Anti-Dither, pre record, and record latch

[vector18 1]

I really want to tweak my recordings to perfection and I've been playing around with these settings lately.

To my belief, anti dither is the amount of seconds of motion until the dvr decides it is a motion even. For example,

if I set anti dither to 5 seconds, there has to be motion for 5 continuous seconds than the recorder says this is a motion

event definitely. I would like this set to like 2 seconds just in case someone in the image moves than stops right away.

 

Pre-record to my belief is the amount of seconds the recorder actually records before the motion event. For example, if the recorder senses motion and you have pre record set to 7 seconds, when you play back, you will see 7 seconds of non motion, and than the motion event. I would like this between 4 and 8 seconds so when you play back, the event doesn't just pop onto the screen.

 

Record Latch to my belief is the amount of seconds the record keeps recording after the motion has stopped. I would like this for about 10 seconds just so the playback doesn't just end abruptly.

 

What does everyone else think and what do you set yours for? Let get this discussion going.

[shockwave199 0]

It depends on how a given dvr, nvr, or IP camera handles it, and how well it handles it. My qsee 408 dvr has auto 10 second pre record built in- it can't be adjusted. It works flawlessly all the time for my home application. I set post record the shortest allowed in there- 30 seconds. That is on the longer side but it too always works out for the best. The only thing I have to control what consitutes a motion event is sensitivity. I don't have a anti dither option in that dvr.

 

The dahua nvr install I did a couple months ago is entirely different. I have all the otions you mention in the nvr, which I suppose is really tweaking the cameras when you set it all up through the nvr. For a business, each of the four locations needed different things. The reception area is so busy during the day that it proved useless to do motion detection. So during normal business hours, it records fulltime. At closing, motion takes over. That camera, I have AD set at 5 seconds, PR set at 10 seconds, and latch set at 10 seconds. Weekends that camera is set for MD full time. The other three parking lot cameras are a totally different story because with a parking lot, it's important to see when vehicles enter and exit as well as any action that might go on closer to the cameras. So, since I found dahua's sensitivity somewhat unreliable with super large fov's, I keep all those cameras sensitity max at 6, AD at 5 seconds for all, but pre record is 30 seconds for each. This almost always backs the event up to when a vehicle enters way out in the front of the lot. I have latch on all those set at 20 seconds, which helps to cut down on the amount of events somewhat so it's less broken up- but still taking advantage of events.

 

This gets trickier at night, when the vehicle/pedestrian traffic gets way less and you can really see how well all the settings are balanced out. And of course, it can't be 100% all the time with MD settings. Some events have been missed or chopped too short up front or at the end. So in order to not miss a single thing ever, I had to schedule full time recording on two cameras out there- one during the day and then that one goes to MD and another takes over full time all night. So all day and night, at least one camera is recording full time while the others do MD. That really gets everything and with MD events as well, it makes for easy searching. So much critical stuff can happen at any given moment in public parking lots, you really have to dial it all in right- so you get everything but still save as much hard drive space as you can.

 

How's that for discussing! LOL!

[vector18 1]

That's definitely a good starting point for this discussion. I just don't trust a 5 second AD time in any scenerio though. If the camera is set to MD and 5 second AD, than a person or car has to be in the FOV for 5 seconds. What if the person just walks past

the camera and is less than 5 seconds? I realize someone might not want or need to record someone just walking past, but I think it should be recorded. I do like the long pre record so you don't just get someone popping up in the middle of the screen and ask how did he get there? And I agree with you that a high traffic area should get constant recording. It makes a nicer playback rather than bits and pieces all day long.

[shockwave199 0]

Can AD be set to less than 5 seconds in the dahua? Mine says 5~600.

 

At a normal walking pace for me, I walk about 12' in five seconds. With wide 1080p fov's, I never seem to have a problem getting everything with a default 5 seconds AD. Sensitivity seems to be less consistant the bigger the fov you're dealing with. You have to test and check that carefully.

[icamera 0]

Interesting.

 

Anyone know of which VMS can support simultaneous live recording while still recording motion.

 

This way you have both the the video and the motion events. Set motion events to store for 9 months for example and overwrite live recordings as space needed....

 

Also any VMS besides Avigilon have the pixel search function so you can mask the area you want to search for video events?

 

That's definitely a good starting point for this discussion. I just don't trust a 5 second AD time in any scenerio though. If the camera is set to MD and 5 second AD, than a person or car has to be in the FOV for 5 seconds. What if the person just walks past

the camera and is less than 5 seconds? I realize someone might not want or need to record someone just walking past, but I think it should be recorded. I do like the long pre record so you don't just get someone popping up in the middle of the screen and ask how did he get there? And I agree with you that a high traffic area should get constant recording. It makes a nicer playback rather than bits and pieces all day long.

[vector18 1]

Can AD be set to less than 5 seconds in the dahua? Mine says 5~600.

 

At a normal walking pace for me, I walk about 12' in five seconds. With wide 1080p fov's, I never seem to have a problem getting everything with a default 5 seconds AD. Sensitivity seems to be less consistant the bigger the fov you're dealing with. You have to test and check that carefully.

 

Yes, my nvr has 0-600 and I was concerned about my PTZ when fully zoomed in. You have a small FOV while zoomed in tight.

[vector18 1]

 

Right now, it's not that zoomed in, but I can get a pretty tight shot of my truck if I want. But the more and more I think about it, I think I'm better off zoomed out slightly so I can get more information if something happens to it. But in this NVR and I checked, all other ones I have, AD can be 0-600.

[lowpro 0]

I have a QC808 NVR and I hated that I wasn't able to go below 5 seconds on any of the settings. The nvr would just revert back to 5 seconds. But I found out that via web access into the nvr, it allowed me to change AD to 3 seconds. It appeared to have saved the setting. Later that evening I went back to check the setting on the nvr and it was reporting 5 seconds for the AD. So i don't know if it really took the setting or not.

 

MD was really unrealiable for me at home. Since then i've just installed 2 Optex motion sensors outside and have been waaaaay happier.

[shockwave199 0]

Can AD be set to less than 5 seconds in the dahua? Mine says 5~600.

 

At a normal walking pace for me, I walk about 12' in five seconds. With wide 1080p fov's, I never seem to have a problem getting everything with a default 5 seconds AD. Sensitivity seems to be less consistant the bigger the fov you're dealing with. You have to test and check that carefully.

 

Yes, my nvr has 0-600 and I was concerned about my PTZ when fully zoomed in. You have a small FOV while zoomed in tight.

I wonder what the heck the difference is? The one installed is a 4 channel rebranded dahua, but I wonder if it's a firmware update or something. And if you have the time, you should really put the settings to the test to see if they're working properly all the time- especially the AD. For instance, set 5 seconds and see if you can fake it out so it doesn't record. At 5 seconds, I don't seem to miss a thing. If I get a chance, after work maybe I'll swing by the place by me and try as well. It would take running to actually make it through a fov in less than 5 seconds. Or, standing in front of a camera and not moving so the event ends, and then doing a quick hand wave and then be motionless again. Hmmmm.