Canon Community Canon Community

Esteemed Contributor
Posts: 3,048
Registered: ‎06-11-2013

Re: Tamron Lense Distorts Colors

Chromatic Abberation (CA) occurs because the curved shape of the lens elements work like prisms in that they "bend" light rays.  But it turns out the amount of bending is based on the wavelength of the light.  


Short wavelength light (blues) bend much more than long wavelength light (reds) and this results in "dispersion" (seperation of colors ... basically "white" light will split into it's "rainbow" of colors.  But mostly what you'll see is that the colors are just starting to separate -- so you'll see a red fringe / blue fringe on the edges of things.


At the pin-point spot in the very center of the glass, light doesn't disperse... it passes straight through.  This means you should get no CA at that point.  The degree of CA should get worse as you get farther from the center.  


BTW, there is something called "axial" CA which is more difficult to correct.  "Transverse" CA is a bit easier to correct.


If you were to imagine that only one wavelength of light passes through the lens (say it's only "green" light) then you get a "green" image on the sensor at some focused distance.  If you also took only a "red" image and then only a "blue" image and compared them... what you'd find is that the three focused images are all technically slightly different sizes.  In other words if you stacked all three images on top of each other, you'd discover the blue image is just a tiny fraction smaller than the green image which is a tiny fraction smaller than the red image.    You can get them all to align at the center point... but the farther you are from the center (e.g. if you inspect the corners) then you clearly have three different sizes.


A computer can fix this simply by separating the color channels and resizing them back to the same size... then re-combine them.   


Several programs have features that let you correct CA and this is basically what they do.


But axial CA is more difficuult to correct because in that version of CA, the light focuses at different distances from the center.  So maybe the "green" light is focused but the red and blue wavelenght light are both a little out of focus.  Simply sliding them in or out a bit (resizing) isn't going to help because those channels are actually blurry (if you were to open the image in Photohsop, go ot the "channels" section and inspect each channel you'd notice that they aren't actually uniformly focused.)


Lens makers add extra lens elements into the image path which are meant to compensate for these problems.


The simplest lens just has a single convex lens element (like a simple mangifying glass or magnifying loupe).  If you look through a magnifying glass you'll notice the image distorts near the edges and looks best near the center.  


It was discovered that a second lens element with a concave front and nearly flat back could be stacked behind the object lens element (the normal convex lens) and this combination (called an "achormatic doublet" arrangement) causes some (but not all) of the CA to be corrected.   Adding more elements can help even more (apochromatic triplets, etc.)    Some lenses use "low dispersion" glass.  But these "glasses" are often made from exoctic elements ... such as florite crystals ... that have to be "grown" into large enough crystals (using a kiln and the process can take months) to be able to grind the crystal into a lens.  If the crystal is grown too fast, it'll have optical flaws in it.  It's a slow process (which is why many of these lenses are very expensive.)


Anyway... once upon a time, photographers shied away from "zoom" lenses for any serious photos because they were optically inferior to non-zoom lenses.  It's one thing to put in extra lenses to control the light at just one focal length.  But try to do this for a broad range of focal lengths ... that's harder.


Lens technology has come a long way and today's zoom lenses are capable of competing with non-zoom (aka "prime") lenses.  But there is a range of quality.  The best zooms -- the ones that do the best job optically -- aren't cheap.  There are cheap zooms... but their optical quality doesn't compete.


It also turns out that controlling these optical issues is a bit easier when the zoom range is more limited.


Most zooms tend to have a range where the narrow end has a focal length roughly 3x that of the wide end  (e.g. a 24-70mm is appoximately 3x (24 x 3 = 72).  An 18-55 is another example (18 x 3 = 54) or a 70-200 (70 x 3 = 210).  You get the idea.


And then you get to the "super-zooms"... such as your 16-300... where the narrow end is often 10x or more as compard to the wide end.  A 10x range would be 16-160... but your lens is 16-300 (300 ÷ 16 = 18.75).  That's even ambitious for a super-zoom.


Super-zooms are very ambitious w.r.t to the zoom range and they never compete (optically) with less ambitious zooms that tend to limit the zoom range to about 3x.


Lens selection is always a game of trade-offs.   The whole point of having a removeable lens camera is that no single lens is best at everything so instead of forcing you to choose... the camera is designed so you can change your mind at any time ... but attaching a different lens.


When you select a super-zoom, you're basically putting priority on "convenience" over optical quality (and also over many other factors).  


Tim Campbell
5D III, 5D IV, 60Da
powered by Lithium

LIKE US on Facebook FOLLOW US on Twitter WATCH US on YouTube CONNECT WITH US on Linkedin WATCH US on Vimeo FOLLOW US on Instagram SHOP CANON at the Canon Online Store
© Canon U.S.A., Inc.   |    Terms of Use   |    Privacy Statement