Has anyone tried this ?
What's the best practice for this?
Has anyone tried this ?
What's the best practice for this?
FMS live streams are using the RTMP protocol:
ffmpeg -i rtmp://server/path -acodec copy -vcodec copy -y captured.flv
Here, we are saving the whole stream to an FLV file, which is Flash's static movie file format and so can always preserve all RTMP audio and video codecs without conversion.
You can then extract any frames you want, e.g.
ffmpeg -i captured.flv -s starttime -vframes 1 -f image2 -vcodec mjpeg captured.jpg
If you are ambitious and know exactly what time offsets and intervals you want to capture in advance, you can do both steps at once, e.g. one frame every second:
ffmpeg -i rtmp://server/path -r 1 -f image2 -vcodec mjpeg captured%d.jpg
All commandlines have not been tested, will need fixing but give you a good impression