I have uploaded the following gifs to Michael R's ftp site:
and
They show the leading and trailing columns from two images on Disk 5 (images where produced by fv).
The first two are columns from the image (h3r1438.903) that Herb Johnson has already posted results for: Herb Johnson Wrote on Friday, November 05, 1999 05:13
---SNIP--- > > for image H3R1438.903: > > column average standard deviation > 0 42609 471 > 1 40649 2.9 > 2 42299 211 > 3 40650 3.03 > 4 42628 711 > 5 40650 2.96 > 6 44958 822 > 7 44997 753 > 8 44982 808 > (values checked to column 12, similar to col 8) > > 2037 44860 996 > 2038 42501 940 > 2039 46441 890 > 2040 42425 495 > 2041 53348 8628 (not typos) > 2042 42383 907 > > Tom's quote below suggests his format is 2 prescan pixels, 2 > dark pixels, > 2034 data pixels, 2 postscan pixel (but 2043 total anyway). My > data suggests a mix of SIX leading dark and prescan pixels, 2034 > probable data pixels, and THREE postscan pixels. > > Columns 1, 3, and 5 > (from column 0) are candidates for dark pixels as they are very > consistent and have almost equal average values. Columns 0, 2, and 4 > have averages and deviations well above the dark columns but different > enough from the data columns that they are not likely to be > image data. > Column 2041 has a very wide distribution and a high average; column > 2042 happens to look similar to the data columns. Column 2040 > happens to have a lower distribution but a similar average to the > data columns: it may be data, analysis of other images would > be informative. None of the trailing columns are likely to be dark > pixels, assuming dark pixels will have a low variance.
in h3r1438.903_cols_2030-2043..gif the second to last column in the array exhibits some banding (2 sets of 4 bands) and as Herb Johnson showed has a large scatter. This is not apparent in the corresponding I band camera image (h4r1438.903_cols_2030-2043..gif).
(Columns are numbered from 1 - as opposed to Herb J numbering from 0) (Sorry - I mucked up the colours for columns on the end of scan gifs)
V Band Columns 1-6 H3R1438.903: Columns 2,4,6 have a very similar scatter (Standard deviation of ~3 from Herb J) and overlap closely on the graph. They are presumably the same sort of column. Visually Column 3 appears to be similar to 2,4 & 6 but Herb J measures a higher St. dev although it is still smaller than the other columns. Columns 5 & 1 overlap. Column 2042 has banding I Band Columns 1-6 H4R1438.903: Columns 2,4,6 & 3 are similar to those in H3R1438.903 Columns 5 & 1 do not overlap Column 2042 has no banding - Herb J also noticed that it has identical values to its neighbouring column 2043
(Sorry - I mucked up the colours for columns on these)
Herb Johnson wrote on Saturday, November 06, 1999 05:28
> Speaking of "funny numbers", I'm seeing something odd in the #5 disk > of Tom's Mark IV data. Image H4r1446.880 has IDENTICAL COLUMNS of data > for columns 2041 and 2042. It's not my program, three other files > do not have this feature: H3r1446.880, Vdark15.fts, H3r1438.903 have > in column 2041 an average value well above the other column, and a > MUCH greater standard deviation (thousands rather than hundreds) than > the imaging columns. Another Mark IV analyist says that the H4 camera > does not have the "banding" of the H3 camera: I interpret that to mean > that this problem may be unique to the H4 camera (or the software > used with data from that camera). Tom, any comments on this and > the uses of the last few columns of the Mark IV images?
For Disk5: All of the I Camera (H4*) images have identical data for the last two columns as do the 3 idark images.
I only checked 8 images from the V Camera and none of them exhibit this feature. Although the ones I have graphed columns for all exhibit banding in the second to last column (the one with the very large scatter).
I&V image on disk 4 appear to have similar start and end scan columns to those on disk5.
