strange artifacts from stacking - Forums

Topic starter

Aug 30, 2021

Hello everyone.
I am hoping to get a bit of assistance from people who use Sequator to stack images. I have noticed that when I stack images in sequator I am getting a background pattern appearing in my images. I am shooting with a Nikon Z6II at this time as I am only getting started in deep-sky photography. I shoot the images in RAW and give an initial process to the images in ACR. I save the images as JPEG for stacking in sequator. I typically shoot between 15-30 darks and process the dark with the same adjustments that were done to the light frames. At this time I do not shoot flats or bias frames either. The images are stacked in Sequator and then processed in ACR and finally in photoshop. The individual frames do not show any background pattern but the stacked image does so I know it's an artefact from stacking. Does anyone know If I am incorrectly using Sequator or if something else is not being done properly? The attached image has been pushed a bit to show the background I am talking about. Any help would be appreciated. This is a stack of 70 light frames and 30 dark frames.

Well Written Respectful Engaging

autonm

Aug 30, 2021

You should keep everything in RAW…
Stack in RAW
Process in RAW (regardless of application you use)
…. only at the end finally save out as JPG or PNG.

The red dots, I think may be dead pixels or sensor noise - which is generally the need for Bias frames. Bias frames are very quick to take - fastest shutter speed (same ISO) - take around 100 or them in a minute or so.

The lines or artefacts I think in this instance is 'walking noise'. You dont say how long your exposure time is - however 'dithering' between every 2-4 frames should help remove this. Its difficult to remove post-processing. Without dithering Bias frames may help towards this.

The only other way really to remove (if you cant dither) - is to try and clip the blacks (but you lose data) in processing.

Helpful

Björn Arnold

Aug 30, 2021

Hi Andrew,

A few things regarding stacking: no pre-processing on the light images whatsoever. Each light frame should be calibrated. A typical workflow contains lights, master dark (dark frames to create master dark), master flat per filter/flats per filter, and master dark per flat per filter/darks per flat per filter. Each light frame is calibrated through a master dark and master flat frame. Don't convert anything to JPEG at this stage.

What you're seeing is walking noise. It stems from a slow drift of the telescope over time. Not noticeable on a single light but over time. It highlights the variations of the sensors response (fixed pattern noise).
I always had this with my DSLR until I started dithering. In short, dithering moves the motif randomly between frames (when and how often and by how much can be controlled) by slightly letting the scope point to different directions (we're talking about a few arc seconds for longer FLs). With that the fixed pattern noise will average out through the integration step.

Cheers,
Björn

Helpful Insightful Engaging

Haakon Rasmussen

Aug 30, 2021

I had these artifacts because the darks had higher temperature than the lights. If you shoot in Raw then you can read out the temp to see if darks and light are aproximately the same. If they are not, you must retake the darks and try to get the same temp…

Helpful Concise

Andrew Murrell

Topic starter

Aug 30, 2021

Thanks for the responses.
The "walking noise" sounds like that's it. I haven't employed dithering yet to my images as I am having trouble setting up my autoguider. I have the scope and the camera but I can't get a good connection to the computer, it keeps failing the calibration, typically claiming a drop in the Dec and asking me to check the cable. (that's another issue I need sorting). The Red dots are sensor noise and I know I should start shooting bias. The exposure times were 1 minute due to no autoguiding.
It seems strange to me that you wouldn't process the initial RAW file before stacking but I will give that a try.

Jacob Heppell

Aug 30, 2021

I found the best way to do darks on a dslr was to create a large library at a range of temperatures (about 5 to 35 Celcius) then, for a given nights worth of subs, use the darks from the library that are within the temperature range of the lights.

Calibration and stacking must always be done on the raw, unprocessed images. Converting to a jpeg won't do you any favours as you're significantly reducing the bit depth and performing a lossy compression. Neither are good for enhancing details on DSO's.

And as others have said, dithering is essential. I do it after every sub but every 2nd or 3rd is fine also.

Helpful Concise

Björn Arnold

Aug 31, 2021

Andrew Murrell:
It seems strange to me that you wouldn't process the initial RAW file before stacking but I will give that a try.

Let me ask the question reversed: „why would you process the raw data before integration?“

You see, the information is there. No matter if linear or stretched. What matters is that one wants to improve signal to noise ratio. In order to do that, there’s quite some mathematical theories (namely statistics) around it. They are only applicable if the data is untouched. Otherwise it would become tremendously complicated.

As others have said: you‘re capturing data with probably 12 or 14 bit (didn’t look the bit depth of your camera‘s sensor up) and by converting it to 8 bits and lossy compression (JPEG) you‘re removing 4 to 6 bits of gray scale. Each bit removed halves the gray scale resolution.

Björn

Helpful Insightful Concise