Yannick Khong: "The Problem with modern Optics"

[simsurace]

The same can be said for many properties of a lens.   A softer lens can be sharpened more, a lens with CA can have the CA corrected, a lens with distortion can have the distortion corrected.  Where you draw the line and what qualities you prioritize are up to you.

I personally would choose a lens with more pleasing rendering of the range of light and tones than one that is sharper.  Reason being, outside of really large prints, the fine detail where the sharpness really shows is lost anyhow.  The way the light/tone renders is not.  Correcting for the squashed light range the Sigma exhibits above is not as simple as bringing your whites and hilights up and all is well.  Anytime you push and pull on an image's exposure, you get closer to running into artifacts.  I'd rather the lens give me an image that needs less to look good off the bat, so that I have more latitude to shape it.  I have no doubts that most find this negligible, but it is something to be aware of.

There is always a question of where to draw a line. Basically, when it becomes very hard/tedious to compensate for lens defects in software, we have passed a line in my book.

In the gif switching back and forth between the Sigma and Nikon lenses, I find the differences in tone very small. I would predict that if you tried to extract the tone curve which transforms one image into the other (neglecting the geometric distortion effects and other subtleties, one could find such a curve), you would find one which is not significantly different from a straight line (no change to the tones). I would also think that given a set of a few dozen images with different scenes you could extract the curve, but applying it would easily even out almost completely any tonal differences between the lenses.

The main point I want to make is not that one should produce such a curve (although it would be a fun experiment), but that any adjustments which we usually make to our images will have a far greater effect.

I think the worry about artifacts is perhaps not required in the case above. When working in 16bit, you have to push tones really hard to see any artifacts. Any white balance adjustment is messing far more with pixel values than the small tonal adjustment to make the Sigma lens look like the Nikon (or vice versa) would be.

[jhinkey]

On a purely scientific basis, I tend to agree. However, that is not the point. We need to learn to understand that lenses are not objective in the way they capture our images.

BUT the method to compare them must be objective in order to be able to make subjective conclusions.  You have to use the same image subject lighting, etc. which the original article did not do at all.

The above .gif has subtle differences that are problemmatic to subjectively evaluate because the focal lengths are slightly different with different perspective, etc..

[bogrod]

I'm about 6 years late to the 'party' on this thread, but I felt like posting a response regarding Yannick's website. 

There are some interesting perspectives that he has regarding some lenses (in particular those that render so close to what a cellphone camera will produce).  However, the Sigma/Nikon comparison was so close to not really even matter to me as a photographer.  It's something that could so easily be changed in less than a minute in post processing.

He also contradicts himself in some ways.  He doesn't care for 'unidimensional' lenses (a lens that performs exceedingly well in one category, to the detriment of others) such as the 58mm f/1.4G, but then in another area of his website, he think that a photographer should have something more than three zooms and three primes.  In other words, a large collection of glass.  A large collection of glass will inevitably contain one or more 'unidimensional' lenses.

I admit that in years past, I was a hardcore gearhead.  However, what I realize now looking back is that being a gearhead simply gave me insight not only to lens performance and performance criteria, but it also showed me lenses that work for the kind of photography that I want to do.  What Yannick looks for for his own photography is different from others.  After so many years, I know what kind of photography I want to do, and I'd rather have a small collection of 'unidimensional' glass that will do an excellent job for the kinds of shots that I want to take.