Sweet-Spot

I could not resist a try with my new toy. The target is by Bart van der Wolf who posts on LL a lot - it is a variation of the Siemens star with sinusoidal spokes, 144 cycles, about 130mm square - shot with my DMC-GH1 from a meter or so at f=45mm, ISO 100. Big files, sorry, but downsizing invalidates the results :-(

f/22, f/16, f/11, f/8:

http://kronometric.org/phot/iq/f22tof8.gif


f/8, f/7.1, f/6.3, f/5.6:

http://kronometric.org/phot/iq/f8tof56.gif

It can be seen clearly that the extremes of aperture are soft as expected but that f/6.3 thru f/11 are quite useable with f/8 being arguably the winner but barely so.

A slant-edge test would perhaps be more precise, due to the moiré in the above images. In fact, slant-edge tests show that f/11 is in fact slightly better at 1620 lpph. f/22 was fair at 1305 lpph and f/5.6 a poor third at 1095 lpph.

It remains my firm belief that a lens "sweet spot" can not be quantified by shots of natural objects. These quantified results also showed clearly a "useable" range of apertures f/11 thru f/6.3 that would allow setting for depth of field without too much concern that sharpness could thereby be significantly reduced.

It remains my firm belief that tests such as these while sometimes interesting to conduct or read about are pointless in the real world.

If...and its a big if by the way...you intend your images to be viewed at 100% from close range or if...pretty much as big an if as the last if...you are printing them huge and viewing them at very close range then the drop in performance at extreme apertures may be relevant.

If you are going to print them and view them from a sensible distance where you will actually enjoy the image crafted as it was intended then it doesn't matter.
If...lots of if's I know...you are going to resize them for sharing online then you could shoot handheld at any aperture with a rat down your pants and still get a sharp final image.

If you need f22 to get sufficient DoF then shooting at f11 to get a technically superior shot is pointless as the areas you are striving to improve with less diffraction will be out of focus which is a darn sight more noticeable than a slight loss of resolution. If you have to shoot wide open as the light levels dictate it to ensure you have a shutter speed fast enough then stopping down to improve resolution will be somewhat pointless as the shot will be ruined by camera shake.

Perhaps what we need to take account of is that what one photographer considers the 'real world' in respect to his own photographic tools, subject and objective are not necessarily the same as what another photographer considers are his.

What is important is gaining the knowledge to enable one to determine whether he needs to consider these minor variables or not.

Quote:

---

Originally Posted by Stagecoach

Perhaps what we need to take account of is that what one photographer considers the 'real world' in respect to his own photographic tools, subject and objective are not necessarily the same as what another photographer considers are his.

---

Thank you, Grahame.

Quote:

---

Originally Posted by black pearl

It remains my firm belief that tests such as these while sometimes interesting to conduct or read about are pointless in the real world. . . [etc. ad naus.]

---

A strong response, Robin, dripping with condescension. However, you did fail to define what, in your opinion, the "real world" is. For me, as can be plainly seen in my signature, the "real world" - albeit small - is watch photography and, BTW, I do not print anything. Outside, I take snaps and I don't print those either.

http://kronometric.org/tcw/mec/vintage_V/dw512x384.jpg

Such tests are highly relevant in the small but real world of bench-top photography where sharpness and depth of field are arguably the most important factors, unlike the artistic depiction of rolling hills under ever-variable lighting conditions, models' phizogs, etc.

To denigrate the post as "pointless" is quite ungentlemanly, especially as it was for interest only and was certainly not recommended as a course of action for everybody.

OK - so "pointless" could better be "of limited relevance to many applications", but I still think Robin's comments useful to those of us less evolved in our understanding than either of you two.

Koren is helpful on this too, in pointing out that the limits of human visual acuity make anything above 55 LP/mm on full frame irrelevant to a resulting 8x10. (http://www.normankoren.com/Tutorials/MTF.html)

This doesn't mean I don't enjoy detailed and careful exposition of lens performance - I do. So I spent some time trying to see for myself what the Siemens Stars were revealing with engagement. It's just that a lot of my images are not so sharp, and I doubt, in any but the rarest of instances, that it any fault of the lens I used, whether a high performer, or not.

Quote:

---

Originally Posted by xpatUSA

I could not resist a try with my new toy. The target is by Bart van der Wolf who posts on LL a lot - it is a variation of the Siemens star with sinusoidal spokes, 144 cycles, about 130mm square - shot with my DMC-GH1 from a meter or so at f=45mm, ISO 100. Big files, sorry, but downsizing invalidates the results :-(

f/22, f/16, f/11, f/8:
(image)


f/8, f/7.1, f/6.3, f/5.6:

(image)

It can be seen clearly that the extremes of aperture are soft as expected but that f/6.3 thru f/11 are quite useable with f/8 being arguably the winner but barely so.

A slant-edge test would perhaps be more precise, due to the moiré in the above images. In fact, slant-edge tests show that f/11 is in fact slightly better at 1620 lpph. f/22 was fair at 1305 lpph and f/5.6 a poor third at 1095 lpph.

It remains my firm belief that a lens "sweet spot" can not be quantified by shots of natural objects. These quantified results also showed clearly a "useable" range of apertures f/11 thru f/6.3 that would allow setting for depth of field without too much concern that sharpness could thereby be significantly reduced.

---

My conclusion from these images would be that the lens is amply sharp enough, as your resolution is limited by your sensor. (For me) the only visible differences are in the contrast, not in the resolution. So based on this, no resolution sweet spot for this lens (other tests can give different results).

Quote:

---

Originally Posted by black pearl

why?

What possible use can you make of the results?

Are you only going to use your superzoom at its sweet spot to get its optimum performance for a minuscule set of parameters or are you going to go out and take pictures making the most of all of its capabilities?

---

:d:d:d

Ted I didn't intend to upset anyone, nor was I being condescending or ungentlemanly. If you felt that way then I apologise - In the UK we have very fluid take on life and humour so I often forget that many Americans don't. My error, if there was one, was in forgetting this is an international forum and not tempering my turn of phrase to suit all.

I was trying to get across that wasting time (and I do still feel it is time wasted) finding a lenses so called sweet spot gains the vast majority of photographers nothing what so ever. I did add a few ifs and I'll agree there are circumstances where it might be relevant but my concerns were in regard to the OP who may have been lead astray on other forums by people who spend too much time mucking about photographing test charts and not enough time taking actual photographs. No one should limit there range of option when taking a picture to a tiny set of parameters in the vain hope that they will get a better photograph because of it.

What they get may well be technically perfect but it is likely to be a poor image in every other way.