[debakarma]

Hello fellow Pentaxians,
Here is a technical question.
(some things I have never managed to understand well – but would like to!).

I was comparing the following 2 lenses on photozone.
Sigma AF 85mm f/1.4 EX DG HSM vs Zeiss Planar T* 85mm f/1.4

Looking at the MTF charts I realised that they were not using the same scale.
The Sig85 has the max Resolution in LW/PH ≅ 4000
The CZ85 has the Resolution in max LW/PH ≅ 3800
Compare the FA77 Resolution in max LW/PH ≅ 2350

What does the Resolution in max LW/PH scales mean? Does higher scale mean different lens class (as in apples and pears) even though they are the same sort of lenses? I read the explanation in Imatest documentation, but I am confused. In other words, if had to choose between the Sig or the CZ (hypothetically speaking) which one should I consider (in terms of sharpness parameter)?

Any guidance would be greatly appreciated. Thanks in advance.

[osv]

klaus makes it clear that all imatest results are tests of the entire imaging system only, so the only way that lenses can be compared is if they were tested on the same model of camera.

dxo attempts to circumnavigate that reality by extrapolating all crop sensor test results to ff before publishing... i don't really get it.

at any rate, were the two lenses in question tested on the same camera?

[debakarma]

at any rate, were the two lenses in question tested on the same camera?

Yes both lenses were tested on Canon EOS FF.

[osv]

when looking at these mtf numbers, you can't use just one number to make a decision like this.

if anything, the zeiss pretty much blows the sigma away in the corners and mid-frame areas.

Sigma AF 85mm f/1.4 EX DG HSM (EOS) - Review / Lab Test - Analysis
Zeiss Planar T* 85mm f/1.4 ZF (ZE) (on Canon EOS) - Lab Test / Review - Analysis

of course there is a price difference, and the zeiss is not af.

if you can live with manual focus on a pentax slr, with it's optical viewfinder, i'd be looking at the samyang/rokinon/bower/etc 85mm primes as well, they are much cheaper, with pretty impressive resolution numbers, if you get a good one.

the fa77 can't be compared to these full frame mtf numbers, because it was probably tested on a crop sensor camera... but it could rock on ff.

[stevebrot]

Ignore the scaling on the graphs. They are meaningless. Instead, compare the numbers for each lens at each aperture. If you wish, put them into Excel and plot your own graphs.


Steve

[Nicolas06]

Hello fellow Pentaxians,
Here is a technical question.
(some things I have never managed to understand well – but would like to!).

I was comparing the following 2 lenses on photozone.
Sigma AF 85mm f/1.4 EX DG HSM vs Zeiss Planar T* 85mm f/1.4

Looking at the MTF charts I realised that they were not using the same scale.
The Sig85 has the max Resolution in LW/PH ≅ 4000
The CZ85 has the Resolution in max LW/PH ≅ 3800
Compare the FA77 Resolution in max LW/PH ≅ 2350

What does the Resolution in max LW/PH scales mean? Does higher scale mean different lens class (as in apples and pears) even though they are the same sort of lenses? I read the explanation in Imatest documentation, but I am confused. In other words, if had to choose between the Sig or the CZ (hypothetically speaking) which one should I consider (in terms of sharpness parameter)?

Any guidance would be greatly appreciated. Thanks in advance.

If the Sig85 && CZ85 have been tested on the same camera then you can compare the numbers. You can't compare with the FA77 that has been tested on an APSC camera with far fewer pixels. The numbers are smaller but are limited by the camera.

You need also to know how to compare theses numbers:
- A few percent error is easy to come in the measurement.
- Sample variation tend to make even bigger differences
- Your eyes would not notice a small variation in the sharpness numbers.

Depending if you are planning to use them on the future FF mainly or on APSC it can make the FF results irrelevant. Different lenses behave differently on different sensor size. You'd need to look at the APSC performance if you plan for an APSC usage and FF performance if you plan for an FF usage.

Also MTF are mesured differently on different review websites. My understanding is that photozone get higher number than some other because they accept more contrast drop than some other website.

Constrast/Micro constrast and the whole rendering (including colors and out of focus transitions) can be much more important and visible than a 5% difference in resolution at low contrast levels. In particular 4000 is very high, The best a K5 can do being 2700 for example and that allows for A3 shoots at 300dpis ! You start to really notice 4000 numbers in 30x40" prints or big crops... And even only compared to much lower numbers like 4000 compared to maybe 3000, not 4000 compared to 3800.

Also please note what is important is also the performance as the apperture you will use in practice and on a area of the sensor that has a probability to be in focus. A 85mm make for great landscape but if you buy a 85mm f/1.4, that's for the the wide apperture performance I think.



Anyway, this mean 3800 vs 4000 is totally ireleavant, too many factors.

[debakarma]

Thanks Nicolas, Thats a whole lot of new info I need to digest.
Yes I will be using the 85s for portraits of elderly people. For the moment I will limit myself to APSC only.
I plan to capture some skin/facial imperfections in great detail - to build the character.
The border sharpness becomes important where background also needs to be relatively sharp in order to balance and compliment the whole frame. I currently interpret the border to center sharpness ratio when I look at MTF charts.
And my questions relates to if I should be looking for more info contained in the MTF and which is particular to specific lenses.
Thanks again.

[Nicolas06]

If you want the background sharp, you want to close down the lense and have the background near your subject. Check dof master here: Online Depth of Field Calculator to lean more about deph of field.

Using a 85mm f/1.4 for a head shoot (say 2m shooting distance), you get 3cm of deph of field at f/1.4. if the eye is in focus, the nose is not. At f/16 you get 32cm in focus just enough to have everything in focus if the guy is on the wall that serve as background. A wall that would need some texture for the sharpness to show.

Even on 5m setting, you get only 18cm of deph of field at f/1.4. You'd need f/4 or f/5.6 to ensure you have a sharp near background and f/22 would still give only 3.2meters of deph of field.

The resolution figure you see are anyway the very fine details of the smaller skin hair that would appear only if you zoom at 100% on a computer or look with a magnifying glass on a 30x40" print.

For getting the character as you want, what you need is micro constrast and the performance of the lense bellow 1500 LW/PH with the best possible contrast at theses settings. Very high resoltion results are typically irelevant for what you are after are they are for all portraiture related work.

You don't need a big lense, you don't need wide apperture. You need a lense with punchy colors/constrast and also to master you post processing skills.

Look to the next shoots why one would get a wide apperture lense. Most of the subject is out of focus and the background is completly blured to make the subject pop. This help soften the rendering, hide the skin defects (even through you can see them on the boy by zooming)

FA77 f/1.8



FA77 f/2



Look then how much more visible and contrasty a picture look like (with the same FA77 f/1.8) closed down to f/4.5... How much more it make all the skin defects more visible. See how the background is still completely blured because it was taken outside:

FA77 f/4.5




To me a constrasty and much cheaper lense would do a better job. That can be a macro lense, a type of lense that excell at sharpness and high contrast (DFA50 macro, DFA100 macro, sigma 70 macro) and is often avoided for portraiture because it is too sharp, too constrasty and not enough forgiving. But that what you are for. This kind of lense is also sharp borders to border.

You can also rely on a DA70. Very constrasty and punchy colors. FA77 or FA43 would add more 3D/pop to the rendering. That can be interresting.

One thing is sure, you should not care of extreme sharpness results of very fine details just to get a eldery portrait with lot of character!

[debakarma]

Nicolas, thanks for the tip on the macro lenses. I will try that out one of these days.

[goubejp]

in LW/PH, PH means Picture Height. It gives more or less the absolute resolving power of the lens. Of course it dépends (among other things) of the sensor size. That's why, on photozone, you'll find different figures for the same lens tested on APS-C and full frame - see Nikon or Canon.
Some figures are given in LW/mm (like lenstip.com) which is a way to normalize result and make comparison possible between lenses, though there are differences between makes and models that prevent fair comparison. That's why Klaus (photozone) writes that cross-brand comparison is not possible and gives LW/PH

[debakarma]

i That's why Klaus (photozone) writes that cross-brand comparison is not possible and gives LW/PH

Thanks, that was really my missing clue.