How do you interpret MTF

When I saw the MTF figures for the Nikon 24-70mm f2.8 on Photozone I had to pick

my jaw up from the floor.

 

A lot of people in the know are all saying good things so I thought, yes, well,

I guess it's possible.

 

Then I started to have a look at my lenses and they were so so, but they seemed

reasonable for what I paid.

 

But when I saw the figures for the 18-55mm II DX, I had to raise an eyebrow.

Some of the figures - say at 24mm - are higher than that from the 17-35mm. In

particular, the MTF values up to 35mm wide open are simply incredible.

 

With today's way of upgrading everything every two or three years, are we about

to see disposable lenses, or are the MTF figures not telling the whole story.

 

(Please, I'm not talking about built quality, vignetting and so on, only MTF

figures)

Andy,

 

When you buy a lens, you are not only buying resolution (more or less, as a quick reponse, MTF). You are buying usability, duration, reliability. With a slow lens, high MTF is useless when you need to take a phot0 in a very low light situation with no flash.

 

The perfect combination would be: high MTF and fast (luminous) lens.

Normally fast and professional lens has high MTF too. But cheaper lens can have, under certain conditions, high resolution, but not high luminosity at the same time.

 

Regards.

"...are the MTF figures not telling the whole story."

 

No, they are not. One issue is that typically they are measuring lenses wide open, so if you compare the 18-55/3.5-5.6 at 18mm to the 17-35/2.8 at 17mm there is a 2/3 of a stop difference (and as we know, stopping down a lens improves its performance). There are other issues as well that are too numerous to go into that almost always translate to better optics for more expensive lenses which are not being measured by MTF numbers; that being said, the 18-55 is considered a very good value as long as you don't beat it around too much -- but hey, it's so cheap it's practically disposable.

<p><i>"How do you interpret MTF?"</i><p>

 

IMO the best way to interpret them is to ignore them. Just my 2 cents.

I don't have any experience with the 17-35 mm lens, but I do know that the 18-55 mm lens has quite a lot of barrle distortion at variou zoom settings. It is possible the 17-35 mm lens does better in t hs regard. Also, the 18-55 mm lens is not as robust a lens and could fail under heavy use. The plastic flange for mounting on the camera broke off within a few weeks in my case, and the lens had to be repaired. Finally, as has already been noted, designers may have to give some on resolution to achieve a larger aperture, as going from f/3.6 to f/2.8.

The MTF you can see in that page is a MTF50, so it is a relative and a subjetive measure. This phenomenon of visual perception in spanish is called "acutancia", do not know traslation to English.

As a simply rule of thumb, not too exact, line widths per picture height (LW/PH)in MTF50, over 2.000 shows an excelent lens resolution capacity, but it not means that a lens with a LW/PH of 1920 is a bad lens. Apparently, the higher LW/PH, the best.

 

Regards

MTF is a reasonable and objective measure of the contrast and resolution of a lens. The data is a continuum, rather than the go-no-go nature of resolution charts (and newspapers taped to a wall). The same data can show the performance of a lens from corner to corner, or as a function of frequency (line spacing).

 

MTF does not, however, put numbers on chromatic and other aberations, nor distortion. All these factors plus build-quality determine the quality of a given lens. Consequently, MTF curves are only part of the information needed to evaluate a lens for possible purchase.

 

Some values are misleading. Canon MTF values are calculated not measured, and tend to be higher than real life (with measurement and manufacturing tolerances). iphoto results are based on measurements using the sensor in a compatible camera, and are lower than either the sensor or lens measured separately (q.v., sum of variances). No matter how scientific the process, someone will find a way to subvert it to support their conclusions ;-)