The RLE algorithm described in this section is used to compress Byte Segments into RLE Segments. There is a one-to-one correspondence between Byte Segments and RLE Segments. Each RLE segment must be an even number of bytes or padded at its end with zero to make it even.
A sequence of identical bytes (Replicate Run) is encoded as a two-byte code:
< -count + 1 > <byte value>, where
count = the number of bytes in the run, and
2 <= count <= 128
and a non-repetitive sequence of bytes (Literal Run) is encoded as:
< count - 1 > <Literal sequence of bytes>, where
count = number of bytes in the sequence, and
1 <= count <= 128.
The value of -128 may not be used to prefix a byte value.
It is common to encode a 2-byte repeat run as a Replicate Run except when preceded and followed by a Literal Run, in which case it's best to merge the three runs into a Literal Run.
Three-byte repeats shall be encoded as Replicate Runs. Each row of the image shall be encoded separately and not cross a row boundary.
Pseudo code for the RLE decoder is shown below:
Loop until the number of output bytes equals the uncompressed segment size
Read the next source byte into n
If n> =0 and n <= 127 then
output the next n+1 bytes literally
Elseif n <= - 1 and n >= -127 then
output the next byte -n+1 times
Elseif n = - 128 then
output nothing
Endif
Endloop