Page 1 of 1

Edge bleeding/ghosting quest

Posted: January 23rd, 2014, 10:02 am
by tomczak
What follows has little to do photography (as Jonathan recently pointed out), but I've been struggling with it for a while now and became scientifically obsessed as to the reasons for it.

In short, I found that dpReview's new comparison tool in low light has one element (the metal brush ferrule in the LHS upper corner) that mimics the condition: an overexposed (in one or more channels) edge spills over the dark background. While I had to look for it in the dpReview well lit scene, it happens often outdoors, and can make viewers angry (it makes me angry...).

Please take a look, switch cameras: some show it much more than others, RAW or JPG. Dpreview comes as close to standardizing everything else (lenses, aperture) other than the sensor, as they come. Small sensor have it more than larger ones, but there is no clear pattern (look for instance at Fuji X-E2 vs. Fuji X-Pro1 - the first one has it, the second one much less, while the sensors are almost the same.

Beer for anyone who can explain it!

Cheers!

Re: Edge bleeding/ghosting quest

Posted: January 23rd, 2014, 1:29 pm
by den
I wonder if this DPReview comparison is not more of a test for Adobe Camera Raw's [ACR] ability to RAW convert a very small image area ("...the metal brush ferrule in the LHS upper corner...") where the dynamic range exceeds the sensor's range of a single exposure... ie, ... ACR's ability to set the black and white point blend/clip percentages for a variety of camera manufacturer raw file types...

I would think for the purposes of comparing 'flare' resolution that the manufacturers’ raw converters/editors should be used.

...den...

Re: Edge bleeding/ghosting quest

Posted: January 23rd, 2014, 2:23 pm
by jsachs
Could be caused by the antialiasing filter in front of the sensor or it could be blooming - a tendency of charge from bright areas to leak into adjacent pixels which is a sensor/electronics issue.