[MEncoder-users] ATSC closed captioning

[Scott W. Larson]

Reimar Döffinger <Reimar.Doeffinger at gmx.de> wrote:
>Yes, probably the best way. It does make sense since there will be
>one CC set per channel/video stream, using this struct instead of
>global data should work better if we ever want to support
>processing of multiple video streams at the same time.

Does this assume that the captions will only be in the user data of displayable frames? 

My understanding is that MPEG user data can also be included in the sequence and GOP headers. Will video_read_frame() always be able to pass this data to decode_video()?