Then I better use record only and delete data after some days if nothing happens. And about acceleration use? I read something that blue iris is using it and is a huge improvement. Can motioneye use it? I can`t find something about it.
Motion detection analysis video frames so this will drive up CPU usage. Recording actually doesn’t do much in comparison.
Reduce camera resolution at the camera will decrease CPU
Reduce camera frame rate will reduce CPU
NO. I don’t believe acceleration available for motioneye.
I try blue Iris and meh!!!
windows is maintenance/security worry for me and I don’t feel it ultimately worth it. Also general software complexity was unnecessary. I like motioneye simplicity.
I actually find that recordings where not very useful so I decide to use image with reduced recording(less sensitive or controlled by some external automation)
I use Doods to capture images using image processing.
I found this captured more relevant images (even if many useless images captured). Images take up less space as well so added benefit. I save images from Doods into motioneye image folder for review within motioneye. To do this It required I set motioneye to capture images each 10minute to ensure date-time folder created
Image processing work very well in low traffic area. For example, I have camera at front door that not point to publicly walk/street and it perfectly capture relevant event. I have another camer facing driveway entry/highway and it capture many irrelevant image(but many still usefull).
I use motion recording also but lower the sensitivty and place more dependency on image. In future I may disable motion record and trigger recording using motion sensor or Doods as trigger
Yesterday I add Coral TPU and huge improvement. I have beefy server (dual i7/>70GB ram) but old (2011cpu) so CPU is not concern but video processing it has poor efficiency. I test in recent 2015 i5 laptop and Doods was working well. I think this is use case dependant but may prove useful for you
EDIT
I did a little quick math and came to folowing
(1)photo in motioneye average <200k
(86400) second in day
approx 18G to capture photo each second of day
(1) video file I estimate <1mb per minute. I see 7minute file at 5mb
(1440) minute in day
< 2GB to record camera 24hr
storage is super cheap and based on this I plan to likely set most camera to record 24hrs with 5minute max lenght recording.
I will still used doods for image capture and as an easy way to catch the important event. This will allow quick review of photos for events then go to video for more detail of occurance. Even with many camera I feel this will easily support 30 day file storage if using 2TB drive. even (10) camera I expect less than 20GB/day or 600GB/month. maybe even longer storage time is possible
Thx for insight. I`m thinking to buy Coral I think with my nas will be a nice improvement. I think is the best solution for now.
I made mistake. actually 5minute file is 200-250mb so 30 day is almost 9tb for single camera.
Maybe continuous snapshot with trigger recording may be best
Try Shinobi for continuous recording, far better than MotionEye, less CPU and RAM use.
Shinobi is also great.
I set “movie quality” to 50% and size is down to 50MB or less for 5 minute recording and the quality is more than acceptable for 2mp camera capturing subject within 30’
After many tweaks and settings I still have a huge cpu load with motioneye even with recording and motion disabled so tonight Im gonna try Shinobi and ZoneMinder dockers to see how it goes and if still Im not happy I have to test Blue Iris and MIlestone XProtect Essential. My last resort will be synology surveillance station but the license is quite expensive for 6 cameras what I have now.
This may be hardware issue.
Honestly all are pretty much same resource wise.
EDIT
Actually you need to limit it CPU access. This will be true for them all or else it will eat of available resources unnecessarily
I’m assuming you run docker sontry adding
`--cpus=<value>` Specify how much of the available CPU resources a container can use. For instance, if the host machine has two CPUs and you set `--cpus="1.5"` , the container is guaranteed at most one and a half of the CPUs. This is the equivalent of setting `--cpu-period="100000"` and `--cpu-quota="150000"` . Available in Docker 1.13 and higher.
`-m` or `--memory=` The maximum amount of memory the container can use. If you set this option, the minimum allowed value is `4m` (4 megabyte).
Honestly I forget which variable to set as I now use portainer to set these values
So I have tried ZoneMinder and I don`t like it. Shinobi is awesome. All cameras on and my cpu is 2% and ram 400mb. With motion eye I have tried to limit cpu/ram but still is killing mu nas cpu.
After removing stream: from hass I have same high cpu load meaning that stream: is using ffmpeg and without is going to mjpeg. If I understand this right motion eye is failing to use ffmpeg ? Shinobi is streaming so smooth with a cpu usage very low but I can`t integrate it with hass and the hassio addon is no longer maintained.
for motioneye: passthrough increases CPU addon usage - but here’s the thing - it doesnt really affect the overrall system CPU usage ( maybe a bit ) so question: there a way to aassign more CPU cores or something to the addon, just seems like its only grabbing 1 core or something? Thoughts on this ?
You can set CPU Core usage at build time in docker. Memory and other parameters also.
addons don’t provide for that option. that just build with some default parameters I believe
okay gooood to know thanks! - docker supposedly defaults to run unlimited so i am guessing either HASS - when clicking that addon switch enable limits it when it runs the docker run command …Or (and i am not sure if it is possible) perhaps the addon maker can limit the cpu usage during docker image build
hummm is this normal with 3 cameras?
Seems like motioneye is cpu intensive on my rspb3+ with hassio os installed ?
Hassio slows down considerably with motioneye enabled… even freezes from time to time.
I am also running into the same issue above, see screenshot by fillppos. CPU usage with a few cameras running ~70% for addon CPU usage, and whole OS is 99%. Pages slow to a crawl when loading anything. Need an answer for how to use this for motion detection without maxing CPU?
Yes
Motioneye is cpu intense especially on RasPi and HD video due to video decoding.
I had older dual core intel w/ 5cams and it ran CPU high.
You could try something like Frigate, especially for motion detection as Frigate allows using lower def video for motion detect but can record without decode. Also it offload object detect to Coral tpu. Currently I run >8 cam with little affect on CPU.
humm seems quite interesting but I can only install frigate through HACS.
I don’t have HACS.
Lost many hours troubleshooting my hassio because of HACS in the past.
Separate RasPi
Integration wouldn’t require hacks if you wanted to use it else mqtt provides everything you need from frigate