Hi JS, I'm just back online after a trip so just starting to catch up. Given there is lots left for us to learn about the ffmpeg API and the issues surrounding video access I think we'll need to subscribe to the ffmpeg mailing list and see if that will provide some good advice.
Robert. On Tue, Apr 14, 2009 at 3:25 PM, Jean-Sébastien Guay < [email protected]> wrote: > Hi Robert, > > I have merged your error code changes and the additional formats to >> ReaderWriterFFmpeg.cpp, but haven't merged the http & rtsp as I'd like to be >> able to test them directly myself and to would like to invesitgate the >> appropriate ffmpeg way to handle this type of input source. >> > > Go ahead, you just need to set up your streaming source to serve the stream > using these protocols. By the way, I did mention that it doesn't work 100% > even for me right now, I've just hit a roadblock and would like someone else > to see if they can get further, so I'm not surprised you didn't merge those > changes. > > What makes me more cautious about the code path is that my own /dev/ code >> is hardwired to a low resolution for the stereo webcam I've been >> experimenting and plan to remove this code and replace with more generic >> code down the line, and you've just bolted extra code on to this >> experimental code path that. I'd prefer to see a device code path kept >> separate from a network streaming path. I also tested watching movies over >> http using the original ffmpeg code, it had problems but it worked so I'm >> curious to as why you needed to use this other route. >> > > It's not a file, it's a stream. You're right, for http those changes were > not necessary so it was the addition of formats to ReaderWriterFFmpeg.cpp > that did it. But even then, it still stops after 15-20 seconds for me, as I > said. > > But for rtsp, we need to open the stream using that format (pass "rtsp" in > av_find_input_format() ) and even then, it doesn't work for me. I'm thinking > we need to pass in an additional format to specify it's MPEG-4 inside the > container, but I really don't know enough about ffmpeg to figure out how. > > In any case, I think we'll need to have different paths for a *file* served > through http versus a *stream* served through http, but once again I don't > know enough about these things. Perhaps I'm wrong and the same code path can > handle both. > > Could you provide example files for http and rtsp so that we can all test >> against the same sources and can observe the same issue. >> > > I cannot publicly serve video streams from our cameras, no. These are > streams, not files, so you really need to have a camera that can serve using > those protocols... (unless someone can set up a stream that loops over a > given file and serves that, but once again I'm way over my head here) > > J-S > -- > ______________________________________________________ > Jean-Sebastien Guay [email protected] > http://www.cm-labs.com/ > http://whitestar02.webhop.org/ > _______________________________________________ > osg-submissions mailing list > [email protected] > > http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org >
_______________________________________________ osg-submissions mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
