What would cause BI to do this?

[Rockford622]

I have two HikVision DS-2CD3332-I cameras on the front of our house, one on the north end and the other on the south end. Both basically view the road but from different angles. Here is a short clip caught by camera A (north) that worked perfectly:

You see a car from the left, and then a van from the right, and a black car in there as well. Now here is the video from camera B:
BI got the car coming from the left, but didn't even see the van until it started to move from being stopped, and totally missed the black car in there...

Both cameras have the same settings, .5s make time and 10s break time, 30 MB camera buffer. The entire field of view is Motion A on both...

I just don't get it. Most of the time BI works perfectly, but every once in a while it just fails to see motion.

Anyone else have issues like this? I find it hard to trust this software when it does stuff like this.

 

[fenderman]

I have two HikVision DS-2CD3332-I cameras on the front of our house, one on the north end and the other on the south end. Both basically view the road but from different angles. Here is a short clip caught by camera A (north) that worked perfectly:

You see a car from the left, and then a van from the right, and a black car in there as well. Now here is the video from camera B:
BI got the car coming from the left, but didn't even see the van until it started to move from being stopped, and totally missed the black car in there...

Both cameras have the same settings, .5s make time and 10s break time, 30 MB camera buffer. The entire field of view is Motion A on both...

I just don't get it. Most of the time BI works perfectly, but every once in a while it just fails to see motion.

Anyone else have issues like this? I find it hard to trust this software when it does stuff like this.

The software is extreamly reliable. Most of the time is a user setting that is incorrect. Post BOTH motion and record settings - remember these are profile specific.
What is your pretrigger time set to?
Is the iframe rate set to match the fps in the cameras?

 

[Rockford622]

Iframe and FPS are set to 20 on both cameras. Pretrigger frames is 40 so I get 2 seconds of recorded time before each alert. Max bitrate is 4096 and variable bit rate on both cameras.
Here are the settings for camera A:



And for camera B:

 

[fenderman]

What are your zones set to?

 

[Rockford622]

Here you go:

Camera A:


Camera B:


Both use just Motion A, and there is a little square in Camera B that I don't care about because there is a little flag there that waves in the breeze (when it's not buried under snow).

 

[fenderman]

Video is A is marked private, can you make it viewable?

 

[Rockford622]

Apparently I forgot to publish it...You should be able to view it now.

 

[fenderman]

Apparently I forgot to publish it...You should be able to view it now.

Is the video jerky as it appears, when you play it back on your pc?
What version of BI was running when this occurred?

 

[Rockford622]

I would have to say yes to that. A tad bit worse in the YouTube verison (I think just because of the compression, but the jerkyness is about the same), but fairly representative. Should it be smoother at 20 FPS?

I currently have the latest version, 4.3.0.7 x64

 

[fenderman]

I would have to say yes to that. A tad bit worse in the YouTube verison, but fairly representative. Should it be smoother at 20 FPS?

Yes, its should be smoother....do you see that in the live view?

 

[Rockford622]

Live view I have set to 30 FPS...I just changed it to 20 FPS to match the cameras so I will monitor it. I would say the live view while it was at 30 FPS was very smooth.

Also, when I experience this type of effect, I generally see something like this in the alert list:


You can see a truck drove by, and I have highlight motion turned on, but the highlighted motion almost leaves a shadow behind. Generally just the object that is moving is shown highlighted in the alert list. Not sure if this helps or not.

 

[fenderman]

If the camera is set to 20fps, then setting the live view higher will not have any effect in blue iris.

 

[fenderman]

What version of blue iris were you running when this happened?

 

[Rockford622]

The latest, 4.3.0.7 x64. Things like this have happened with earlier versions as well.

 

[fenderman]

The latest, 4.3.0.7 x64. Things like this have happened with earlier versions as well.

When you say earlier, do you mean before 4.3? There were new changes introduced in 4.3 that may be causing this, though i havent seen anyone else with hikvisions report this so it may be the camera or network. I would set it to record continuous, with motion triggers and see if the video freezes. Then you need to figure out if its freezing in blue iris or the camera is sending corrupted video.

 

[Rockford622]

Yes, prior to 4.3...before the new implementation of motion zones.

We had an incident that happened at my neighbors house (to the left from the perspective of my cameras above). He had his rims and tires stolen off of his car at 2:20 am one morning at the beginning of this past February. Once I checked my video footage, I found that I did have some good video that showed partially what happened. The criminal parked his car in front of my neighbors house and walked across my sidewalk and driveway (from left to right, then back right to left) because he used some of my landscaping bricks to prop up the car while the rims and tires were removed. My Camera A as I call it in the above posts, captured him first walking left to right, then back right to left with a load of bricks. My Camera B caught NOTHING, even thought it points at the same sidewalk! He walked right in front of it and it got nothing at all. I am amazed at this. So, because of this I started to pay a little closer attention to what one camera captures vs. what the other one captures. Both of the cameras had basically the same settings on the cameras and in BI. This has lead me to the conclusion that they both miss various things when I compare one to the other. I need to figure out what is going on here.

I suppose I can try continuous record and see what that tells me.

 

[fenderman]

Yes, prior to 4.3...before the new implementation of motion zones.

We had an incident that happened at my neighbors house (to the left from the perspective of my cameras above). He had his rims and tires stolen off of his car at 2:20 am one morning at the beginning of this past February. Once I checked my video footage, I found that I did have some good video that showed partially what happened. The criminal parked his car in front of my neighbors house and walked across my sidewalk and driveway (from left to right, then back right to left) because he used some of my landscaping bricks to prop up the car while the rims and tires were removed. My Camera A as I call it in the above posts, captured him first walking left to right, then back right to left with a load of bricks. My Camera B caught NOTHING, even thought it points at the same sidewalk! He walked right in front of it and it got nothing at all. I am amazed at this. So, because of this I started to pay a little closer attention to what one camera captures vs. what the other one captures. Both of the cameras had basically the same settings on the cameras and in BI. This has lead me to the conclusion that they both miss various things when I compare one to the other. I need to figure out what is going on here.

I suppose I can try continuous record and see what that tells me.

Yes, you might find there is an issue with the video its providing. Also note that you cannot directly compare motion settings because the target appears different in size to each camera based on its mounting location/angle and the distance from the camera. The freezing of your video every second or so leads me to believe there is something else going on.

 

[Rockford622]

Thanks for your help so far Fenderman. Here is an example of a continuous record from this morning:

There was an alert generated and here it is:


When I double click the image, it take me to when the car is where you see it above. Generally it would take me to a few seconds before the actual alert because of the pre trigger video buffer. Is that because I am doing a continuous record so it take me to the actual alert, and not a few seconds before?

You can see the video still looks laggy and choppy.

Not sure where to go from here.

 

[fenderman]

That is normal. When on continuous it will take you to the actual alert not the pretrigger. You need to wait until there is video that is actually missed. A better way to test would be to setup cloned cameras. One on continuous the other motion.

 

[Rockford622]

That does make sense.

I did export my 2 front camera configs and re-import them as a continuous record (changed the originals back to record on motion). The only odd thing was I had to stop and restart the Windows service before the 2 new continuous cameras would create new clips. No big deal, just took me a few minutes to figure out.