IPVMU Certified | 03/01/13 12:29pm
The QoS requirements on the network immediately come to mind. Human communications via VoIP or Video conferencing is a two way communication system, and therefore demand network designs that are v low in jitter and low delay. Networks designed for one way video dont care much about delay nor jitter, as long as the bandwidth is there, netwotk delay in cctv doesnt matter much.
Sure, after all there is only so much difference between an IP security cam and a video conferencing cam.
The details are in the software. Dealing with firewalls, network latency issues, ensuring audio/video sync, being able to prioritize audio over video in the event of network congestion, etc.
IPVMU Certified | 03/01/13 05:45pm
The video support is probably straightforward enough, but the audio support might be tricky.
Bi-directional audio is not a common feature for surveillance cameras, and integrated mics do not typically have great fidelity, so some other equipment is sure to be required.
In addition to what the Brians have said, remember that two-way audio in surveillance is usually intended to be half-duplex, like a walkie-talkie: push to talk, release to listen, not do both simultanesouly. Videoconferencing systems, which are essentially a speakerphone with a camera, have carefully placed mic(s) and speaker(s) with appropriate cancellation systems to prevent audio feedback when using full-duplex audio.
I'd suggest, in contrast to BK's statement, that you'd want audio to be a priority over video in such a system as well, since most people don't lip-read - being able to see someone talking without hearing them would be pointless.