Architecture
Last updated
Last updated
Input: target web URL (the edited video)
Processing:
Request metadata from an external source (e.g. Internet Archive)
Data parsing
source locations (e.g. URLs)
time (start and end of source, start and end of target URL)
other extra information (filmed by, title etc.)
captions metadata
Generate a manifest file - Have an chimera HLS/Dash manifest file (.m3u8 or .mpd) from the data, such that your playlist contains the metadata for segments and the rendering is prioritised. JWPlayer supports embedding manifest files https://support.jwplayer.com/customer/en/portal/articles/1430240-adaptive-streaming Example: https://bitdash-a.akamaihd.net/content/sintel/hls/playlist.m3u8 Note: It was pointed out the exact segment difficult to extract out exactly, but this is just a nice-to-have feature so the segments can be matched up with best effort.
Output:
Captions
Make the captions go the right way... (Top to bottom)
Video
Render a poorer quality video for the source file as well. Just because the video is a lot smaller.
Pop up video frame that would be easier to close / be able to toggle between the top and bottom video.
What happens when.. fade? Overlapping video? Silent?
Data model will change based on the API endpoint of the database we get our information from. For now we have a mock manifest placed in the repository.
See .m3u8 / mpd
Not considered as this is the mock of a mock of a mock.