind, I'll do a bit of research myself :-)

(Not @ Chris - just in general)
On the "who makes what for who" I find the statements to be a loaded derogatory slam at some country/region/company that has no basis on actual performance or quality. Any company's manufacturing no matter what country they reside in, or if it is farmed out to another company, can make the best products in the world. The only time the quality suffers is when WE THE CONSUMER demand prices that can not sustain the highest quality. Quality suffers because WE ARE NOT Willing to pay for it, not because they do not posses the skill to create it.

If the desirability of the Zeiss ZF lenses are any indicator, I would say they have done a masterful job. Cosina makes Zeiss glass for SLRs, Voigtlander products, the Nikon FM-10, and has made Rollei branded lenses and bodies, Vivitar, Canon lenses, Nikon lenses, as well as lenses and cameras for damn near everyone that has been in the photography business in the last 60 years. To make the leap that they are incapable of producing quality products is unfounded in fact. If you want to see what it costs to have Zeiss branded made in the desired Germany, just look up the prices for the CP.2 prime series. A 50mm f/1.5 lens is $4,500 and the zooms are almost $20,000.

Rant over

BTW from what little I have seen examples, the CP.2 prime 50mm f/1.5 probably is the sharpest 50 created for Nikon. I guess we will see what the new 55mm one is like as well shortly.

chris_weinert said:
Where did you see that? In the DxO field maps, the Nikkors all get better until f/5.6 at least, the Sigma get's better until (I'm speculating) f/4 or so, it loses slightly between 2.8 and 5.6, but gains dramatically when stopped down from f/1.4, too.

Or what did you mean?

Edit: DxO Database, Camera: D3X, Measurements -> Resolution -> Field Map

If you do the comparison or look at the scores, DXO places the highest score settings in the title. "Reached for xxxx" is the title for the scores. From DXO that is where the highest resolution was met. Maybe that has changed since I last read the "how we test" section.

adamz said:
spraynpray - sure You can live, but for 2.8 You have zooms, and I at least would expect 1.4 prime to get it's maximum wide open (f2 the most)

So are you saying that the lens doesn't actually achieve f1.4 or that it needs to be sharper at the largest apertures? I find 2.8 to be quite wide enough to give me decent bokeh in portraits without worrying about which eye to focus on. As for the 2.8 zooms - what a price difference - I got my 50 1.4G for £250!

spraynpray - sure You can live, but for 2.8 You have zooms, and I at least would expect 1.4 prime to get it's maximum wide open (f2 the most)

"The Nikkor 1.4 reaches it optimum at 2.8 and experience has shown me it stays Near the top till f8."

For my use? I can live with that in the real world.

Hope one of you guys reaches out to R8R, this ain't life or death, only images of it.

TaoTeJared said:
Note that the Sigma is at it's highest resolution at 1.4, which for lenses is odd as they always become sharper as you stop down. What Sigma did is set the optimal focus at 1.4, and it degrades from there. The Nikkor 1.4 reaches it optimum at 2.8 and experience has shown me it stays Near the top till f8.

Where did you see that? In the DxO field maps, the Nikkors all get better until f/5.6 at least, the Sigma get's better until (I'm speculating) f/4 or so, it loses slightly between 2.8 and 5.6, but gains dramatically when stopped down from f/1.4, too.

Or what did you mean?

Edit: DxO Database, Camera: D3X, Measurements -> Resolution -> Field Map

Ah ok, it's good to specify that, I thought you were talking about the current (Cosina) f/1.4, which is a completely different story – it's not more than average, just for more money, and without the comfort of AF. The old Contax Zeiss 50mm f/1.4 is a gem, I used that with an adaptor on my 5D II some years ago.

Many of the current lenses marketed as Zeiss are manufactured by Cosina, including the 50mm 1.4, not sure about the Makro-Planar, but I'd think that, too.

I was talking about 50mm f/2.0 Makro-Planar not about Cossina nor Contax. Zeiss and Cosina is not the same company.

Hey Adamz, are you talking about the latest-generation Cosina 50mm 1.4 or about the older Contax 50mm 1.4?

P.S.: Sorry if my post about "I doubt anyone would second that" came across smart-assy, I always forget how small wording details can make a big difference in forum talk. I certainly didn't intend to doubt anyone's, i.e. your, experience. I hate that myself. Same goes for TTJ and his experiences. Hope we're cool :-)

@chris - that's maybe Your experience with zeiss, certainly not mine - based on personal data.

TaoTeJared said:
That it is that soft. Mine is not even close to that at all. I do have fine tuning adjustment set for it.

Ah ok. Well, I had three copies from very different batches, and all were the same results, everything was set up in a way that it was maximally comparable, and again: I of course DID set AF fine tuning, and I describe that in the thread as well. After all, the fine tuning doesn't do anything but shift the auto focus calibration bach and forth, so whatever you've set, you will always get a sharpest level somewhere on that chart. So AF fine tuning isn't not responsible for softness.

Anyone not comfortable with the result (i.e. you :-) ) can of course just do the same thing and post it in comparable size (100% crop D800 max resolution, same 45°-setup, approx. same distance), but I don't see where my test had faults and the conclusion would be wrong, technically.

I agree with you, as I did in the thread, that the practical relevance can be more limited (and a *lot* more limited) than one would assume at first, and that on the other hand, it really depends on what, how and for what end product you're shooting. For me, also described in the thread, the things I tested all matter not only theoretically.

That it is that soft. Mine is not even close to that at all. I do have fine tuning adjustment set for it.

Hey TTJ, are you talking about the AF Finetuning? I did that with all lenses (which was how I got into that test chart thing in the first place). Which conclusion are you referring to as being not fair?

If you recall Chris many of us went through saying we needed to use the micro adjust to get the focus corrected on the 1.4g. Not to say your tests do have some merit, but to say it was the conclusion is not fair.