A lossless compression algorithm is only the result of a shorter description of a binary combination that usually appears several times in the uncompressed data.
To differ and mark the uncompressable data from the compressed you need something I call the "overtake" sign or short ovtk. If something compressable appears you write a sign if an ovtk appears before or not. If there is any you need to encode the length of the bytes that is not compressable and has to be used without any change for decompression.
The way you mark something as an ovtk-block with the length can have a big influence on the overall compression efficiency. So sometimes it is rather useless to compress a 16 Bit value if the ovtk-sign + ovtk length + compression sign + compression code is bigger than the 16 Bit and some kind of skipping algorithm or efficiency analysis for this case results in a better compression.
Now my question: Is there is another expression for this "ovtk" marker and what different kind of length encoding for the uncompressable block exist? In most cases compression discussions only talk about the other part: The compression algorithm itself and my interest is on getting more information about the encoding of uncompressable data. Are there any sources talking more about this?