From a technical standpoint it's pretty much impossible for the dose files or even the software to contain actual voices... the file sizes are just too small. The only way to work in subliminal messages on such a small file size would be with speech synthesis, and synthesized voices are very obvious...