Disturbing test image on D-70: yellow edges

[haris]

I was trying to replicate characteristic curves from

www.photogenetics.com on my D-70 and noticed a very strange thing in

the images from +3 1/3 to +4 exposure compensation. I was shooting a

gray binder back through the whole Exp Compensation range and the

edges of the mostly white images in this compensation range were dirty

yellow. I're not talking about minor discolorations--up to 1/10 of

the lower-right corner was affected as were the bottom and the right

side of the image. Barely noticeable in the top-left.

 

I ran a couple different trials with the kit lense at 70 mm and

wide-open aperture. Matrix and spot metering yielded the same result.

Using +.3EV custom curve from photogenetic changed the output

lightness but did not make a difference in terms of this strange

yellowing (looks like an old photograph). BTW, light fall-off to the

corners was very noticeable but the area affected by the discoloration

was larger than the apparent light fall-off. Most pronounced at +3.5.

I was using Fine JPG quality Small image size.

 

Anybody has seen the same? Ideas what this could be? Please try to

shoot an evenly lit mid-toned object with +3.5 exposure or thereabouts

and let's compare notes. Help me diagnose if this is my camera

problem or a larger issue for all. I'll try the same with different

image quality settings and in raw and post the results.

[craig_bridge]

Try raw mode and see what you get. The conversion from 12 bit channels to 8 bit channels can do some strange color shifts when a channel is at max. You need the raw channel values to make any sense of what is really happening! If the red and green channels are maximum and blue channel isn't in the affected area, it is going to be yellow! Over exposure color shifts are "normal digital film" behaviors and the white balance has a significant effect.

[frank_skomial]

You use crappy curves from www.photogenetics.com, then you do a +4 exposure compensation, and on the top of it you shoot Small image size JPEG. What do you expect then?

[kenneth_logan]

A useful test could be with a non-zoom lens at f/8 and no more than +2EV exposure compensation, with no third-party curve applied. When would you ever want to take a mid-toned subject area up +4EV?

[haris]

Here's the image I was describing<div></div>

[haris]

I don't believe the lense could be to blame because I was shootint a card that was about a foot away and set the focus to infinity. The images were way blurry!

 

As for +3 values: I really cannot think of a situation when you'd want to set the compensation that high, but values in that range will occur in pictures all the time. The camera has no idea how bright the subject is. It simply measures light and then sets the exposure accordingly. Exposure compensation should not interfere with image processing, it simply biases the meter and changes the exposure time. In the case of the image posted above, the exposure time is 3 seconds.

 

The image was shot with regular Matrix Metering and no Custom Curves. This is the worst occurrence of the yellowing problem of all the shots I took.

[rustys pics]

Looks like vignetting. You've got the lens at infinity and you're shooting a white card one

foot away with the aperture wde open? Why would you ever do that? What real world

shooting situation are you trying to replicate? The D70 has its faults, but this seems trivial.

 

See this blog by the infamous Ken Rockwell:

http://www.kenrockwell.com/tech/7.htm#measurebator

[haris]

I am surprised no one is concerned about this. If you do photography with a white background involved (snow), you probably will be setting exposure compensation. Probably not as high as +3.5EV, but the ugly yellows was the case in <b>MY</b> camera. Yours maybe different and worse. I do not think it's measurebating if I want to make sure the camera can handle a situation before I take a shot.<p>

 

Will someone try exposing a monotone subject in a series of shots with +2EV-+4EV range and post the outcome?