Sample Input Data Files

The following files may be useful as sources of data for certain assignments, listed in order of increasing ( jpeg ) size.

Note: Either the raw or No 0s formats are probably the best choice as data files. See below for discussion.

Base Filename	File Formats					Histogram	Notes
Base Filename	raw	jpeg	bmp	gif	No 0s	Histogram	Notes
light_drops	48 kB	5 kB	49 kB	20 kB	48 kB	light_drops	Small and fairly monochromatic
diver_resize	81 kB	37 kB	82 kB	23 kB	81 kB	diver_resize	Small with a variety of ( mostly dark ) shades
Chevalier_473	2.0 MB	202 kB	2.1 MB	460 kB	1.9 MB	Chevalier_473	Wide variety of bright colors, with fine-grain resolution
GrandeJatte	1.2 MB	575 kB	1.3 MB	480 kB	None!	Grande Jatte	Similar to Chevalier_473, but with softer colors
LUG_Newbies	1.2 MB	614 kB	1.2 MB	222 kB	1.1 MB	LUG_Newbies	Mostly black and white
Esher	1.8 MB	669 kB	1.8 MB	741 kB	1.7 MB	Escher	Black and white, should contain sets of 3 identical values
diver	9.2 MB	804 kB	9.2 MB	1.5 MB	9.0 MB	diver	The full-size version of diver_resize
VRUPL_Logo	14.9 MB	1.3 MB	14.9 MB	252 kB	8.9 MB	VRUPL_Logo	Only 4 colors, should have STRONG locality of reference.
JohnBell	18.0 MB	3.0 MB	18.4 MB	3.2 MB	17.8 MB	JohnBell	A diver doing survey work in Lake Michigan
calibration	21.1 MB	3.0 MB	21.7 MB	4.2 MB	10.9 MB	calibration	A standard printer calibration image.

JPG was the original file format of the pictures. Some research into the JPG format shows that it goes through a number of transformations, including conversion from RGB into YCbCr, a cosine transformation, and Huffman encoding.
BMP does not do compression, and preserves the RBG bytes of the original image. However it does include a header at the beginning of each file.
GIF files are in indexed-color format. At the beginning of the file is a color table with RGB values, and then for individual pizels there is just a single integer reference to the table for each pixel. ( GIF may also compress sequences of identically colored pixels using run-length encoding. )
RAW files are the raw R, G, and B bytes, with no compression, headers, or processing. These are probably the preferable files to use.
No Zeros are raw files with all null bytes removed. If the presence of ( potentially lots of ) zero bytes in the data set is a potential problem, then these files should be used in lieu of raw files.