Blue Iris Integration Tutorial

Tags: #<Tag:0x00007f7c5d3d26e8>

I’d be interested in seeing how well that performs. BI can use quite a bit of resources when you’re dealing with multiple cameras. I’ve got BI hooked into 5 of them, and it’s constantly using at least 25% of my I7 setup running Windows. I don’t know how well that’d translate when you’re going through a virtualization layer running in Docker.

I don’t use Blue Iris, I was just pointing out there is ways to run it on a non Windows machine.

I use shinobi, it uses under 10% CPU usage on an i5 2400, with 4 cameras, running in docker. Far less taxing on the machine than MotionEye or Blue Iris.

I’m trying to get this working and cannot get it to work for some reason. Anyone have any ideas? I’ve tried with limit IP addresses and just removed it so I could diagnose why I can’t get an mjpeg stream. This is what I see in Home Assistant:
image

This is the configuration
image

And Blue Iris
image

image

You can try disabling authentication for Home Assistant. On the advanced tab where it says “Limit IP Addresses”, add your range starting with a + sign, e.g. +192.168.10.*

I assume your configuration is indented correctly and just a bad image.

I use the generic platform like this

- platform: generic
  still_image_url: "http://192.168.1.106:81/image/Garg"
  stream_source: "http://192.168.1.106:81/h264/Garg/temp.m3u8"
  name: Garage Door

This is actually what I ended up doing, this worked great for me. Now on to image processing!

Do you passthrough the GPU to use hardware acceleration? I’m also thinking about using BI on my Win 10 VM on Proxmox, but I’m not sure if my i5 NUC can handle it without passing through the GPU (which is doable but quite a hazzle).

Im using the same setup as RichardU. Running BI on a Lite win10 install. Currently not using GPU passthrough, however I have made the modifications to my Dell Poweredge R510 server to allow me to install a small GPU. Ill let you know how it goes.

I’m not sure how NUC will work without GPU I’m using Intel® Core™ i7-3770K CPU @ 3.50GHz
with 5 1080p cams running 15fps recording constantly and I don’t need a GPU yet.

Thanks for the info! I just installed the test version of Blue Iris in a Win 10 VM (Proxmox as the Hypervisor) on my 8th generation NUC i5. Works like a charm with my 3 cameras (2x 3MP and 1x 4MP). When the Blue Iris app is minimized, but running motion detection on all 3 cameras, then the system is riding along at 30% CPU utilization even though the Win 10 VM is only using 4 of the 8 CPU cores. When it detects motion and starts recording then utilization increases to 60-70%. And this is without GPU passthrough. I plan to add 3-4 additional cameras. Then it might be necessary to passthrough the GPU or at least dedicate 6 cores to the Win 10 VM. However, a newer generation i5 seems more than capable of running Blue Iris.

I’m having a strange problem, and it’s not related to HA at all, rather using MQTT alerts in BI. I’m hoping someone here may have some insight to help me in my troubleshooting quest as I evaluate this as an alternative option to MotionEye

Once my ‘On Alert’ is triggered, an ON payload is sent, shortly followed by an OFF payload. However, after the motion event has cleared, the MQTT alerts continue to send

Here is a snippet after I restarted BI and (eventually) walked through the living room:

2019-10-28 15:42:39.614 BlueIris/status schedule=Default
2019-10-28 15:42:39.617 BlueIris/app starting
2019-10-28 15:42:41.616 BlueIris/app running
2019-10-28 15:42:48.941 BlueIris/status signal=1
2019-10-28 15:43:01.953 BlueIris/status signal=2
2019-10-28 15:43:01.955 BlueIris/status schedule=Default
2019-10-28 15:43:01.956 BlueIris/app starting
2019-10-28 15:43:03.957 BlueIris/app running
2019-10-28 15:43:09.960 BlueIris/status signal=1
2019-10-28 15:43:23.972 BlueIris/status signal=2
2019-10-28 15:43:23.975 BlueIris/status schedule=Default
2019-10-28 15:43:23.978 BlueIris/app starting
2019-10-28 15:43:25.976 BlueIris/app running
2019-10-28 15:43:31.976 BlueIris/status signal=1
2019-10-28 15:43:45.987 BlueIris/status signal=2
2019-10-28 15:43:45.991 BlueIris/status schedule=Default
2019-10-28 15:43:45.995 BlueIris/app starting
2019-10-28 15:43:47.992 BlueIris/app running
2019-10-28 15:43:53.994 BlueIris/status signal=1
2019-10-28 15:44:08.000 BlueIris/status signal=2
2019-10-28 15:44:08.002 BlueIris/status schedule=Default
2019-10-28 15:44:08.004 BlueIris/app starting
2019-10-28 15:44:10.006 BlueIris/app running
2019-10-28 15:44:16.010 BlueIris/status signal=1
2019-10-28 15:44:30.015 BlueIris/status signal=2
2019-10-28 15:44:30.016 BlueIris/status schedule=Default
2019-10-28 15:44:30.018 BlueIris/app starting
2019-10-28 15:44:32.016 BlueIris/app running
2019-10-28 15:44:38.020 BlueIris/status signal=1
2019-10-28 15:44:52.030 BlueIris/status signal=2
2019-10-28 15:44:52.031 BlueIris/status schedule=Default
2019-10-28 15:44:52.032 BlueIris/app starting
2019-10-28 15:44:54.031 BlueIris/app running
2019-10-28 15:45:00.034 BlueIris/status signal=1
2019-10-28 15:45:07.617 BlueIris/living_room/motion ON <--- I walked through the living room
2019-10-28 15:45:13.621 BlueIris/status signal=2
2019-10-28 15:45:13.624 BlueIris/status schedule=Default
2019-10-28 15:45:13.627 BlueIris/app starting
2019-10-28 15:45:15.625 BlueIris/app running
2019-10-28 15:45:18.447 BlueIris/living_room/motion OFF <-- event ended

But the alerts are continuously being sent. This is a snippet taken shortly after I walked through the field of view of the camera, with absolutely no motion (including shadows) afterwards

2019-10-28 15:49:53.820 BlueIris/living_room/motion ON
2019-10-28 15:49:55.820 BlueIris/living_room/motion OFF
2019-10-28 15:49:59.826 BlueIris/status signal=2
2019-10-28 15:49:59.829 BlueIris/status schedule=Default
2019-10-28 15:49:59.831 BlueIris/app starting
2019-10-28 15:50:01.828 BlueIris/app running
2019-10-28 15:50:03.831 BlueIris/living_room/motion OFF
2019-10-28 15:50:03.834 BlueIris/living_room/motion ON
2019-10-28 15:50:15.841 BlueIris/living_room/motion ON
2019-10-28 15:50:17.844 BlueIris/living_room/motion OFF
2019-10-28 15:50:21.846 BlueIris/status signal=2
2019-10-28 15:50:21.848 BlueIris/status schedule=Default
2019-10-28 15:50:21.849 BlueIris/app starting
2019-10-28 15:50:23.851 BlueIris/app running
2019-10-28 15:50:25.852 BlueIris/living_room/motion OFF
2019-10-28 15:50:25.853 BlueIris/living_room/motion ON
2019-10-28 15:50:29.858 BlueIris/status signal=1

This continues to happen until I restart BI, then it will happen again after motion is detected.

Someone told me once that there was a way to display the video or snapshot of the triggered camera on my touch-pad running Kiosk. Don’t remember how to do so or if it was possible…

I was reading that BI was a CPU hog at times. Do you all see that? I’m thinking of getting away from my old Hikvision NVR and going to BI on a SFF-PC.

Thanks,
Ambi

Yes.

I’ve got a newer windows pc with 16gb ram (can’t remember the cpu off the top of my head) and BI usually takes up 30% to 50% of cpu and it eventually hogs up all of the ram over several days of running. To the point i need to reboot the pc about once a week because everything comes to a crawl.

That’s really the worst thing about BI. I love everything else about it.

I don’t see that. I have BI running on a Win 10 VM under Proxmox. I dedicated 4 of my 8 CPU cores of a 8th generation i5 and 8 Gb of RAM to the VM and the VM runs at around 20% utilization. The Intel GPU is passed through to the VM to take advantage of hardware acceleration. All this is currently with 3 cameras (all 4MP), which all have motion detection running. I plan to add additional cameras soon.

Hi,

Thanks for the input. You make me think of another question. I guess performance is at least partially based on number of cameras, resolution, frame rate, etc. So what’s the most number of cameras we can expect from a BI install on a SFF-type machine? (2-4 core, hyperthreading, 16G memory)

Thanks,

Ambi

There is a document on the BlueIris forums with tips on tuning it to consume less CPU. The biggest one is to avoid re-encoding the video (i.e. write the raw video to disk), but I don’t remember the setting off the top of my head.

I have 10 1080p cameras with BlueIris running on an old i7 based box (3.4GHz, 16 GB) and I rarely see it peak above 20% CPU utilization. Those cameras are all doing motion detection, and one is almost constantly recording because of it (it triggers on cars going past).

Any idea where this is as I am very interested in it.
I am running on a i5-4570 3.2GHz 16GB ram.
Have 6 camera’s with one of them is only clone so that I can just do motion detection when someone is walking up the front steps and when it’s sitting idle it’s probably using 45-50% CPU.
When I have 1 or 2 to HA screens showing the camera it’s hovering around 80% cpu.
This pc also has my Plex server, so if someone is trying to watch something the CPU is pegged to 100.
All cameras use the Blue Iris DVR file format with Direct-to-disc.

I am not sure if I need to upgrade my CPU or maybe get a video card to help this run smoother.

Anyone with tips would be appreciated.

1 Like

Excellent, thank you.
I will see if this helps.