1MRF image format specification(0) MRF image format specification(0)
2
6 MRF - monochrome recursive format (compressed bitmaps)
7
10 This program is part of Netpbm(1).
11 MRF is a compressed format for bilevel (1-bit mono) images. It
13 achieves better compression for some such images than either GIF or
14 PNG. (It's also very easy to implement (about the same difficulty as
15 RLE, I'd say) and an MRF reader needs no tables/buffers, which may make
16 it useful for tiny machines).
17 In case the above hasn't made it sufficiently clear, I'll make this
19 next point explicitly: MRF cannot represent color at all. Nor can it
20 represent grayscale. It's a specifically mono format. (If you want to
21 compress a color or grayscale image, my advice is to use JPEG2000).
22 First, here's what goes where in an MRF file. I'll explain how the com‐
24 pression works afterward.
25 Offset Description
29 0 magic number - 'MRF1' (in ASCII)
31 4 width (32-bit, MSB first (i.e. big-endian))
34 8 height (same)
37 12 reserved (single byte, must be zero)
40 13 compressed data
43 Note that there is no end-of-file marker in the file itself - the com‐
47 pressed data carries on right up to EOF.
48 The way the picture is compressed is essentially very simple, but as
50 they say, the devil is in the detail. So don't be put off if it sounds
51 confusing.
52 The image is treated as a number of 64x64 squares, forming a grid large
54 enough to encompass it. (If an image is (say) 129x65, it'll be treated
55 in the same way as a 192x128 one. On decompression, the extra area
56 which was encoded (the contents of this area is undefined) should be
57 ignored.) Each of these squares in turn (in left-to-right, top-to-bot‐
58 tom order) is recursively subdivided until the smallest completely
59 black or white squares are found. Some pseudocode (eek!) for the
60 recursive subdivision routine should make things clearer:
61 if square size > 1x1 and square is all one color, output 1 bit
63 if whole square is black, output a 0 bit and return
64 if whole square is white, output a 1 bit and return
65 output a 0 bit
66 divide the square into four quarters, calling routine for
67 each in this order: top-left, top-right, bottom-left, bottom-right
68 (Note that the 'output a 0 bit' stage is not reached for squares of
70 size 1x1, which is what stops it recursing infinitely. I mention this
71 as it may not be immediately obvious.)
72 The whole of the compressed data is made up of the bits output by the
74 above routine. The bits are packed into bytes MSB first, so for example
75 outputting the bits 1,0,0,0,0,0,0,0 would result in a 80h byte being
76 output. Any `unused' bits in the last byte output are undefined; these
77 are effectively after EOF and their value is unimportant.
78 If writing that sounds too much like hard work :-), you could always
80 adapt pbmtomrf and/or mrftopbm. That's the main reason their source
81 code is public domain, after all.
82 Above, I said the contents of any extra area encoded (when a bitmap
84 smaller than the grid of squares is compressed) is undefined. This is
85 deliberate so that the MRF compressor can make these unseen areas any‐
86 thing it wants so as to maximize compression, rather than simply leav‐
87 ing it blank. pbmtomrf does a little in this respect but could defi‐
88 nitely be improved upon.
89 mrftopbm's -1 option causes it to include the edges, if any, in the
91 output PBM. This may help when debugging a compressor's edge optimiza‐
92 tion.
93 Note that the "F" in the name "MRF" comes from "format," which is
95 redundant because it is the name of a format. That sort of makes "MRF
96 format" sound as stupid as "PIN number," but it's not really that bad.
97
101 mrftopbm(1), pbmtomrf(1)
102
105 Russell Marks.
106netpbm documentation 1991 MRF image format specification(0)