I currently have 15x 5mp Dahua cameras, but hoping to eventually upgrade to 4k when they become more affordable.
Blue Iris currently runs in a Windows VM on a Proxmox server with Ryzen Pro 5750G, 64gb ECC RAM and Quadro P2000.
Attached are my current AI setting and logs in BI and CPAI.
I've been researching frigate vs CPAI and dual edge TPU vs Nvidia GPU for the past few days, but can't seem to find solid direct comparisons.
For people with experience in above setup, what would be my optimal setup in terms of accuracy(false positive), speed, scalability(in case all cameras are upgraded to 8mp) and power consumption(last and least)?
Apart from the Quadro P2000, I also have a spare Nvidia 1660 super, and don't mind the cost of dual edge TPU + PCIe adaptor if that's the optimal hardware.
I also don't mind setting up frigate in Home Assistant or as a container/docker just for notification and event tagging and leave BI to 24/7 recording only.
Blue Iris currently runs in a Windows VM on a Proxmox server with Ryzen Pro 5750G, 64gb ECC RAM and Quadro P2000.
Attached are my current AI setting and logs in BI and CPAI.
I've been researching frigate vs CPAI and dual edge TPU vs Nvidia GPU for the past few days, but can't seem to find solid direct comparisons.
For people with experience in above setup, what would be my optimal setup in terms of accuracy(false positive), speed, scalability(in case all cameras are upgraded to 8mp) and power consumption(last and least)?
Apart from the Quadro P2000, I also have a spare Nvidia 1660 super, and don't mind the cost of dual edge TPU + PCIe adaptor if that's the optimal hardware.
I also don't mind setting up frigate in Home Assistant or as a container/docker just for notification and event tagging and leave BI to 24/7 recording only.
