EOS 20D Color Fringing & Soft Images

Recently I hired an Architectural Photographer to shoot one of my

projects. Our agreement was to provide me with a file for my use at

high resolution.

 

I received his RAW file as well as a JPEG file. Both files contain a

fairly significant amount of color fringing at high contrast areas.

For example, the edge of a building has a red fringe line on the

right corner and a blue fringe line of the left.

 

My queation is; could these be abberations due to the lens optic or

are they digital artifacts of the 20D. The RAW file (23mb)seems to

have less than the JPEG (13mb) but neither file is good enough for my

purposes.

 

Although the photographer claims the images are tack sharp, they do

not even come close to what I am used to receiving. Zooming in on

different areas of the image show soft details and significant color

fringing.

 

Is this typical of the 20D, could it be the camera or the lenses?

 

I don;t want to hurt this guys feeling, but I really can not use his

work.

 

Regards,

 

Michael

Architect

Sounds like Cromatic Aberation caused by the lens. Is this fringing less near the center of the image? CA will generally show like that. Some digital sensors do create blooming & fringing on there own, but the 20D sensor is pretty good about that. Anyway, CA can be greatly reduced with a decent image editor. Photoshop should do it, as well as Paint Shop Pro, etc.

Their I go again -- putting the wrong "there" in they're.

A 20D RAW file needs quite a bit of post-processing to look its best. It definitely needs sharpening and levels or curves adjustments along with possibly some saturation. What lens did he use? Some Canon lenses cause more fringing than others.

Many of these problems can be solved with a few seconds of manipulation, and in fact are

also present on film. The difference is that one rarely blows up film to the same degree as

digital.

You should have hired a professional photographer. I can't imagine anyone trying to use a 20D for serious architectural work.

<I>For example, the edge of a building has a red fringe line on the right corner and a blue fringe line of the left.</I><P>

 

If it's red on one side and blue on the other it's definitely chromatic aberration. That's C.A.'s classic signature since C.A. is defined as not bringing all wavelengths of light to focus at the same point.<P>

 

That said, the 20D, or its software, ALSO has serious problems with color fringing in high-contrast transitions, especially if the transition also involves different colors. I recently posted some images here of a transition between a dark blue subject backlit witj bright orange that absolutely gave the 20D <B>major conniptions</B>!<P>

 

Also, the 20D seems like a <B>very</B> odd choice for architectural photography! Where did you find this "architectural photographer"?

Sounds like crappy optics combined with the 20D's absurdly soft sensor. 35mm is my lower limit on my 10D and 20D's I've used because I haven't found any Canon glass that can go lower without these kinds of problems. A 1Ds will make life easier due to it's larger sensor, but Peter is against that kind of solution for some stupid reason. I guess making his 20D look bad, which isn't that hard.

 

You will not find fringing on my 10D 50mm shots, and I frequently use it for high key studio shots of chrome parts. 20D not the best choice for architectural work anyways. Ask the Photog to shoot it right, or get a refund and find another with a digital back or better camera.

Do the RAW conversion in CS2 Camera RAW to take care of the CA. Also, it sounds like the files need some Unsharp Mask. That's just the nature of digital. My 20D files are already a lot sharper than what my 10D gave me, but for optimum sharpenss, you still need some sharpening. I use Photokit Sharpner, which gives tons of control over sharpening and offers great results: http://www.pixelgenius.com/sharpener/. Fred Miranda's 20D CSPro (Custom Sharpening for 20D) is also very popular: http://www.fredmiranda.com/shopping/20DCSpro. At only $19.90, I would give 20D CSPro a shot. You'll probably be stunned at how much sharper those 20D pictures are going to get after running 20D CSPro!

<I>Do the RAW conversion in CS2 Camera RAW to take care of the CA.</I><P>

 

How does it fix CA in software? Does it have different profiles for different lenses?

Bob, I wasn't aware that the 20D wasn't up to shooting architecture.

 

In fact, I'm sure it is.

<I>Bob, I wasn't aware that the 20D wasn't up to shooting architecture.<BR>

In fact, I'm sure it is.</I><P>

Obviously you <B>could</B> shoot any subject with any camera. I'm sure there have been a few weddings shot with Argus C3's rangefinders or Instamatics over the years.<P>

 

But why on earth would someone who bills himself as an architctural photographer CHOOSE a 20D? Its 1.6 crop factor limits its wide angle ability to a few zooms. And its small sensor size means there's no point in using a PC lens. And its small dark viewfinder would make it hard to compose many common architectural photography shots.<P>

 

Most architectural photography is done with medium format or 4x5 cameras because it's a lot easier to frame the subject and correct for perspective distortion, and there's a wide choice of PC lenses available (or you don't need them at all if your camera has a tilt-bellows), and the larger format allows you stop WAY down without diffraction issues for use with PC, and because those cameras have much brighter viewfinders, it's easier to compose common architecural subjects.

http://www.pbase.com/parr/pfringe

Michel I found this test at the site above that apparently shows thet PF doesn't have to do with blooming but with the lens

Ron Parr's experiments do debunk blooming (in the sense of electrons flowing from overexposed pixels to adjoining ones) as a cause of PF. However, the angles at which rays are focussed onto a particular point on the sensor is not a constant from one lens to another, even when the same focal length and aperture are in use. I venture to suggest that sensor optics play an important part in the equation - easily verified by comparing the same lens on different bodies, including film ones. Lens based CA should show up on any medium capable of recording it - PF is a different phenomenon.

Chromatic Aberation can show up in ANY DSLR if the circumstances are prime for it. I've seen it on my 20D with a 50mm lens, and a 300mm lens. Fortunately it is very easy to fix in Camera Raw. Two sliders and the ALT key let's you remove it the right way by shifting the colors to align more appropriately. This is the nature of CA. The color convergence gets a little misaligned through the optics of the lens and are captured by the sensor this way. The data is there it's just not lined up properly. When using the CA adjustment tool within Camera raw, you'll actually see the image shift a little as the color fringing disappears.