Blue Iris

Hi Everyone,

New user here. Pretty familiar with VMS software. I don't need to record any of the streams. We have (3) Luma NVRs that handle all the recording and storage.

I want to use BI5 to compile all the cameras onto a single monitor - and then trigger groups based on motion.

We currently have 24 cameras. Cameras range from 3mp to 8mp. I have set the frame rate on all the cams at 5FPS.

I have a couple of old WIN10-64bit machines with SSD drive. One is an Intel I5 with Intel HD4000 graphics card. I started testing with it last night. But once I get up to about 12 to 14 cameras - the system just starts fumbling. And even if I disable half the cameras it still stumbles.

The network is all Ubiquiti, everything is running 1gig. The switches that the cameras come to that link to the NVR's are 10gbs linked.

I'm not one of those folks that just asks a question without looking first. I read the other threads and I also looked at the CPU table several times. But since I am not going to record any of the streams via BI5 I wasn't sure how big a system I need to build.

The goal is to end up with 30 cameras - Output onto 4k monitors.

Thanks for taking the time to share your knowledge and experience.

Dave B

A couple of points...

• BI uses a decent amount of CPU processing to decode the video streams and perform its trigger detection. Which generation I5 are you using?

• If using a fully licenses version (ie not the demo) and you are using a CPU with QuickSync capability you can use Intel Hardware Acceleration.

• There is a list of voluntary submissions of various systems and how many pixels/sec they are all processing that might be helpful https://biupdatehelper.hopto.org/default.html#stats

• When you say that you want to Trigger Groups upon motion detection are you saying that you want to change the displayed cameras based on detection of any camera within that group? I am not sure about this but others may comment if you clarify.

• Make sure that you are NOT recording because the default recording method involves re-encoding all of the streams before writing to disk. This can 2x your CPU usage.