Retrieved the contents of a css file: (http://gizmodo.com/assets/stylesheets/app-ecbc6044c59319aab4c2a1e31380ef56.css)
Detected the encoding with mb_detect_encoding
... says UTF-8.
Viewed the page in a browser, looks fine (readable), and declares @charset "UTF-8";
Tried to output the string, got garbage. Tried to save it to a file, got garbage.
Tried to convert the encoding to ASCII, ISO-8859-1, and HTML-ENTITIES. No luck.
Any ideas here how to determine why this string is garbage, and how to fix it?