Sorry if I was not clear. The first code snippet I posted is a line in navit.xml that refers to another file, namely /usr/local/sbin/navit-speech-wrapper. This file should contain the contents of the second code snippet (altered as necessary to suit your espeak command line preferences). It must have execute permission, and can be located anywhere and named anything, as long as the path specified in navit.xml is correct. Navit will then run that command whenever speech output is required..