Strange grainy artifacts in background when applying NXT - normal or error?

5 replies118 views
Derek Vasselin avatar

Is this normal or am I doing something wrong? When I apply NXT, I get really strange grainy artifacts in the background. It looks like colored specs all over the image (see below).

Doesn’t appear to matter whether I do NXT early in processing on linear data, or near the end on stretched data. And doesn’t seem to matter if I do more or less NXT (I’ve tried 0.35 up to 0.9 denoising and still get these artifacts). I’ve tried messing around with the Version 3 AI settings as well (HF, LF, intensity, color) and haven’t figured out a solution.

Details (let me know if you need more):

  • Celestron EdgeHD 8

  • ZWO ASI2600MC

  • Optolong L-Enhance HaOiii filter (~18 hours)

  • Altair SiiOiii filter (~28 hrs)

  • Used DBXtract to create Ha, Sii, Oiii and create an SHO image

Is this just an issue with the data or limitation of dual-band imaging? Or is there a specific method to avoid?

📷 image.pngimage.png*Edit: I don’t apply any BXT or NXT until after I combine the 3 channels. Could this also be an issue or should that not matter?

Well Written Engaging
andrea tasselli avatar
That's one of the defects of NXT (V.3). Use sparingly and at maximum iteration (5). Best is to use V.2 to avoid those artifacts.
Derek Vasselin avatar

Ah ok. That explains why I never remember this with V2. Thank you

Quinn Groessl avatar

Yeah I don’t get why RC hasn’t released an update either fixing it or getting rid of AI 3 because there are a ton of people who use it and don’t notice/don’t care/don’t know that it’s leaving a bunch of artifact.

Derek Vasselin avatar

Yeah it’s been driving me crazy for weeks and couldn’t figure out if it was me, a software issue, or a data issue.

Tim Ray avatar

I usually make a smaller preview and run NXT in preview mode while I change HF/LF Pixel Scale slider. Most times this “for me” is 5.5 to 6.0. But keep moving the slider until you get satisfactory results. I have had 4.0 to as high as 8.0. It really depends on the image.

The Denoise sliders for HF and LF color I set to 1.0

The HF Intensity I leave at .9

The LF Intensity I set somewhere between .62 and .82

Russell has a nice description on how to use NXT on his site.

Good Luck and CS

Tim

Helpful
Related discussions
Artefacts near coma-corrected stars after applying "Correct only" in BlurXterminator
Hello dear AstroBin community, I'm posting this to see whether anyone else encountered this issue after applying the BlurXterminator and whether there is some way to avoid it. Below this post, I've attached four images showing the lower-left ...
Directly addresses similar artifacts from NXT processing, likely same root cause.
Oct 24, 2024
Struggling with gradient artifacts around stars in mosaic merge
Hello, Im learning how to do mosaic but Ive been struggling quite a bit with the merge. I’m trying to follow the usual steps “ MosaicByCoordinates → dnalinearfit → GradientMergeMosaic”. I’ve tried to do merge with and without BE/Denoise and play arou...
Grainy artifacts in background may relate to merge gradient issues.
2 days ago
Multiscale Adaptive Stretch MAS in PixInsight
Hi, Two days ago I heard about MultiscaleAdaptiveStretch I have been playing with it and as an advanced beginner I like it. It saves me a lot of pre-stretching time with HistogramTransformation (Sorry, Never understood GHS 😵‍💫? I also tested VeraLux…...
NXT artifacts may relate to post-processing stretching issues similar to MAS workflow.
Jan 1, 2026
Blotchy images after noise reduction
I've noticed lately I struggle with blotchy images during processing. Specifically after using NoiseXTerminator. You can see what I'm referring to in the top left corner of the image below. I am using a ZWO ASI2600MC, with an HaOiii dual-band...
NoiseXTerminator artifacts discussed; similar blotchy/grainy background issues reported.
Oct 16, 2024
Dual band processing as OSC vs HOO
I'm curious as to how other people like to process their dual-band data on OSC cameras. I have about 20h of HaOiii on M101, my main goal was to do this to boost the Ha to add to upcoming broadband data. In my OSC image (Stretching + green noise r...
Discusses processing artifacts; relevant to author's stretching and noise removal challenges.
Mar 6, 2024