is the DB set up to store unicode? is what you store in DB unicode? is what you get out from DB encoded as unicode?

placing a binmode(STDOUT, ":encoding(utf-8)"); (thanks Tux) at the beginning of your script will display unicode to STDOUT and complain if it contains invalid bytes. But you may need to tell your db to store unicode (for mysql it used to be something like this during table construction: mysql_enable_utf8mb4) and make DBI::connect aware with mysql_enable_utf8 => 1. For some cases you need to explicitly encode the bytes received from DB into unicode using Encode::encode... or investigate if it's a case of double encoding using Encode::decode...