Help With Specking a 25 4k Camera NVR

[Maverick9661]

Hello,

I am looking at building a 25 4k camera system for my office and need help specifying what hardware we need for our NVR Server.

We are planning on recording constantly using a low framerate 1080 sub-steam with motion-triggered full res. Additionally, we are looking at using CodeProject AI for Person and vehicle detection although we may run this on other hardware.

In testing I have used a system with an intel 12900k and followed the substream and other optimization guides on the forums. However, I haven't found a software configuration that will scale well among 25 cameras without separating into multiple NVRs for the final deployment. But I'm all ears for suggestions.

Its worth noting that we will deploy this in a server chassis and will have a separate NAS that is mounted for the recordings directory.


Camera Specs:

EmpireTech IPC-Color4K-T and IPC-Color4K-X
15 or 20FPS
3840x2160



Let me know if you need any additional details. I greatly appreciate any advice or pointers as I'm fairly new to BlueIris.

Thank you,
-Maverick

 

[looney2ns]

IMHO, 4k cameras are overkill. Takes lots of horsepower and storage space to accommodate them.
The best all around cameras at this point in time is THIS varifocal or THIS fixed lens.

Not sure what you mean here:
"In testing I have used a system with an intel 12900k and followed the substream and other optimization guides on the forums. However, I haven't found a software configuration that will scale well among 25 cameras without separating into multiple NVRs for the final deployment. But I'm all ears for suggestions."

A 12900k machine will handle 25 cameras without breaking a sweat.

You need to decide on what goals you want to accomplish with the cameras.
Cameras need to be chosen for what they will be used for.

 

[wittaj]

+1 above.

Do you mean you were using Blue Iris? If so, like mentioned 25 cams on a 12th gen is easy. People run 50 cams on a 4th gen.

If all you need are human and vehicle AI, go with the camera AI and leave CodeProject out of it.

The true test....I have found the AI of the cameras to work even in a freakin blizzard....imagine how much the CPU/GPU would be maxing out sending all the snow pictures for analysis to CodeProject LOL. My non-AI cams in BI were triggering all night. This picture was ran through AI (without the IVS or red lines on it) and it failed to recognize a person in the picture, but the camera AI did. This pic says it all and the video had the red box over it even in complete white out on the screen:

See this thread on how using just Dahua AI may be sufficient for your needs:

Who uses Dahua AI capable cameras? Reliable AI for triggering events? Pro's/con's?


Unless your subjects are all within 15 feet of the camera, 25 cameras of all fixed lens cameras will fail you.

See this thread for the commonly recommended cameras (along with Amazon links) based on distance to IDENTIFY that represent the overall best value in terms of price and performance day and night.

The Importance of Focal Length over MP in camera selection