I am trying to understand how the user defined fonts are encoded within the prg files. I was told that a RLE compression is used but I was not able to decode the data so I need some help of math experts... 
Here's an example of a user font source bitmap...
...and this is the result within the compiled program...
Here's a closer look how some pixel lines have been changed from the original glyph to the compiled data...
...and that's all I've found so far - does anyone know, how the compiled data can be expanded again to get the original font?