Control Characters within Original Data

A supplementing mechanism must be specified to work with any combination of data, i.e. data also containing the control character. The simplest variant uses a run of the length 0 containing the control character as run value itself.

   Counter   Sequence Length
       0          1
       1          2
          ...
     255        256

   uncoded data:   abc#dddddefg
     coded data:   abc##0#d4efg

The introduction of an offset is also an option for this procedure. If a lower number of control characters appears, the sequence (##0) has to be repeated (## -> ##0##0).

The most important question for the efficiency of this coding scheme is, how often the control character is used within the original data. In text files for example, code words greater than 127 rarely appear like any code word above the 7 bit limit. On the other hand, uncompressed graphic files frequently contain certain code words like 0 or 255 (FFh), e.g. for white (0) or black (FFh) pixels.

 <   ^   >