0.00
#...
·
2
likes
|
|---|
|
I have been struggling sometimes with Noise reduction. Not getting likes, comments, and such on the socials and wondering if I am doing something to the images that might be affecting this. I have done get a few hours of tutoring in PixInsight and I have learned quite a bit. I am after some constructive Critique on my Pacman nebula shot: https://www.astrobin.com/full/s07cug/0/ Processing entirely in Pixinsight. STF, MLT, Histo, SCNR, Starnet, TGVDN, Histo2, ICC_Trans, LHE_50, MLTSHARP, Starmask, Morph(Star reduction), L-eNhance_Hue, Curves Thanks all, Jeff |
 #...
·
1
like
|
|---|
|
I suggest you to split channels, mix G and B in pixel math with max function max(G,B) where “G” and “B” are the name of your G and B images then use linear fit to equalize Halpha (R) brightness taking as reference the G+B image. now you can combine as a HOO |
|
0.00
#... |
|---|
Alessio Pariani: I tried this but not sure how to combine as HOO, I used LRGB combo but it doesn't look like HOO. Also tried the Light Vortex Method and still didn't look like it. I expect the internal nebula to be blue, right? |
0.00
#... |
|---|
|
Hi Jeff, not sure I can help much but I thought another reply might be in order! I definitely think your narrowband(?) version is much better. I have just done this object myself as my 1st full nb image and think it looks so much better than broadband. Your stars are all blue, mine went green until I got them back to neutral white. Arguably with nb star colour is going to be unnatural but it can be distracting, that's why I turned mine white. As for noise in your image, I am not seeing it unreasonably in a low res laptop display - it depends on what you want out of it. However, like I say I am definitely a rank amateur in these things - I don't even use PI as I resent paying so much for something with such a lousy UI! I use Startools, an outsider but it seems to make pretty pictures, at least in the right hands (not mine). As for likes etc, I think you need a lot of people looking which takes lots of followers which takes time or awards (I look at awards images when considering a new image, gives me an idea of what is considered a good image on Astrobin). However in the end I want an image that I like, and what other people think is irrelevant. Sorry cant be any more helpful than that. 8-/ My version is here: Pacman in SHO ( philhilo ) - AstroBin |
3.71
#... |
|---|
|
Hi Jeff, I know noise is the devil that haunts all of us. I saw a review in the British magazine, Astronomy Now, of a product from Topaz Labs called Denoise AI (artificial intelligence). I downloaded the free 30-day trial, and it works wonderfully. I shoot Ha only, LRGB, and OSC, and Denoise helps immeasurably on all of them. The price for the app is $79 US, and, for me, is absolutely worth it. I've also been scanning some old family photos and cleaning them up in PhotoShop, and Denoise is the perfect finish to reduce graininess. Nothing to lose with a free trial. It reduces or even eliminates the background sky noise, and can sharpen also to a selectable degree, if you wish. |
9.48
#... |
|---|
|
Topaz Denoise AI is absolutely amazing when it works well. However, it has known shortcomings for astrophotography. 1) It can erode the smaller stars such that they are not round, or even filter them out completely. 2) It provides such a strong effect that it can easily be abused, turning your images into "plastic." 3) Because it uses artificial intelligence to "guess" what the image would look like without noise, and it has a sharpening feature that cannot really be turned off, there is the possibility of introducing details or features that are not really there. There is potential for abuse, especially in planetary imaging. I purchased Topaz Denoise AI, and I found it to be incredibly useful for reducing ugly noise in the dimmest parts of an image (sky background, dark nebulae), where there is little useful information in the data anyway. I typically apply standard noise reduction on the highlights & midtones with Topaz AI on the darker, most noisy background areas, with careful masking of smaller stars. I have found it to be a wonderful product for narrowband images in particular, when used judiciously. |
9.53
#... |
|---|
| Hi Jeff, I looked at your image. Looking great to me. Keep doing what you're doing! |
0.00
#...
·
1
like
|
|---|
| If you post your calibrated un-stretched files out onto a share drive I will have a go at them. |
6.80
#... |
|---|
|
Hi Jeff, Ok, echoing (agreeing) what others have said about not judging yourself by likes, I'll give a go at providing some feedback. I think it's really helpful to search for the object (using quick search) to get a feel for what others have done. I personally also think it's nice if you're doing narrowband (but not doing a color paradigm shift like SHO) to be aware of what the object looks like in broadband. I'll offer my take on the object (broadband) for comparison. Note that I acknowledge up-front I have too few subs (only 25 minutes @ f/2.2). My background is too black and is also too noisy. I'm still working on my version. However, I use PI's photometric color calibration, so my colors aren't too far from where they should be. Maybe we could debate whether color saturation is too much or not. In comparison, I think your take has some things to work on. These include too much blue in the background and stars. The reds don't quite "pop" the way it seems they should. You've lost some orange/red color in the field stars, so the object looks a bit flat. Hopefully these things will help you. Best wishes as you push to your own "great take" on this object... cheers, Doug https://www.astrobin.com/full/kj8n4x/0/ |
0.00
#... |
|---|
| Both versions look very nice to me. Their technical qualities are done well. Aesthetic qualities generally vary with the preferences of the imager, but I think these look good and not overdone. |
