[Recercare]

When looking at Yoshihiko Takinami's resolution tests it seems like many K/M/A lenses have a sweetspot between F11 and F16. This contradicts my experience with manual focus lenses. I usually get the sharpest results between F5.6 and F8. Is this your experience as well? Perhaps Yoshihiko tested on film/analogue and that DSLR gives a different result ?

Kind regards
.lars

[danielausparis]

I had the chance to compare real life lenses and Takinami's results only for three 28mm's (A 2.8/28, M 3.5/28 and K 3.5/28) for which I was able to confirm his findings (K being the best of breed, followed not that far away by the M). However my comparisons were only qualitative by making side-by-side shots.
As for the sweet spot my experience with the K 3.5/28 is still developing; however I have the impression that this lens is very sharp from f/4 on, were it is able to produce one of the sweetest bokehs - which is a purely subjective issue. That's where the figures of Takinami's tests have their limits, because they measure only one aspect (absolute sharpness).

[m42man]

Off the top of my head, I know for a fact that my copy of the SMC-M 50mm f1.7 is sharpest at f4-f5.6 - certainly at the centre of the frame. And the same is pretty much true for all my other 50mm lenses, Rikenons included.

As an aside, I'd say that if you want a cheap, sharp, fast lens, then look no further than a 50mm "standard" lens. The SMC-M (and probably K and A versions) f1.7 isn't, in my opinion, quite the sharpest of these (though it's close), but it's unbeatable for flare control.

[bdery]

I believe those shots were made with film, it might be a little different with digital because of the way the sensor interacts with light. And each lens might be a little different. The sweet spot might also refer, in this case, to the most even result across the frame, or the sharpest, or the one with the more true-to-life colours, or any combination of those things.

I'd say use the lenses in any way you feel is good

[farfisa]

Because of the tight pixel spacing in modern DSLR's, diffraction (the point at which image softening occurs) happens sooner compared to film. Or something like that--google 'diffraction.'

So stopping down past f/8 usually gives you a softer image. Have a look at any of the photozone tests, and they always mention diffraction. The sweet spot on DSLR's is usually somewhere between f/4 and f/8.

[bdery]

Because of the tight pixel spacing in modern DSLR's, diffraction (the point at which image softening occurs) happens sooner compared to film. Or something like that--google 'diffraction

Sorry but that's innacurate. Diffraction is a property of the lens, not the sensor.

[Recercare]

If you compare Yoshihiko's and Photozone's resolution tests of the K 135mm F2.5 there's huge difference. Both are measuring centre and border sharpness. According to Photozone the sweet spot is F5.6 and according to Yoshihiko it's F16.

However, I just stick to 2-3 stops from wide open. Seems to work fine for F2.8-F4 lenses at least.
Fast lenses like F1.4 and F1.7 often need between 3-4 stops for the borders to follow, but I haven't tested real gems like A* 85mm F1.4 though.

Kind regards
.lars

I believe those shots were made with film, it might be a little different with digital because of the way the sensor interacts with light. And each lens might be a little different. The sweet spot might also refer, in this case, to the most even result across the frame, or the sharpest, or the one with the more true-to-life colours, or any combination of those things.

I'd say use the lenses in any way you feel is good

[hangu]

I do not go above f8.0 unless I were shooting macros or the subject matter dictates the extremely wide depth of field. There's a noticeable softening at F16 and F22 on certain ones of my lenses, it almost seems to spit out doubles.

[farfisa]

Sorry but that's innacurate. Diffraction is a property of the lens, not the sensor.

True, diffraction is not a property of the sensor, but the effect on images is also related to pixel size.

That's what this guy says anyway:
Diffraction limits of Resolution | Andre Gunther Photography

[bdery]

I believe you're refering to the circle of confusion, which in turn relates to the resolving power. any optical effect will be affected by the resolution of the sensor, of course, implying that this is particular to diffraction is not accurate.

[whojammyflip]

When looking at Yoshihiko Takinami's resolution tests it seems like many K/M/A lenses have a sweetspot between F11 and F16. This contradicts my experience with manual focus lenses. I usually get the sharpest results between F5.6 and F8.

I think the Takinami resolution tests are of limited use, as light piping in the film will have compromised the resolution measurements. The only way to measure resolution of a lens is using a microscope at the back of it, measuring the aerial resolution, removing the degredation of the sensor/film. I've tried measuring resolution from test charts using film and it was much harder than just using my DSLR, but the DSLR is limited to about 100lp/mm whereas many prime lenses will go higher than that.

There's an interesting review on Ken Rockwell's site of various 50mm Nikon lenses and a Zeiss thrown into the mix. His conclusion is that there is more variation in a lens according to the aperture being shot at, than amongst lenses. Its worth a read:
Nikon 50mm Lens Comparison

Stopping down a tad means you are using a smaller area of the elements, so that chromatic aberations come under control. Conversely, diffraction kicks in at smaller apertures.