I'm looking at building a faster machine to support up to 10 cameras. I've seen other CCTV software use the Intel Quick Sync which allows very fast decoding on it's 6th Gen and above chips. I believe that Sighthound cannot use this due to it's software but I had another thought.
Why decode a 4MP camera stream for movement when decoding a 1080P stream would be good enough. My cameras support Stream 1 and Stream 2, so why not have a configuration that you can setup the motion detection on stream 2 (low res) and have stream 1 (high res) being recorded. That way the processor overhead must be radically reduced.
