Oh no..Another motion detection question!

General discussion about Blue Iris
Post Reply
BlueRIdge
Posts: 19
Joined: Thu Dec 26, 2019 3:58 pm

Oh no..Another motion detection question!

Post by BlueRIdge »

What could cause BI to completely miss a motion event?

I’m a long time user, with motion detection working well. I use zone crossings and have things tuned up well.

Occasionally, BI will completely miss a motion event. Identical zone crossing events happen. BI at times will completely miss one, after recording the exact same motion hours earlier. I’m certain that zone motion is set up properly, as the exact motion will record 99% of the time, but be missed very occasionally.

An example is one event recorded when people walk across zones. A few hours later, the same motion occurs, with even more people walking across zones for a longer time, however the second event is completely missed.

I’ve got no time exclusions via schedules, so BI is always active. My server is set to never sleep. Any ideas what could be causing missed motion events?
User avatar
Thixotropic
Posts: 744
Joined: Wed Sep 04, 2019 7:20 pm
Location: Low-Earth Orbit

Re: Oh no..Another motion detection question!

Post by Thixotropic »

What happens if you take one of the clips where it didn't trigger and run it through the "Test run video through motion detector option"?

Does it see the motion and trigger or does it miss it? Run it repeatedly to see if it misses or hits every time. If it's not consistent, something's wrong.
Blue Iris 5.x x64 | Windows 10 Pro x64 | 16GB RAM | i7-7700 3.6 GHz | 1TB HDD | 2TB RAID NAS | 9 Cameras | Almost Dual NIC | 2KVA UPS
HeneryH
Posts: 692
Joined: Thu Jul 18, 2019 2:50 pm

Re: Oh no..Another motion detection question!

Post by HeneryH »

Are the missed ones within the time window for ignoring events within x minutes of each other? I'm not at my screen at the moment so I forget the exact name of the setting.
BlueRIdge
Posts: 19
Joined: Thu Dec 26, 2019 3:58 pm

Re: Oh no..Another motion detection question!

Post by BlueRIdge »

I've got no clip to run through a test trigger option - it didn't trigger, therefore no recording. I've tried repeating the same motion by walking through the area and it will trigger OK.

The missed motion was hours after a previous trigger, so not subject to any retrigger time.

It has me stumped. The same motion will trigger 90+% of the time, yet occasionally miss.
atreyu
Posts: 56
Joined: Fri Nov 27, 2020 7:22 pm

Re: Oh no..Another motion detection question!

Post by atreyu »

Some quick things to check:
  • uncheck “limit decoding unless required”
  • ensure your camera has a decent FPS
  • I can’t remember the setting but it’s something about the interval to motion check frames
  • The there is playing with the trigger contrast, size, etc
I struggled with this exact scenario when I first got BI. It ended up being the limit decoding unless required. It limits analysis to Key frames and Reolink has 0.25keyfps so it was only analyzing a frame every four seconds. It my cpu usage was great :D
BlueRIdge
Posts: 19
Joined: Thu Dec 26, 2019 3:58 pm

Re: Oh no..Another motion detection question!

Post by BlueRIdge »

atreyu wrote: Mon Dec 21, 2020 10:51 pm Some quick things to check:
  • uncheck “limit decoding unless required”
  • ensure your camera has a decent FPS
  • I can’t remember the setting but it’s something about the interval to motion check frames
  • The there is playing with the trigger contrast, size, etc
I struggled with this exact scenario when I first got BI. It ended up being the limit decoding unless required. It limits analysis to Key frames and Reolink has 0.25keyfps so it was only analyzing a frame every four seconds. It my cpu usage was great :D
Thanks for the suggestions. I've had "limit decoding" unchecked already. Last missed motion was over 30 seconds of motion all over the view field, eliminating a lot of variables. I'm starting to wonder if my server is at fault. I believe that I've disabled all Windows 10 sleep/power save functions and am saving clips to an SSD drive.

Odd problem. As I've said earlier, missed motion is very infrequent, once or twice per week. In all cases, there was obvious motion that should have been easily detected. There have been cases where missed motion was identical to earlier motion that did trigger. I've considered recording continuously and examining the missed triggers, but it would take days of recording, as misses are very infrequent.
Post Reply