Fixing character encoding conversion using Encode and other small fixes

[PhobosK]

Hi,
Here are the changes I propose:

• I did change the way the encoding conversion is done. It doesn't use piconv by a system call anymore, since piconv is actually Perl Encode. I removed all system() calls since they are not needed and have problem with file naming (if unquoted, encoding etc). The logic now is that if the char encoding conversion fails, we at least will have a valid non-char-encoding-converted file... All this closes Error: function tmpnam undefined. #2 as well...
• I did fix the subtitle formats naming mix-up, so this will close README: subtitle types mixed up #3 too...
• Some spelling and formatting clean-ups
• A version bump up to 0.5.6

If you like you may accept this pull request

Thanks