I will in my "spare time"... Ive got a bad taste in my mouth about BI right now, and am going to let it lie for a few days before I pick it back up.
Failed Motion Detection - Car broken into...
- Thread starter corkangel76
- Start date
fenderman
Staff member
- Mar 9, 2014
- 36,892
- 21,408
Its impossible to make glitch free software ...I can tell you from experience with many machines that is simply runs great...
This is precisely why I record 24/7. I actually learned this when a neighbor put a sign on my car one night and didn't catch it. He did it for fun and it was a joke, but it was a good trial run.
Recording on motion is nothing more than a convenience to make it easier to see what went on, but it is a major security hole due to unreliable detection. I have my cameras recording 24/7 but I also have motion detection to take images. I can use these images sometimes to help find out when an event happened (I don't use any video recording software--just what's on the hikvisions), but they are only a supplement.
No need. I live in a very safe area with essentially no crime, and even in the decade in this spot we've had cars broken into one night (a bunch of them), and then again one car a couple of years ago for a stereo. In all cases cars were opened if unlocked (including mine hehe); no windows broken. I just assume they were teens who live in the area looking for quick cash. There is a major difference between going into cars looking for spare change and entering homes, and the latter hasn't happened and is the only thing that really should worry a person.
As far as software is concerned, in theory it's not possible to make perfect software, though for its practical use it can be done. Case in point windows calculator. I'm sure if you spent enough money on it you could find a bug and cause it to create a bad result, but it works for what we need.
I do not use BI, I use Exacq and it uses the motion detection built into the Hik cameras and it works great. It is in fact the best motion detection on any of the cameras I have, though it is almost overly sensitive as I have to keep the motion detection set to "1" in Exacq and it works perfectly. Once you get 100 feet away is is more tricky I suppose for people on foot but it still does really well.
I have no complaints, and in fact cannot wait for the 9 3mp ACTi cameras at my house to die so I can replace them with the more sensitive Hiks. They may well be "broken" here soon regardless....
I have no complaints, and in fact cannot wait for the 9 3mp ACTi cameras at my house to die so I can replace them with the more sensitive Hiks. They may well be "broken" here soon regardless....
Last edited by a moderator:
Thank you for sharing your story with us, @corkangel76. I have, since reading your post, set all of my cameras to record 24/7 since storage is so cheap these days. Now BI does 24/7 and marks the events by motion in the timeline allowing for me to easy comb through events, but never missing a beat.
Last edited by a moderator:
Totally sucks that, sorry to hear it.
My cameras are always slightly too sensitive rather than being slightly under sensitive. If you have no false motion detections then you're also more likely to miss a real motion detection event. It's easier to fast-forward through one or two extra false positives a day then playback an event that wasn't recorded!
My cameras are always slightly too sensitive rather than being slightly under sensitive. If you have no false motion detections then you're also more likely to miss a real motion detection event. It's easier to fast-forward through one or two extra false positives a day then playback an event that wasn't recorded!
Rather than cloning the camera, on the Record tab, set Video to Continuous, and on the Motion/Trigger tab, enable and configure the motion sensor apropriately. My i5-4690K with 4x1080p, and 2x720p, is sitting at <50% CPU, recording all cameras 24x7, with me in the BI console, connected via Teamviewer, and a rain storm triggering all of the exterior cameras pretty much continuously.
When I close the console and drop the Teamviewer connection, the Android app shows CPU at ~18%.
Last edited by a moderator:
If you were correct then an assembler level register add command would occasionally fail to do 1+1=2, it does not because the on chip microcode function that the assembler calls has been created and tested to be error free. The same can be done to assembler and higher languages, it will be expensive and it might be impracticable but it is not impossible.
fenderman
Staff member
- Mar 9, 2014
- 36,892
- 21,408
A human still has to account of every practical possibility in a complex system. There are simply too many permutations to test every single one. It would take longer than your lifetime. Regardless, it has never been done to date. The point is it does not exist in the realm of VMS software.
Your statement was that it was impossible to make glitch free software all I was doing was pointing out to you that the statement was false.
The thing about glitch free software though is like proving a negative. The best you can do is test for all scenarios you're able to think of, but until all permutations possible have been tested you will never be able to bet everything on it. I wouldn't bet $10 vs the end of the known universe that the add command mentioned above could not result in a 3 under a certain scenario; some new variable introduced that hasn't been considered.
Go and research dependable computing and then you will understand how to construct error free code and how there is enough time to test.