semi-automatic region recoding?

Hi everyone,

I have configured Emacs and ERC to prefer utf-8 as the coding
system for most IRC channels, but there are a few channels where
both utf-8 and iso-8859-7 are used.

My current Emacs and ERC coding system setup is:

    (prefer-coding-system 'utf-8)
    (setq erc-server-coding-system (cons 'utf-8-unix 'utf-8-unix))

When someone writes in iso-8859-7 with my current setup, I see:

    \360\357\353\375 \341\361\343\374

If this was recoded from iso-8859-7 to utf-8, it would be:

    πολύ αργό

I think I've seen irssi automatically `guess' the coding system
of incoming text and recode it to the terminal locale.  Is
something like this possible with ERC too?

TIA,
Giorgos