Actually the database in Oracle is setup and functioning correctly in UTF-8 through the ORA_NLS. The application we wrote using various web service technologies (except for anything .NET :-) ) works wonderfully. It is only the extract we have with problems with.
In answer to the other reply above, yes we are using Oraperl. I was told we could not upgrade anything on the box originally. However, I did identify the possibility that Oraperl is to blame and have found the right Unix admin to add DBD:Oracle to the system so we can use it. Unfortunately the box is on the other part of the world for us, so this takes time for us to implement. Admitedly this is one of my only foray's into Perl and I love what I see so far. I am a C++ / Java person and now this has my interest totally.
Thank you for support and helpful hints!
Akira