Security camera quality issues, what's wrong?


Staff member
Ok, this is something that has been bugging me for a long time now, and I want to know why this is happening. As you know, I am running my camera (VisionTech cam from Martin, really nice unit) over cat5e, probably a 175ft run, and I run power over that wire as well. The vido feed is pretty good when things aren't moving, but once there is a moving object, then the moving object itself will look pretty bad. Some examples:




Is this normal, or is there a technical issue. Thanks!
I am wondering if there is a type of compression going on where only the changed pixels actually get sent over the Cat5. With motion, there would be more pixels and possibly a processing or bandwidth issue to redraw all that have changed. Therefore, poor resolution.
I'm assuming this doesn't behave this way when looking directly at the camera over a short coax run then?

I agree with DavidL. I think you are experiencing bandwidth limitations of the balans. Not sure about this as this is not my area of expertise though! ;)
I'm not sure if I see the problem. Are you talking about the steps in objects that move horizontally? The stair-stepping on the car's spoiler in the first picture, and the stepping between the bus's front tire and fender?

If so, you are seeing an artifact of interlace. The even scan lines are displayed 1/60 of a second after the odd scan lines, so things appear moved by that amount.

You can test this easily. Get something with high contrast. I test with a large white board with black vertical 3-inch lines (created with black tape). Move the board horizontally, and you should see the scan lines. Move the board vertically, and you won't.
Originally I figured it had something to do with interlacing, but this defeats the purpose of having a snapshot of whenever something moves, since I can't see many details. I haven't seen anyone else mention this issue, so figured I would post. I'll have to do some more testing I guess, thanks for the tip.
There is a tradeoff between single-field and full-frame captures, as Ski's post illustrates perfectly.

Motion is a problem when capturing a full-frame. But when capturing just a single field, you are discarding every other scan line in your image, thereby lowering the vertical resolution of the still objects in the image.

Compare the banding of the rack in the background, or the ceiling-to-wall joint behind it.
As Ski showed, it's possible, it's just that you give up details when doing this. I use Active Webcam, but settings are driver related, not application from what I can tell. I decided to leave it as it is.