FX1 and color fringing of the object edges

[Alex Raskin]

What's the latest word on why does that occur and how to avoid this?

I only bumped into this situation twice so far, but it's very annoying from the otherwise great camera.

In one case, it was a black gun against grey ceiling tiles (shot from below the gun) - and the front circle of the gun's body around the barrel had a green fringe on the right. Shot under fluorescent office lights.

In other case, it was a telephoto shot of a tall brown building against the blue sky. The building had a very pronounced magenta border along the left vertical edge and an equally disgusting green edge on the right.

In the 1st case I noticed that fringing INCREASES if the object is slightly out of focus. Since 2nd case involved a hard-to-focus telephoto shot, I can speculate it could be the case there also.

Ideas?

[Radek Svoboda]

There is nothing you can about it. Lens has a chromatic aberation. All cameras in this price range with fixed lenses have it.

Radek

[Thomas Smet]

This can also sometimes be caused by pixel shift which is why you see green on one side and purple on the other. When you shift the green CCD by half a pixel and you have a sharp color edge one half of the green CCD will show up on that edge giving you a green fringe. On the oposite side there is a lack of green data from the shift so all you get is a blend of the red and blue CCD giving you a purple fringe on that sharp egde.

This can also be caused by other problems such as low quality lens, edge enhancement, and bad chroma sub-sampling.

[Radek Svoboda]

The Italian guy put on FX1E Fuji lens, color fringing was gone.

Radek

[Dan Diaconu]

Alex (and all other Z1/FX1 or other camcrders owners)
Do the following test (before blaming the lens again):
1.frame a white cup against a white wall and observe if there is any color fringe.
(notice the iris setting)
2. frame a darker color cup against the same wall (at the same aperture setting as in the first case). Now (for the second case only): open the iris till you get reasonable details on the darker cup. Question: did the color fringe increase as well?
I am 99.9999999% sure the offender is NOT the lens, but the CCD's limited dynamic range. Check the above in a room on any EVEN lit scene and then turn on a bright light. ALL shinny edges of all objects WILL show a color fringe.
Check all hight contrast areas (mirror reflections) and decide:
http://dandiaconu.com/gallery/album05/IMGA0412
http://dandiaconu.com/gallery/album05/IMGA0402
Adding a Nikkor and a GG should have made matters worse adding even more distorsions and so on, right? But it did not. It only reduced the contrast ratio in a "range" that CCD's "feel" more confortable with. My 2c.

[Barry Green]

The color fringing is chromatic aberration in the lens.

You can see this by going to the eidomedia site, where they took the stock lens off and replaced it with a Funinon lens. With the stock lens, there's purple/green fringe. With the Fujinon glass, no fringe.

Nothing else changed -- just the lens. So it's not exposure, it's not pixel shift, it's chromatic aberration in the glass.

http://www.eidomedia.com/hdve/ziess_fuji.htm

[Dan Diaconu]

Originally Posted by Barry Green

The color fringing is chromatic aberration in the lens.

Could be, but.... that does not explain the absence thereof when the contrast ratio of a scene decreases....(using the same lens!)
(I saw those images quite a while ago Barry and they seem to point in the same Hi Contrast zone of the pictures, thus confirming my wilde guess; check again the roof detail and notice the absence of fringe on the roof line just under the dark "whatever-that-is" section. Is the only place where the roof line is brown, the rest being purple) My 2c again. As for the lens..... it could be Zeiss sharper than Fujinon ?(just enough to become offensive....)or.......????? a "deadly combination" of both... (or even a third something we do not know....)

[Jimmy McKenzie]

... and I haven't been happier since I dumped my still cam with it's highly touted Ziess lens for a DSLR for just a few bucks more.
Sometimes contrast is part of what I do, and to see the unacceptable green and purple destroy nearly every shot was the tipping point for me.

[Thomas Smet]

Thanks Radek and Barry. I have never looked at that test. I agree the amount of fringing shown there isn't from the pixel shift. Pixel shift can cause this problem but not that much! I was talking a very slight half of a pixel fringe that you can only really see in certain conditions and zooming into the image. The effect shown in those images is much worse than what the pixel shift does.

I actually am writing a program to test out fake pixel shifting to enlarge images and that is where I first noticed the green and purple fringing.

[John Jay]

when you look at the lens with the lens hood removed you can see the reflection of say a window frame with purple and green fringing caused by the coating on the first element (glass ornamental). This is the relection in the coating and must be filtering purple and green light from high contrast objects leading to the fringing that we are observing

[Radek Svoboda]

I saw images made with HDCAM with HD lens, had color fringing too.

Radek