odbc stripping multi-byte characters - help!

How can I make calls to the odbc driver to tell it what encoding the client
is.  So when perl dbi runs it will make the connection, it knows already
what the content encoding is, and needs to make a call to the odbc driver
what to convert the data to so odbc does not default to the locale setting.
Basically I am tying to extract ntext and nvarchar unicode data from MS SQL
7 and 2000 using dbi but it is showing up as "????" - basically odbc is
stripping the multi-byte characters.   If I set the locale on my machine to
the appropriate locale it works, but I need my locale to remain US English,
not Korean, Japanese, etc.  On Oracle there is the NLS_Lang variable but SQL
Server always looks to the locale.  What can I do?

Thanks for your help!

Nikko

• PPM for DBI 1.20 by Richard Chiswell
• Re: PPM for DBI 1.20 by Simon Oliver
• odbc stripping multi-byte characters - help! by Nikko Odiseos
• New DBD::ODBC 1.33 release by Martin J. Evans