Thanks for your article.

Secondly from a technical point of view resulting values (such as definition or resolution) can be compared using the numbers, but artistically seen, who can say that a particular 'flaw' of your lens couldn't make all the difference in a picture.

As is true know thyself it is also true know thee equipment.
And thanks for your article.

Charles: MTF graphs will probably show it; a graph is the contrast between black/white at a given resolution (line-pair/mm) at aperture. Along the X-axis you have the distance from centre of image, and then two curves in the range 0-100% for the contrast measured radially (along the radius) and saggitally (perpendicular to the radius). Normally these lines wiggle around a bit and then drop off towards 0 as the radius approaches the coverage of the lens. So you want a graph for each aperture to compare, and to be looking for the lines to be high-up % and stay consistently close together (no use being sharp radially but not the other way) as far towards the right as possible.

I don't know if you know about Matt Klaskowski? (lightroomkillertips.com) He suggest in a video with dtowntv.com that you always should shoot landscape at 11 - 22... I just laughed. 11 ok, but higher, no way. Then its better to merge two images together

This statement implies that less of the front element is used when stopping down and that implication is just not true. A large portion of the glass is used regardless of the aperture used. Although it is true that each image point uses only a small part of the front element at a smaller aperture, the image as a whole still relies on a large part of the front element.

It is also true that localized imperfections in the glass will affect fewer image points at smaller apertures, and therefore, less of the overall image, than larger apertures. Maybe this was what you meant, but I just wanted to expand on that point.

Contax bothered to make the Contax Zeiss Planar 50mm as fast as f/1.7, so that it would out perform other lenses at it's best - f/4 for 'most' uses. Just like Nikkor made the 85mm f/1.4 so that it would rock at f/2.8.

The only exception I know of, is my Nikkor 200mm f/2.0. I often shoot that monster wide open with outstanding results. However, just like the OP, I fall back on the old rule of thumb often: the optimum aperture for sharpness and detail is about two stops away from wide open - and/or closed.

For large prints, as John rightly points out, it is important to pay attention to aperture selection; however, processing and printer capabilities/settings are equally important to obtain the sharpest details in the print.

Generally, the writer is spot on with the optical science, but readers should not go overboard avoiding use of stops at the extremes. Diffraction effects come into play when the physical aperture diameter approaches the wavelength of light used. But in reality the sensor pixel spacing is also important to note.

The basic formula for lens resolution (applying the Lord Rayleigh criterion) is R=2000/f#. For f/64, the resolution would be ~30 lines per mm; for f/2 it would be 1000 lines per mm. What this means is the lens set to f/64 can resolve two lines spaced ~33 microns apart. For f/2, the lens can "resolve" 1 micron spaced lines. This assumes no image degradation from other lens aberrations and is a simplistic view.

Ultra-high pixel count sensors (25 megapixels) have pixel spacing ~4.5 (Canon 5D) to 5.94 microns (Nikon D3X) (1 micron = .001mm). So you can see that somewhere below f/11 the lens resolution is roughly equal to the pixel separation and the sensor dimensions become the limiting factor. For apertures smaller than f/11, diffraction "spreads" the edge or spot, thus limiting the sharpness of detail recorded.

Lens designers try to optimize lens resolution at a particular aperture (usually f/8). As you stop down or up from there, aberrations (like chromatic aberration, coma, astigmatism, etc.) reduce max resolution. The diffraction limited aperture (DLA) is always higher (like f/16 or f/22).

This is getting complicated!

Bottom line, take a series of pictures of a detailed object with a full range of f/stops. Compare and decide for your camera sensor-lens combination where diffraction and/or your lens aberrations limit the detail or sharpness of the scene.

But don’t be afraid to use widest or smallest aperture settings. Other factors (pixel spacing, camera or subject motion, sensor noise, lens aberrations – so use expensive glass if you can afford it, defocus because the subject moved slightly, etc.) often come into play.

Dear John Sevigny, ask yourself this: why did the people at Contax bother to make the Contax Zeiss Planar 50mm as fast as f/1.7, if it performs at it's best at f/4 ? Because "best" is a flexible word, and f/4 is often not the best aperture.

The classification "best" is not just about sharpness and crispness.

Furthermore sharpness is really only one of the factors in a photo and in fact it's the most overestimated factor (except when it comes to catalog product photography stuff, but that lacks any form of photographic creativity imho)

I do hope you understand that the f-number does not only exist for getting the "best EQ" out of a lens, but more importantly for two other reasons:
- *control* the depth of field
- *control* the amount of light entering

So if you want to make a photo with a shallow depth of field in a certain composition using a 50mm lens, you'll have to go to for instance f/2 (or you move your subject farther from the background....but that changes the composition and not the DoF!)
Also, at dusk even ISO6400 is quickly too little if you keep insisting to take that shot at f/5.6 ; you'll simply need to open up that aperture.

But I believe you go overboard in a few places. For example, you state: "f16... is useless at 35mm or for DSLRs. Diffraction... destroys images at f16. You might rule out using f11 for the same reason."

No. On a full frame DSLR (or 35mm film camera) f/11 produces excellent and very sharp images. Yes, diffraction does decrease the maximum potential sharpness at these apertures, but the difference is very tiny - to the point of being negligible in most case. In addition, a number of lenses continue to become sharper in the corners as you stop down, even as center sharpness might decrease a bit due to diffraction. Which is sharper: the lens that gets the optimal center sharpness but noticeably softer corners or the lens that has ever so slightly less sharpness in the center but is more uniformly sharp across the frame?

That said, I agree that it is very useful to understand the real world performance of your lenses at various apertures - to know how and where they are strong/weak and to understand how this can be used to make more effective use of your lenses.

Take care,

Dan