Help with troubleshooting artifacts? on my 4K-T180s

[Ri22o]

I made some updates to my sky rig which involved turning the whole thing 90*. In doing so it changed the scenes and I am seeing artifacts I didn't see before, possibly due to lighting. Any thoughts on how to troubleshoot if it's a setting issue or a hardware issue?

I am seeing this with three of them, and it just happens that three of them were refurbs, so maybe it was an existing issue, but not sure. I thought maybe it was BLC, but the issues persisted with the camera when moved to a new scene.

BLC is ON.
All other settings are as expected.





I am seeing no issues with this camera. It is what originally pointed West. I see no issues looking back at those captures.




If you look at my roof, the houses across the street, and the tree line you can see artifacts in the left camera. This camera was originally pointed North. The capture below this one is from when it was pointed North.


You can see how the left lens is very heavy with dots and the right side shows only a few stars.




In the right lens you can see artifacts all of the way down through the tree line and onto the street. This likely had artifacts when it was originally pointed South. The camera at the top that was originally pointed this direction had no issues.


This is when the camera was pointed South, much like the North capture, there are a lot of dots in the right lens but not many stars in the left lens past the splice.




This one is more difficult to see, but there are some on the house in the lower right. Looking back at when this one was mounted facing East, and actually looking for them, there were some visible but a bit more difficult to see because of the lighting in the scene.

 

[wittaj]

You found UFOs

That looks to me like typical reflective glare off lights and the way it happens to be hitting the lens and protective "glass" in front of the lens.

And some of the dots are pixels getting a little overzealous trying to produce a bright image.

You just have to adjust the camera to minimize it. I have that in one of my views as well.

 

[Ri22o]

You found UFOs

That looks to me like typical reflective glare off lights and the way it happens to be hitting the lens and protective "glass" in front of the lens.

And some of the dots are pixels getting a little overzealous trying to produce a bright image.

You just have to adjust the camera to minimize it. I have that in one of my views as well.

I understand the glare from the lights, but what I am talking about, and I should have circled some of them, is the white dots appearing where there aren't stars. These dots moved with the cameras from the previous direction to the current direction, even if the camera in the previous direction had no issue or an issue on the other lens.

West camera is now pointed North. No issues when pointed West and no issues pointed North.
North camera is now pointed East. It appeared to have dots in the left lens when pointed North, and still does now that it is pointed East. The camera pointed North has no issues.
South camera is now pointed West and has dots in the right lens. The original West camera had no issues, lighting and darkness hasn't changed.
East camera is now South, and has some of the same dots, noted below.

Here are captures from when the camera was originally pointed East.




Here is the same camera now pointed South. If you use the two circled dots as reference you can find some of the others to still be there. Should this be expected if it is looking at a completely different scene?

 

[wittaj]

Yeah those are just rogue pixels that got overzealous trying to create an image in a dark background. There are several posts of people having the same experience in dark colors.

You can prove they aren't dead pixels by pointing the camera to say an all green wall or something and you will see they then take the color.

My grass in the backyard has a few of those where the image is too dark.

 

[guykuo]

Sadly, those are white pixels from the sensor, just as Wittaj says. It's an unfortunate aspect of pushing the sensor to the max of their sensing capability. Some of the pixels will simply be bright when gain and exposure are high enough. Even astronomical cameras suffer from this type of artifact. In that usage case, the astronomer use cooled sensors and take some completely black field images to intentionally capture those bright pixels. Image processing software uses those black images to subtract out the bright pixels. We don't have that luxury for our live video feeds. You can drop the exposure and gain to drive some of them back into black, but then you lose some dimmer features.

Interestingly, you may be able to reduce those pixels that are "stuck" as white by increasing gain until the dark regions of image have some noise. The noise essentially erases the errant white spots at the cost of some pixel noise. Then drop brightness in image settings do push that back down into black. I know this is counter to our usual tactic of decreasing gain to reduce noise, but adding some noise obliterates most of the stuck white pixels that accumulate. Give it a try and see whether that is helpful for your situation.

 

[guykuo]

It is a clear night so I spent some time with my 4KT to investigate the white pixels in dark areas issue. Whereas cranking up the gain to obliterate them with noise does hide them, that also hides stars.
So, I tested changes in gain to see how that affected things. With gain lowered, the white dots don't accumulate in dark regions of image. It turns out, just two steps of gain can make the difference between the white pixels being accumulated vs not appearing.

On my particular camera, a gain of
52 avoids the white pixels.
53 creates very dim white pixels
54 or higher creates very bright white pixels in black regions of image.

I think it is a good exercise to single step up gain from too dark to find the sweet spot.

Take a look through these. The stars jump around as you step through the gains because I didn't capture them sequentially as I initially made big steps in gain and finally understood I needed to single step to find the highest gain at which white pixels were not being accumulated. The best setting may vary between cameras and ambient lighting. My best may not be the best for you.

Also, I'm not attempting to make the scenery visible in this sequence, Just focusing on getting the most stars without adding white pixel artifacts.

Gain 40


Gain 46


Gain 48


Gain 50


Gain 51


Gain 52


Gain 53


Gain 55


Gain 57


Gain 60

 

[EMPIRETECANDY]

That white dot is the gain from the camera at low light place, esp at dark place, not dead pixels.

 

[guykuo]

Here is time lapse star field with gain 52. No white pixels in dark regions of capture. There are still some faint, fixed pixels in the sky, but stars down to magnitude 7 are visibly in motion

View attachment stars gain 52 fast.mp4

 

[CCTVCam]

Agree with Andy. When you increase a gain you amplify every little electronic signal in the sensor. Some of these currents are just background transient currents not signals generated by photons hitting the sensor. At low gain they aren't large enough in amplitude to be processed into a pixel so get ignored. However, when you increase gain you risk amplifying these rogue signals to the point where they exceed the threshold at which they are considered genuine pixels and are processed and become a part of the picture. The only answer is to reduce the gain until they disappear (or live with them if you need such levels of gain to get a usable picture).