Histogram of full image (full301.353)
FROM: Mike Malin (malin@msss.com)
DATE: 19 April 1995
SUBJECT: This is the "Noise" Problem Based on an e-mail description Scott Murchie sent me yesterday, I went back and looked at some images from the first CCD. This brief memo illustrates the problem. Note that what is being called a "noise problem" really isn't "noise" so much as a loss of gray-level resolution (a quantization problem). We saw something like this in the MOC, which turned out to be a problem in the analog-to-digital converter.
Basically, the problem appears to be in the lower three bits: at low DN levels, these bits are fully occupied; at middle DN levels, the lower three bits are partially populated; and at high DN levels, they're severely under-populated or not populated at all.
Histogram of full image (full301.353)
Low DN portion of histogram, showing values every DN
Middle DN portion of histogram, showing multiple DN gaps (note the
primary spacing is 8 DN = 2^3, or the lower 3 bits of the 12-bit range)
High DN portion of histogram, showing only a few populated DN
levels (note that the spacing is dominantly 8 DN = 2^3, or the lower 3
bits of the 12-bit range)