Multi-Character Set Support

I need to set up a instance which support English and Chinese data BOTH.
How I implement it, what I need?
1. create a new instance
2. convert a existing instance with NLS_CHARACTERSET = WE8ISO8859P1

I know I need to setup up NLS_lang, NLS_character set...
I understand I have 2 options, UTF8 and simple/traditional chinese. which one I should use?

thanks in advance