The camera is a TP-Link SC3230, the universal camera input module is using 2 streams, 1 low quality with no sound for the multiview window and a second one in higher resolution with sound. The URLs are:
Preview: rtsp://<user>:<passwd>@<ip>/media.amp?streamprofile=Profile5&audio=0
Store: rtsp://<user>:<passwd>@<ip>/media.amp?streamprofile=Profile1&audio=1
I have selected the options "Enable H264 recording without conversion" and "Use H264 Stream (from archive) in single camera mode". So instead the .xem files in the archives i found .mp4 files. When i reproduce those files i found no audio stream. When i say reproduce I mean copy and play I don't export them because I have a lots of problem with that (wait for future post on that problem).
After testing different configuration, options, adding the audio stream URL (that i can't make it work, but may be the camera), I found that Xeoma is spawning (or forking) 2 FFMPEG process for each camera that I have using 2 streams. This are:
Preview
/usr/local/Xeoma/Additional/ffmpeg_213c -loglevel quiet -probesize 100000 -analyzeduration 10000000 -i rtsp://<user>:<passwd>@<camera IP>:554/media.amp?audio=0&streamprofile=Profile5 -f fsimstream -vcodec fsim -an fssmem:58957732418446744072392849060
Archive
/usr/local/Xeoma/Additional/ffmpeg_213c -loglevel quiet -probesize 100000 -analyzeduration 10000000 -i rtsp://<user>:<passwd>@<camera IP>/media.amp?audio=1&streamprofile=Profile1 -map 0 -dn -f fssegment -segment_format mp4 -vcodec copy -acodec aac -strict experimental -b:a 32K /usr/local/Xeoma/XeomaArchive/Preview+Archive.20/2014-07-20/h264/1173_53.mp4 -f switchablefsimstream -map 0 -dn -vcodec copy -an -r 15.000000 fssmem:58956519918446744072365759244
Both process (and all other FFMPEG process) have the parameter "-an" in it. And it mean DISABLE AUDIO in the FFMEG documentation. It shouldn't be on the archive process. Or am I wrong?.