PHOTOOG Photography writings by Olivier Giroux

13Jan/098

MTF in Depth

I want to point out this excellent background document on MTF, written by a Dr. Nasse from Zeiss.  The document is full of rich information, you really should set aside some time to read it.  The introduction on the point spread function was a very nice way to start.

I won't summarize the document, I think it's already very dense as it is, but I did find two surprises inside which I'll relay to you.

Surprise #1: wide-spectrum MTF versus monochromatic MTF

mtf3002

The two charts above pertain to the same lens, a 300mm telephoto in this case.  The chart on the right looks only at monochromatic green light, while the one on the left uses the spectrum of white light.  Among other things, this makes comparing Zeiss MTFs with other manufacturers very difficult.  Imagine the case where two lenses appear to be close competitors but their MTFs use different light spectra, one of them would be far superior to the other.

Surprise #2: wavelength-dependent MTF

mtf85

This chart says two things.  The first one is that the distance to the projected plane of focus is wavelength-dependent, i.e. the red image doesn't form at the same distance as the green image.  This is normal for all lenses that aren't apochromatic (APO).  I knew this, and it is a concept I observe in daily shooting with fast telephotos.

The second thing is that the contrast of the lens itself is dependent on the wavelength, and the contrast of white light (the black line) follows the lowest common denominator.  This is the same as saying that if you could find a B&W sensor camera and mount a green filter in front of it, then the useful resolution of the lens could basically double with all else being equal.

Comments (8) Trackbacks (0)
  1. I’ve wanted to send you a link to the paper a while ago, but I’ve just forgotten..

    As a side note, I really want to see zeiss making some apo’s…

  2. I also want to see Zeiss ship some APOs in the ZF/ZK/ZE line. They currently produce them for Hasselblad and they have a number of APO designs for SLRs sitting in drawers from the Contax system.

    One issue I can see is that APOs are typically telephotos (where longitudinal CA is more of an issue) and the competition in that space from Nikon/Canon is really fierce. Auto-focus & image-stabilization really matter for telephotos, and not having them puts a real damper on enthusiasm.

    They had an opportunity to make the 100mm f/2 an APO, which they chose not to take. The next opportunities will be 200mm/300mm lenses if they add some to the line.

    I think I would like to see a conservative lens spec to keep the price down and mitigate the importance of AF and VR… maybe something like a 200mm f/2.8-3.5 APO or a compact 300mm f/4-4.5 APO. Slower is fine if the quality is fantastic wide-open and the lens is made more compact.

    I don’t want to see things like a 200mm f/2 or 300mm f/2.8 priced into the stratosphere to serve no purpose other than glory.

  3. I’d rather like to see a ZF 50/f1.4 APO chiped and recover that lost crown to leica summilux 50… Ah, that’s a dream..

  4. Oh ABSOLUTELY! That would be marvelous. I would wait in line for that.

    This said, there already are 2x 50mm lenses in this line-up and I don’t think enough people are cognizant of the differences to pressure the market to make a response to Leica. It’s looking like the 2/50 ZF is the best option there for the foreseeable future.

  5. By the way, did you take any decision between 2/50 ZF and sigma ? 🙂 I guess you need a 50 now that you got d700

  6. I haven’t and it torments me quite a bit.

    I think I would get the 2/50 ZF if I had the spare money lying around.

  7. I must say that it amazes me that we haven’t seen a monochrome body yet. Take a modern sensor, originally designed for Bayer, and make it without the coloured filters. This would:

    – give maybe 1.5x the resolving power
    – have a higher base ISO (less light loss to the filters

    On the downside, less flexibility in post. But hey, suddenly orange and green filters might make a comeback.

    Dragging this idle fantasy back to APO lenses, have you tried the Voigtländer SL lenses? (90/3.5, 125/2.5, 180/4)?

  8. Thanks for the comment Chris.

    There was one camera from Kodak which was built around a digital B&W sensor. I believe it served 6MP images, and it resolved way more detail than its contemporary 6MP color sensors. The product was a flop unfortunately, for a variety of reasons, and the rest of the industry is probably too scared to repeat the experiment.

    It looks like the industry is going through a “maturity regression”. It used to be a highly-diversified environment with products for non-overlapping audiences. The switch to digital re-set the environment so all but one camera genus survived -the one-size-fits-all design that updates every XX months.

    I think we’ll see maturity come back, and diversity flourish again. I think it’s approaching. I would be fine with D700-level pixel quality for a long time.

    When this much quality comes to the 1000$ price tag, what is the reason why not build niche cameras that appeal to 1% of the market? This should be around the time when we’ll see a Zeiss DRF for instance, they’ve basically said so themselves already.

    For some reason I expect Voigtlander to be part of a digital B&W come-back… :^)

    Which segways into the next topic. No, I haven’t put my grubby hands on a Voigtlander SL. I am sure we will see them re-launched very soon… if not at PMA this year, then for sure at PMA/Photokina next year.

    The APO SLs are apparently _really_ great. It’ll be a treat.


Leave a comment

No trackbacks yet.