The International Telecommunications Union, ITU-T
(formerly know as CCITT), is a multinational union that provides standards for
telecommunication equipment and systems. Here we will explain two standards for
facsimile compression, T.4 and T.6.
Recommendation T.4 defines the characteristics of Group 3 facsimile terminals
which enable black and white documents and also optionally colour documents to
be transmitted on the general switched telephone network, international leased
circuits and the Integrated Services Digital Network (ISDN). Group 3 facsimile
terminals may be operated manually or automatically and document transmission
may be requested alternatively with telephone conversation.
T.6 facsimile coding schemes consist of the basic facsimile coding scheme and
optional facsimile coding schemes. They are defined in Section 2, 3 and 4.
Facsimile coding schemes are specified assuming that transmission errors are
corrected by control procedures at a lower level. The basic T.6 facsimile coding
scheme is the two-dimensional coding scheme which is in principle the same as
the two-dimensional coding scheme of Group 3 facsimile specified in
Recommendation T.4. Optional facsimile coding schemes are specified not only for
black and white images but also for grey scale and colour images. Facsimile
coding control functions are used in facsimile user information in order to
change facsimile parameters or to invoke the end of facsimile block.
The CCITT T.4 and T.6 specifications outline many different kinds of
requirements for the facsimile apparatus, including the resolution of scanning
and printing, dimensional tolerances, timing constraints etc. The sections which
tend to be used in the imaging world are those which describe the compression
schemes, and these are often referred to as Group 3 and Group 4 in their own
right. The compression and decompression algorithms are quite complex, as will
be seen, and must be implemented either in hardware or sophisticated software in
order to achieve good performance.
So how do the Group 3 and Group 4 compression mechanisms work?
Both schemes encode the source image on a horizontal scanline-by-scanline basis,
corresponding to the way in which documents are scanned and printed on a
facsimile machine. The difference lies in the way the two standards handle
successive scanlines - in Group 3, each scanline is encoded independently,
whereas in Group 4, scanlines are encoded with reference to the previous one,
resulting in improved compression ratios.
An interesting side-effect of the way that Group 3 and Group 4 work is that the
size of compressed images is roughly linearly proportional to the resolution at
which the image is scanned, whereas that of the original is of course
proportional to the square of the resolution. This is because the total
compressed size is a function of the number of scanlines, with roughly the same
number of codes being produced for each scanline irrespective of resolution.
Compression ratios achieved using Group 3 of course depend upon the nature of
the source image. Under certain circumstances, it will be obvious that the
encoded version could be larger than the original (consider the case of single
alternating black and white pixels), but in general compression ratios of 10:1
can be achieved on pages of typewritten text.
Compression ratios achieved using Group 4 are several times better than Group 3,
yielding upwards of 25:1. Most imaging systems now use the Group 4 compression
standard, as a result of the improved compression and the development of Group 4
software compression and decompression which executes at high speed.
Here you can find these standards explained :