Wednesday, February 13, 2008

Thoughts on fringing

I took some pictures last night of the ice stuff, then drove my car around an icy parking lot like a madman. It was fun, but I left all of that at home because of the water. Big mistake, there was a really cool "tree laden with ice that's assaulting the sidewalk" thing going on outside. Oh well.

I'm writing this, though, to muse about "purple fringing" as an image aberration. Someone posted a comparison of the Voigtlander (sp?) 40/2 and the Pentax DA 40/2.8, which garnered a response from someone who had compared the lens to both the DA 40 and the FA 43 (which I have). Their conclusions were that the Pentax lenses both produced more purple fringing.

This sticks out because I've never noticed the 43 to produce purple fringing - it has standard lateral chromatic aberration, but it is relatively well behaved until I get into such high contrast situations that sensor blooming takes over the boundary regions.

Then it struck me - people don't understand the difference between lens aberrations and digital sensor woes.

Why would this have hit me? Simple - the wikipedia article makes it apparent that the internet believes that purple fringing is more of an issue for digital photography than film. If what we were seeing was due to lens errors, it would be just as present on film exposures in the same conditions as the digital exposures with the same lenses.

I've looked at some of my shots with film that I still have online and am not seeing anything in that nasty purple range at all in my images, let alone at high-contrast boundaries. Even in a shot with the DA 10-17 on my MX, I don't see what could be described as purple fringing on the right side - the trees simply start disappearing into the sunlight. There's another thing that leads me to believe it's sensor related more than anything else: it's the same color as the nasty artifacts that come off to the north, south, east, and west of the sun in images like this one. Fortunately, in that case, most other people don't notice the problem, but it screams to me "DIGITAL! DIGITAL!"

I don't know, I'm thinking it might be fun to set up a test with some lenses "known" to cause purple fringing and shoot the same images on digital and on film and see if you get the same results.

Basically, I don't buy purple fringing as a strictly optical phenomenon; I believe it's the medium that makes or breaks it.

4 comments:

  1. Nat,

    I look at it as a dual problem. Some lenses are a source of purple fringing way more than others. My 18-55 kit lens produced substantially more purple fringing than my 17-40/4L is similar conditions.

    But... Purple fringing does seem to be much more common on digital than film. On film, it looks like the lens has to be a really bad fringing offender before the problem shows up. Even great lenses exhibit fringing in extreme conditions, but as a matter of degree, something about digital sensors seems to lower the bar and let more fringing show up.

    - Josh

    ReplyDelete
  2. I think it's a contrast issue - lenses that have really high contrast are the ones that produce the problem. At least, that's how it should work. Of course, we don't want to lower contrast in response, so it's probably one of those delicate balance things.

    ReplyDelete
  3. I can post as my AIM name?

    Good lord, the internet is stupid.

    ReplyDelete