First, check out the links to good documentation that correctly explains pseudo-streams, byte range requests, and mp4 fragmentation. Please note that I will only use the mp4 container (h264 codec) and HTML5 video (no flash).
My understanding of the pseudo-stream is that the client can send the initial parameter, which the server "searches" for an answer in it. MOOV data must be in advance, and this implies that buffering of the source source stops in favor of a new response starting from the start / search position. How is a client forced to make pseudo-calls? Should MP4 format be formatted in a special way?
Requests for a range of bytes vary, not just the initial parameter that is sent. Sounds more like a progressive download. How to “look for” a job? Does it have a range of bytes? Can I pre-determine the size of a segment with movie information?
How does MP4 fragmentation work? It looks like a design originally developed by Microsoft for Silverlight. But is it applicable to other browser html5 video implementations?
Finding sophisticated information on the Internet. Looking at a live channel and occupy the historical segments of h264 files created from rtp camera streams. Got a bunch of timed files in MongoDB. Created my own h264 decoder in JavaScript and can build an mpeg-dash box on the fly from a range request. Use Chrome support for MSE to add segments. It works fine, but it is not a universal solution. Want to abandon other methods besides flash, but with html5 video.
source
share