RPM database codeset

I'm writing an utility to retrieve information from RPM installation
databases. On the machines where I'm testing the name, description and
vendor fields always contain ASCII characters.
Is it simply because my machine is an English one? Or does RPM accepts
only ASCII characters?
If not, is there a unique language/codeset for the DB? Or does each
entry have its own one? If the latter is true, how can I retrieve the
DB (or entry) codeset and language?

Thanks,
Ber