Day/Night Thresholds: Do You Find Them Problematic?

I was recently reminded of issues I've had in the past with day/night thresholds. Best example of this: on an office building, 8 cameras were mounted outside, and they practically all switched from day to night at a different point. This is not unusual, because if the sun is going down on the front side of the building, obviously it will be darker in the back. But we found that the two cameras on the back switched to night at vastly different times, even though light was about the same, and both cameras were set to the same settings for day/night switch.

Do you ever find that cameras switch too early or too late to monochrome mode?


Ethan, does any VMS support automatic camera D/N control? I thought we talked about this. It would be a nice feature for a VMS to support and in line with other types of integration.

Your way sure seems like the better method. I like that path-forward, but am not as versed on pushing API events from a central server.

I think for a lot of people running an additional wire pair to each camera is a very big deal. Plus, most IP cameras today do not support serial interfaces so even if you run that cable, you can't connect it to the IP camera.

Doing this via software integration would likely be less costly and physically complex. For instance, the Axis VAPIX API supports toggling the IRcutfilter (see p15 of their event handling API) and the ONVIF Imaging Service mentions a control as well. The software control could be triggered by an input from a controller or importing a sunrise/sunset table.

If your camera has a dedicated connection for day/night switching then all you will need is an additional wire pair to each camera to control the IR cut filter from a central location (although RS485 might become necessary to facilitate long-distance switching). You don't import the tables to the cameras; you use a central controller driving a relay that switches all your cameras simultaneously (the same way you would switch them with a single photocell). I have seen a standalone industrial lighting timer/controller with a built-in sunset/sunrise table that works well in applications like this. I don't have the make/model of controller nearby, but I can look through some old notes this week and try to dig up a recommendation.

There are certainly IP-based devices to switch relays between central and remote locations but they would bring their own issues, cost, and would probably still require a central master timer.

Is it just me because I cringe when suggestions of RS485 connections are suggested? I understand that sometimes they are necessary but they are kind of a necessary evil, especially if you are using IP cameras, where it's a lot of work to add.

I think local sunset tables would be great. How do you import that into a camera or use it to message the camera to cutover?

Ethan has the right idea, here. Many cameras have an input relay or RS485 connection that allow remote D/N switching from a master location. The photocell is optimal because you can set the real-world light level where you want your system to go into night mode without having to mess with camera make/model/sample variation differences.

A secondary (and probably less effective) solution would be to trigger off local sunset timetables. One could trim the data to always go to night mode 30-60min before sunset/after sunrise. However, it wouldn't allow night mode under degraded daytime conditions and would be a more complex logical and electronic connection than just using a master photocell attached to the exterior of your facility.

I have a couple people I could ask!

I've considered testing it, but lighting control is the obvious issue. Also obviously, as we showed in the lowering light levels post, different manufacturers change at different times, and pretty much none of them give any indication what their switch thresholds mean.

If I had my way, I'd like to see day/night switching integrated to VMSs, and then control it with a single photocell attached to an input, instead of letting cameras guess at it.

You should find someone who has a full camera test setup to try this out.... ha ha ha

I am sure there are big variances and, in all seriousness, it would be worth a test. The main thing needed is a way to smoothly control lighting from say 50 down to 1 lux. Anyone have ideas? With that, the whole test wouldn't take more than a day or two.