1. Reminder: Please use our affiliate links for holiday shopping!

Interesting: Sensor stack affecting quality of adapted lenses immensely

Discussion in 'Open Discussion' started by Poki, Jun 7, 2014.

  1. Poki

    Poki TalkEmount Hall of Famer

    Aug 30, 2011
    Austria
    Roger Cicala, known for his service lensrentals.com, just came up with a very interesting theory (prove of it included in the article).

    It basically states that if you adapt a lens from another system to your camera with an adapter which doesn't use glass (or uses glass which isn't there for correcting the difference in sensor stack thickness) you'll loose a huge portion of the image quality the lens is capable of. So if you shoot with adapted lenses, it might be a good idea to invest into an adapter which corrects for this, although they are expensive.

    Source article: http://petapixel.com/2014/06/07/glass-path-using-adapters-may-hurt-image-quality/
     
    • Like Like x 4
  2. rdfisch

    rdfisch TalkEmount Regular

    181
    Nov 13, 2013
    Northern NJ
    Rick
    A fascinating read ... thanks for the link. For me it raises many questions.

    I wonder if there are variations in the sensor stack thickness on the sensors within the Sony E-mount line ? If so then it would seem that E-mount lenses might be at a disadvantage depending on which Sony body they were mounted on. I look forward to the author's upcoming database of thicknesses.

    I have always assumed that when a company (for example Zeiss), makes a lens for different systems (for example the Touit line for Fuji-X or Sony-E) that the only difference was in the mechanical flange, but if the findings in the article are significant then this may not be the case, or sufficient. I would be curious to know the relative sensor stack thickness between Fuji & Sony, and if much different, are the Touits different beyond the mechanical mate.

    Fun theoretical stuff, but the most important question of all: "are the measurements significant in the real world ?". While the author refers to some of the data as "awful", this forum certainly has many examples of excellent results using legacy lenses with glassless adapters ... would they have been much better when used on the camera system they were designed for ? Do the one's shot wide open with a fast lens suffer from "huge ... spherical aberrations, which are apparent even in the center of the lens’ field" ?
     
  3. Poki

    Poki TalkEmount Hall of Famer

    Aug 30, 2011
    Austria
    No, there isn't. The layout and thickness of the sensor stack is part of the E-Mount specification. Even the A7r, which lacks an AA-filter, features the same amount of glass in front of the sensor due to a piece of clear optical glass.

    The overall optical design is exactly the same. And as the X-mount is comparable to the E-mount in pretty much every way I don't think there's a big difference in the sensor stacks. But even if there is, when you know the specs while developing the lens, you can correct quite easily for it by just aligning the focusing group a little different.

    Don't forget that Roger tested the Zeiss Otus, a lens much better than any legacy lens out there. So a 20% drop in image quality is actually quite significant, while the same drop in a Minolta lens might not even be noticeable by the user, and even in direct comparisons you could assume it's just sample variation. I am quite confident though that you could see the difference without magnifying an image on most of the better legacy lenses.
     
    • Like Like x 1
  4. rdfisch

    rdfisch TalkEmount Regular

    181
    Nov 13, 2013
    Northern NJ
    Rick
    Thanks for the answers.

    One more question ... Can the types of aberrations we are talking about be remedied by an appropriate lens profile in lightroom, or is it "too late" to fix in PP ?
     
  5. Poki

    Poki TalkEmount Hall of Famer

    Aug 30, 2011
    Austria
    Due to the spherical aberration, you loose a lot of resolution and contrast, and as we all know, you won't be able to actually gain true resolution with any software. Also you might loose some color gradations due to the degraded contrast which you also won't be able to recover with software.
     
  6. NickCyprus

    NickCyprus Super Moderator

    Oct 11, 2012
    Cyprus
    Nick
    Interesting theory...that could put off all legacy lens users out there.
    But I'm sure all of us have seen the great results over time than can be achieved with adapted legacy lenses on cheap (no glass) "dump" adapters :)

    Don't understand much of this technical stuff but personally I'm taking this article with a grain of salt...
    What's an example of the appropriate adapter that could be used out of curiosity?
     
  7. xXx1

    xXx1 TalkEmount All-Pro

    Jan 15, 2013
    That may be part of reason why some adapted lenses work very well and others don't. I have adapted method of testing the lenses in practise and not reading tests. Seems to be working, I have got some pretty nice lenses that way.
     
  8. addieleman

    addieleman Passionate amateur

    Nov 13, 2012
    Netherlands
    Ad Dieleman
    Very interesting findings in the article. However, to me the only possible approach still is simply testing if I like a lens on my camera or not. At least now we have an explanation why so few older lenses do well on digital cameras at large apertures.
     
    • Like Like x 2
  9. Poki

    Poki TalkEmount Hall of Famer

    Aug 30, 2011
    Austria
    Given that all old lenses were developed for film without anything in-between the rear element and the film plane, testing a lens on the camera you want to use it sure is the only good way to see how it performs. And even though you can correct for the spherical aberration with optical adapters, there could be many more problems like shifting colors etc. Also, let's not forget that additional layers of glass always come at a price - additional internal reflections, degraded contrast and resolution and less light throughput. So unfortunately, legacy lenses will never be as good on digital sensors as they were in film, although you can get pretty close using the right combination.
     
  10. Damovich

    Damovich TalkEmount Regular

    25
    Jun 6, 2014
    I also thought of it as an interesting read though I cant help but wonder about the promotional value this article incorporates. While I have no doubt the author and the Metabones inventor to have a vast amount of knowledge concerning optics the general impression left by the article, to me personally, was that it is a clever piece of advertising promoting at the least newer versions of Metabones focal reducers (the designation "perfect focal reducer" gave it away I guess).

    (interesting on another level)
    In the comments beneith the article there was a somewhat interesting discussion going mainly kickstarted by one unscientific user stating that every piece of glass being put into the equation reduces the amount of light reaching the sensor, while other users sorta claim such a statement to be "unscientific/uneducated" while their own reasoning also lack any real scientific arguments. To me it only indicates that one can never be to sure about the true outcome of such test(s) unless executed by an indepent reviewer and which the lensrental owner clearly is not hence his connection to the Metabones inventor.
     
    • Like Like x 3
  11. ilovehatephotography

    ilovehatephotography TalkEmount Regular

    153
    May 30, 2014
    Los Angeles, CA
    Ed
    This is what happens when you own a lens rental business, with all this optical testing equipment on hand, coupled with a need to market your business. Pixel peeping to the extreme. :)


    Sent from my iPhone using TalkEmount
     
  12. Poki

    Poki TalkEmount Hall of Famer

    Aug 30, 2011
    Austria
    Of course there is promotional value in there. But keep in mind that the post is written by Roger Cicala, who is well known for more or less scientific analysis about different photography related topics. Also, his measurements make sense for sure.

    Also, of course you loose light with every additional layer of glass. The amount of light lost is another topic. I read a book written by Zeiss' renowned Dr. Nasse which also covered this topic. He basically said the light loss is very small due to modern aspherical designs and lens coatings. So with an adapter with, say, three lens elements, you probably loose somewhere around 0.2 T-stops of light.
     
  13. Poki

    Poki TalkEmount Hall of Famer

    Aug 30, 2011
    Austria
    Looking at the MTF graphs, you can clearly see that this is NOT pixel peeping. Well, it may be technically, but looking at the results, you should be clearly able to see the difference without even magnifying the image.
     
  14. ilovehatephotography

    ilovehatephotography TalkEmount Regular

    153
    May 30, 2014
    Los Angeles, CA
    Ed
    It may very well be that the maximum optical potential of the lens is not full realized after having been adapted to another camera, but I wouldn't worry about it. I've enjoyed my adapted lenses on all my cameras and have captured beautiful images with them.
     
    • Like Like x 4
  15. rdfisch

    rdfisch TalkEmount Regular

    181
    Nov 13, 2013
    Northern NJ
    Rick
    If by adapted lenses we mean lenses originally intended for film cameras, these were all designed, as Poki has pointed out, for use with nothing between the lens and the focal plane. Considering this, I would think that they are all on a level playing field and I don't see how the difference in results (when used with the same sensor) could be attributed to this phenomenon. I would also think that, in this regard, the thinner the sensor stack, the better.
     
  16. MAubrey

    MAubrey TalkEmount Top Veteran

    If we're going to call it a 'theory' then I think we need to accept it as 'established theory'--a something that has already been empirically tested and fits the current evidence such that we should accept it, notwithstanding the introduction of new data. Or to put it another way, this isn't just Roger's theory. This is Roger's discovering and playing around with something that optical designers like Brian Cadwell already know and work with every day.

    As for the question of legacy lenses, Brian Cadwell answered that question in the comments: The more telecentric a lens is, the less the optical stack matters. So telephoto lenses, retrofocus wide angles should be be be giving up essentially the same performance or better as they would have on a film camera. And we've already known that for some time now. He also says that lens designers probably started taking these questions into consideration in the mid to late 90's, so that whether the lens was used on film or digital, it would work equally well.
     
    • Like Like x 3