Red edges (relatively)

Something I have been noticing recently (but presumably it was always the case) - the left edges of my pictures come up red when I'm trying to balance levels in shots which are too green or blue. (Of course this very rarely happens ). To avoid these red edges I then have to back off from what I think would otherwise be the optimal set of adjsutments to the levels.

At first I thought it's just a case of there being a light gradient across the picture, but almost always I find red tinges on the LEFT edge when I try this, and sometimes red tinges all the way around.

Here's an example, where the effect has been highlighted by using the "auto levels" in Photoshop.

I seem to remember reading somewhere that the sensitivity of the CCD falls off quickly as the angle at which the light strikes it increases. So is this just an effect of a combination of green & blue light falling off at the edges of the shot (where the angle is greatest) while the red, being very noisy and then highly amplified by the "auto levels" process, is relatively less affected by the change in angle?

If the effect is most noticeable on one side, does it mean the CCD is not positioned centrally with respect to the lens?

Any thoughts?

Perhaps the camera lens is not square with respect to the port? Hard to see that on the PT-015.

Interesting ... I think you're right, but I don't see why that should produce this effect.

I think you're right because I actually have a modification to the housing which I use sometimes, and I can't find any examples of the effect for dives where I remember that I didn't use the modification.

The modification consists of a replacement for the rubber boot which surrounds the LCD screen. Sometimes I replace it with a smaller piece which doesn't go across the top of the screen, so I can put the camera down very low and still see the screen from above for those awkward camera angles.

But recently I noticed that the camera doesn't in fact sit exactly square to the housing with that system, so I didn't use it for the last 30 or so dives (ie my last week at Wakatobi). In none of those dives can I find a picture with this effect.

Why does not being square to the lens port produce this effect? Does it result in a difference in the fall-off of CCD response on the two sides of the sensor? I can't quite visualise it, but it seems reasonable that that could happen, and it wouldn't be the same on the two sides. But shouldn't the effect go one way on one side (emphasise the red) and the other way on the other side (emphasise the blue)? That doesn't seem to be what was happening ... hmmm. :freak:

Well that's one topic for us to mull over in Lembeh whilst sipping coconut milk laced with Bacardi!

OK, but then what do we ponder on the second glass?

Also Dave,

It serves you right for getting in that position

Sorry I'm late responding. There's a explanation. Otherwise I wouldn't have suspected!

What you are seeing is chromatic aberration due to the port not be parallel to the lens surface. Imagine light entering the lens directly down the center. That light should enter the lens perpendicular (normal) to the plane of the lens. Underwater, that same light must pass through the port the same way. If it does not, there will be a spreading of the color spectrum just like in a prism. Moving off-axis in one direction the effect will cancel out natural CA within the lens and will amplify it in the other direction. So... what you are seeing is essentially exaggerated CA.

Thanks Craig - Now what are we going to talk about

OK, but then what do we ponder on the second glass?

Carefull guys.

They say that drinking too much coconut milk can have a diahretic effect on you.

**Better dilute with more rum**

I think it goes like this ...

1st drink: Why is there a red fringe at the left?

2nd drink: If the red fringe is on the left, does that mean the camera is pointing too far to the left or too far to the right?

3rd drink: Given that there's almost no red light down there without the flash, in what sense can we say what amount of red is correct in the picture anyway?

4th drink: Why are there two red fringes now?

5th drink: If the problem is that the camera's mounted in the housing wrong, then maybe I'd better not use the housing ...

6th drink: I notice the red fringes at the end of the day. I see red clouds at sunset which is also at the end of the day. Is there some cosmic connection between these phenomena?

-David