Encoding itself isnt the big deal and wouldn't add much, the development time would come from the implementation - in making the video output something of use. See it could just stream the live race out from your driving position (on a sufficiently powerful computer) or it could do it during a replay. The time will come in adding features like camera control and editing etc to make the resulting video of use for a purpose of some kind - of which videos in LFS have numerous purposes (from broadcast, to artistic videos, to crash videos, to stunts etc etc), so lots of features would be needed to make the feature actually useful in any way, and that's where a substantial amount of development could take place ... when we already have all the interfaces we need to do all these things anyway.