why use instructions instead of source code for control-flow graph?

Back when I was learning CFG in school, we've been applying CFG on program source code as from the book "introduction to software testing". But now that I tried to use CFG generator tools such as CFG Factory by Dr. Garbage, all CFGs have been generated using bytecode instructions. I do know that bytecode instructions are probably more accurate, but isn't it going to be difficult for people to understand the source code from each block for debugging?