Issue with Image After Adding Stars Back into Image

edited August 2024 in PixInsight
Good evening all!

I am trying to follow the steps in Adam's video on NGC1491 with an image I have of M16. The images were processed with WBPP, no diagnostic errors when run. 

When I open the Master light that was created, use STF, I get an image similar to Adam's of NGC1491. The green bas is now under the file name. Then I  use StarXTerminator, and all seems ok. I get the stars image, with the green bar under the file name on the stars image, and the nebulae has no stars, looks good, and still has the green bar. After using DBE on the nebula image, I then tried to merge them using PixelMath. Both images still have the green bar under the image. But the combined image after running PixelMath, still has the green bar under the name,  but its clearly not right - the nebula isn't visible, and if I hit the STF button the result is that collection of over bright "stuff". 

Trying to figure out why, I used StarsXterminator, abd went straight to using PixelMath. Same result.

I'm clearly doing something wrong, but I have repeated the process several times, following video 4 after the color combination sine I have an RGB image, each time with the same result.

Any ideas or suggestions what I might be doing wrong, or setting up incorrectly.

Thanks,

Rich

  

Comments

  • Having a "green bar" doesn't meaning anything other than a screen stretch is in place.

    If you RESET the STF process (with the reset button at the bottom right) and the image goes black- this means your values are small and you are likely looking at a linear image.

    If you RESET the STF process and the image is visible (likely how you wanted it when you permanently stretched it with HistrogramTransformation) then the image is non-linear.

    The initial stars image (extracted stars by using SXT) are LINEAR. .
    The starless image is also LINEAR. 

    Yes you would use DBE on the Linear starless image if necessary.

    The part that appears to be missing before you combine any images is doing a permanent stretch with HT (or GHS).  You attempted to combine your linear stars and nebula images with Screen blending.. this isn't what you want to do I think.

    Can you please indicate the link to "video 4" ?
    (Are you watching the Workflow videos??)

    I will eventually be generating a PDF with a step-by-step workflow which follows the video. However, I just re-recorded these videos on the 12th because of a mistake I made.

    -the Blockhead


  • Adam,

    First off - I apologize for the confusing short-hand referral to the video. It was the fourth in the series working with the images of NC1491. The title was BXT(Correct Only), DBE and SPCC.

    The M16 image was Linear - I had already checked that by reset SFT and the image would go black.

    This morning I used the images that you provided of NGC1491. I got the same funky result when I added the stars back in with PixelMath. So I knew it was not something I had messed up with WBPP. 

    But I did FINALLY realize what I was doing. I was using PixelMath incorrectly. A very stupid type error on my part. It was so stupid, I have been laughing at myself all afternoon at the time I've wasted fumbling around. But if repetition helps cement learning - I'm in concrete not at least on these steps.

    The PDF would be a great addition. I know how much time and effort that takes. Both before I retired as Process Engineering manager, and since, I work on training the young engineers in both "real world, practical engineering" and the simulation software tools we have to model refinery and chemical process units. I do alot of powerpoint slides - and supporting notes and I know very well it's no small amount of effort to prepare and do these programs.

    Question - I built a new machine last summer that is pretty powerful. The graphics card I used is a Gigabyte Nvida 4070 Ti. Not a low end card. But when I run SXT, it crawled compared to yours in the video. If you don't mind my asking - what video card do you have ?

    Also - I upgraded the camera to a ASI2600 MCC. Could not go the full frame route - but I have other options to handle larger images - like Andromeda for example. Found some good videos on proper settings of gain/offset and the images I'm working on for M16 were my first shots.

    Thanks for the help - and please let me know what card you have if you don't mind.

    Rich
  • Remember you need enable CUDA acceleration to get your GPU to do this trick.

    Russ Croman has made this multi-step process a single operation with a repository he created.
    This is the repository to add:

    Read this post about it:

    I believe people have been having good luck with it. Once you get it working... you will really enjoy your GPU.

    I was using the same GPU on my other machine. My new machine has the NVIDIA GeForce RTX 4090 :)

    Your card will do just fine when you get the CUDA bit going. 

    -the Blockhead

  • Adam,

    Thanks for the followup - I got the information from Russ' site. Need to do a little more digging on it. If I understand getting it setup also means loading the CUDA toolkit from NVIDA's site - correct ?

    Something else to learn - but I'm really loving it.

    Made more progress this evening on my M16 images this evening - working thru the BXT Small Star De-Emphasis now.

    Thanks !

    Rich
  • Yes.. you are correct. But loading the CUDA stuff is automated by Russ' repository link. It automates everything I believe. 

    -the Blockhead
Sign In or Register to comment.