I don't know, but my hunch is that the service provider is trying to ensure nobody can capture the stream instead of just rendering it (not that that would stop the determined hacker). So the idea may be that only if the streaming client announces itself as a TV (ie display only, no recording capability) does the server send the content.