(Page 1 of 1 pages for this article )

Filed under: Cameras •Hardware •

Sharpness is a Complex Thing

Limiting resolution is easy to determine, but there’s a whole lot more to image sharpness and overall quality than resolution. For example, you can optimize a pinhole camera for maximum resolution, OR you can optimize it for a sharper-looking image!

Check out “The Pinhole Camera Revisited, or, The Revenge of the Simple-Minded Engineer”:

http://www.biox.kth.se/kjellinternet/Pinhole.pdf

If you want to get into it more, look into Zeiss’ “How to Read MTF Curves”, at:

http://www.zeiss.com/c12567a8003b58b9/Contents-Frame/d9a30a166f326924c125751a004ab770

There are two PDFs there; read the first one (the link at the bottom of the page, not the one in the sidebar), then follow the link to “Part 2” and read the PDF linked at the bottom of that page.

If, after you read them, you still want to fight the pixel-count wars, be my guest… but leave me out of it; I’m more interested in pictures than in pixels!

[With a tip o’ the hat to Don Craig for the Zeiss links, and Gary Powe for the Pinhole PDF.]

(Page 1 of 1 pages for this article )

Very nice read, thanks for sharing. Be sure also to watch the excellent 7-part video conference by Larry Thorpe (Canon Broadcast) and John Galt (Panavision) at http://media.panavision.com/ScreeningRoom/Screening_Room/Demystifying_Part1.html - it can serve as a valuable introduction to the more technical whitepaper by Zeiss. I had found that conference slightly frustrating in depth though, and Zeiss complements it very nicely with necessary conclusions and all the picture examples.
In the paper by Zeiss, there’s the first insight you mention: resolution down-to-the-pixel doesn’t account for overall picture sharpness as perceived by the viewer. But you could also have mentioned those 2 other priceless insights:
- The explanation for the iris ‘sweet spot’: severe MTF drop (= contrast halos and chromatic aberrations) at full open iris, diffraction-limited blur at closed iris.
- Focus shift, which moaners would seem to ordinarily call ‘back focus issues’ but is apparently just a natural limitation of cheap zoom lenses - which you need to be aware and work around.

Posted by Stephan  on  05/01  at  02:50 AM

---

I like the Galt/Thorpe presentation; just be aware that it definitely comes from a certain viewpoint, grin.

The sweet spot on small HD cameras is quite an issue, as visible diffraction losses can appear at f4 - f4.8 or thereabouts (with full-res 1/3” sensors).

Focus shift is *always* present even on top-end true zooms (the focal point plot describes a gentle curve, not a flat, straight line). The “cheap zoom lenses” aren’t true zooms, but varifocals, and consistent back focus is achieved through computer servo control of the focus groups as the zoom groups are moved. Almost all built-in lenses are of this sort. From the user’s perspective, though, the technology is irrelevant: he or she just wants focus to hold as the zoom is used.  Either type of lens—true zoom or computerized varifocal—should be able to hold back focus within acceptable tolerances, so the moaners are justified in getting the problem addressed when it occurs.

Posted by Adam Wilt  on  05/04  at  06:39 PM

---

Lets see… Pana 170/200A has less pixels, its less sharp and its lower in resolution than EX1. Generally, reviewers think that EX1 is better: Better lens, larger chip, more pixels (no pixel-shift), and 10 bit contributes… though I am not entirely sold on EX1’s 10 bits nor am I on some aspects of its rendered image… may be it needs a serious SW update.

But I agree, you can’t just count pixels and claim that higher is better.

Regards,
Jay Bala.

Posted by .(JavaScript must be enabled to view this email address)  on  06/10  at  08:08 AM

---

Name:

Email:

Location:

URL:

Smileys

Remember my personal information

Notify me of follow-up comments?

Submit the word you see below: