The reasons are time & CPU. I’m recording files constantly for hours, stopping and starting the record roughly every 10 minutes. Meanwhile, I’m capturing and streaming in HD too.
After the files are done recording, I need to move the files off box, from multiple boxes, to a single box on the network. All that moving files around takes up network bandwidth, and actually get’s me congested to where I can’t keep up if I do that. I could transcode to H.264 before I move the files, but that introduces a delay, and consumes CPU power that I don’t have to spare in my application.
That’s why the ideal situation is to record to hard disk in H.264 in the first place, using the hardware so I don’t take the CPU hit.
By your asking why I’m not doing this, should I infer that recording straight to h.264 using the GPU isn’t possible?