| 1 |
\ |
| 2 |
.\" This man page was generated by the Netpbm tool 'makeman' from HTML source. |
| 3 |
.\" Do not hand-hack it! If you have bug fixes or improvements, please find |
| 4 |
.\" the corresponding HTML page on the Netpbm website, generate a patch |
| 5 |
.\" against that, and send it to the Netpbm maintainer. |
| 6 |
.TH "MRF image format specification" 1 "1991" "netpbm documentation" |
| 7 |
|
| 8 |
.SH NAME |
| 9 |
|
| 10 |
MRF - monochrome recursive format (compressed bitmaps) |
| 11 |
|
| 12 |
.UN description |
| 13 |
.SH DESCRIPTION |
| 14 |
.PP |
| 15 |
This document describes the MRF format recognized by |
| 16 |
.BR "Netpbm" (1)\c |
| 17 |
\&. |
| 18 |
.PP |
| 19 |
MRF is a compressed format for bilevel (1-bit mono) images. It |
| 20 |
achieves better compression for some such images than either GIF or |
| 21 |
PNG. (It's also very easy to implement (about the same difficulty as |
| 22 |
RLE, I'd say) and an MRF reader needs no tables/buffers, which may |
| 23 |
make it useful for tiny machines). |
| 24 |
.PP |
| 25 |
In case the above hasn't made it sufficiently clear, I'll make this |
| 26 |
next point explicitly: \fIMRF cannot represent color at all.\fP Nor |
| 27 |
can it represent grayscale. It's a specifically mono format. (If you |
| 28 |
want to compress a color or grayscale image, my advice is to use |
| 29 |
JPEG2000). |
| 30 |
.PP |
| 31 |
First, here's what goes where in an MRF file. I'll explain how the |
| 32 |
compression works afterward. |
| 33 |
|
| 34 |
|
| 35 |
.TP |
| 36 |
Offset |
| 37 |
Description |
| 38 |
.TP |
| 39 |
0 |
| 40 |
magic number - "MRF1" (in ASCII) |
| 41 |
|
| 42 |
.TP |
| 43 |
4 |
| 44 |
width (32-bit, MSB first (i.e. big-endian)) |
| 45 |
|
| 46 |
.TP |
| 47 |
8 |
| 48 |
height (same) |
| 49 |
|
| 50 |
.TP |
| 51 |
12 |
| 52 |
reserved (single byte, must be zero) |
| 53 |
|
| 54 |
.TP |
| 55 |
13 |
| 56 |
compressed data |
| 57 |
|
| 58 |
|
| 59 |
.PP |
| 60 |
Note that there is no end-of-file marker in the file itself - the |
| 61 |
compressed data carries on right up to EOF. |
| 62 |
.PP |
| 63 |
The way the picture is compressed is essentially very simple, but |
| 64 |
as they say, the devil is in the detail. So don't be put off if it |
| 65 |
sounds confusing. |
| 66 |
.PP |
| 67 |
The image is treated as a number of 64x64 squares, forming a grid |
| 68 |
large enough to encompass it. (If an image is (say) 129x65, it'll be |
| 69 |
treated in the same way as a 192x128 one. On decompression, the extra |
| 70 |
area which was encoded (the contents of this area is undefined) should |
| 71 |
be ignored.) Each of these squares in turn (in left-to-right, |
| 72 |
top-to-bottom order) is recursively subdivided until the smallest |
| 73 |
completely black or white squares are found. Some pseudocode (eek!) |
| 74 |
for the recursive subdivision routine should make things clearer: |
| 75 |
|
| 76 |
.nf |
| 77 |
if square size > 1x1 and square is all one color, output 1 bit |
| 78 |
if whole square is black, output a 0 bit and return |
| 79 |
if whole square is white, output a 1 bit and return |
| 80 |
output a 0 bit |
| 81 |
divide the square into four quarters, calling routine for |
| 82 |
each in this order: top-left, top-right, bottom-left, bottom-right |
| 83 |
|
| 84 |
.fi |
| 85 |
.PP |
| 86 |
(Note that the "output a 0 bit" stage is not reached for squares |
| 87 |
of size 1x1, which is what stops it recursing infinitely. I mention |
| 88 |
this as it may not be immediately obvious.) |
| 89 |
.PP |
| 90 |
The whole of the compressed data is made up of the bits output by |
| 91 |
the above routine. The bits are packed into bytes MSB first, so for |
| 92 |
example outputting the bits 1,0,0,0,0,0,0,0 would result in a 80h byte |
| 93 |
being output. Any `unused' bits in the last byte output are undefined; |
| 94 |
these are effectively after EOF and their value is unimportant. |
| 95 |
.PP |
| 96 |
If writing that sounds too much like hard work :-), you could |
| 97 |
always adapt \fBpbmtomrf\fP and/or \fBmrftopbm\fP. That's the main |
| 98 |
reason their source code is public domain, after all. |
| 99 |
.PP |
| 100 |
Above, I said the contents of any extra area encoded (when a bitmap |
| 101 |
smaller than the grid of squares is compressed) is undefined. This is |
| 102 |
deliberate so that the MRF compressor can make these unseen areas |
| 103 |
anything it wants so as to maximize compression, rather than simply |
| 104 |
leaving it blank. \fBpbmtomrf\fP does a little in this respect but |
| 105 |
could definitely be improved upon. |
| 106 |
.PP |
| 107 |
\fBmrftopbm\fP's \fB-1\fP option causes it to include the edges, if |
| 108 |
any, in the output PBM. This may help when debugging a compressor's |
| 109 |
edge optimization. |
| 110 |
.PP |
| 111 |
Note that the "F" in the name "MRF" comes from "format," which is redundant |
| 112 |
because it is the name of a format. That sort of makes "MRF format" sound |
| 113 |
as stupid as "PIN number," but it's not really that bad. |
| 114 |
|
| 115 |
.UN seealso |
| 116 |
.SH SEE ALSO |
| 117 |
.BR "mrftopbm" (1)\c |
| 118 |
\&, |
| 119 |
.BR "pbmtomrf" (1)\c |
| 120 |
\& |
| 121 |
|
| 122 |
.UN author |
| 123 |
.SH AUTHOR |
| 124 |
|
| 125 |
Russell Marks. |
| 126 |
.SH DOCUMENT SOURCE |
| 127 |
This manual page was generated by the Netpbm tool 'makeman' from HTML |
| 128 |
source. The master documentation is at |
| 129 |
.IP |
| 130 |
.B http://netpbm.sourceforge.net/doc/mrf.html |
| 131 |
.PP |