GNOME Bugzilla – Bug 319872
Contrast adjustment incorrect for white and black
Last modified: 2008-01-15 13:01:44 UTC
Please describe the problem: White (255, 255, 255) and black (0, 0, 0) pixels don't smoothly change color when the contrast is lowered. But when the slider hits -127, they jump to the correct color. Steps to reproduce: 1. Create a new white image 2. Select Layer|Colors|Brightness-Contrast... 3. Slide the contrast slider to -126 At this point the image should be a middle gray (about 130,130,130), but it's still white. 4. Slide the contrast slider to -127 Now the image will jump to the middle gray 5. Undo the operation to return to a white image 6. Select Layer|Colors|Invert to make a black image 7. Select Layer|Colors|Brightness-Contrast... 8. Slide the contrast slider to -126 At this point the image should be a middle gray (about 126,126,126), but it's still black. 9. Slide the contrast slider to -127 Now the image will jump to the middle gray. Actual results: The white and black remain so until the slider reaches -127, when they jump to the middle gray. Expected results: The white should get progressively darker as the slider is moved to the left. The black should get progressively lighter as the slider is moved to the left. Near white (254, 254, 254) and near black (1, 1, 1) appear to work correctly. Does this happen every time? Yes Other information: Win32 version. Intel Pentium M 1.86 MHz processor.
The tools is intentionally designed to leave black and white pixels intact while shifting intermediate values toward gray. The behavior for contrast=-127 is an anomaly. Your expectations are reasonable but I don't think there should be changes in a tool that has existed since the beginning of time and always worked the same way. (Note that you can get the effect you want, with a bit more effort, using the Curves tool.)
How about a checkbox to enable the expected (similar to other image processing programs) behavior?
Actually this is a real bug. A more extensive discussion of the problem is in bug #332068, hence I am marking this one as duplicate. *** This bug has been marked as a duplicate of 332068 ***