When to run BlurX?

Adam - I was just watching your recent video with Russel Croman concerning the new NoiseX Process. Thank you - it is excellent.

I'm not sure if it was just my copy of the video, but there were breaks in the voice feed from Russ's end. I would not mention it except these breaks were at their worst when Russ was, I think, saying that you should always, in NB, run BlurX on the individual SHO masters before you combine them in your color palette of choice.

Is that true?

bob

Comments

  • edited February 11
    Yes. BXT takes into account the relationship between color channels when applying to a single tri-colored image. It is best to operate on the three NB images independently and optimize the results. However, I do not believe *always* was said with such finality. (For broadband...just operate on the RGB ...because it was training to understand this.)

    I assume you are talking about this point in time:
    (right click on show this in another tab or just go to minute 23:44 .)

    There are no audio issues as far as I can tell.

    -the Blockhead
  • OK - still a bit "blurry" on this one.

    Your posted Narrowband Routine progresses thusly:
    1. Create Masters (Ha, S2, O3)
    2. Remove Stars from each individual master
    3. Linear Fit each individual master
    4. Channel Combination
    5. Remove Gradient
    6. NarrowBandNormalization
    7. THEN BlurX
    8. NoiseX

    Is Russ suggesting that BlurX should be inserted between step 1 and step 2 - and performed on each individual master - before channel combination?

    Also, if BluxX should remain where it is, how does it generate a PSF if the stars have been removed? Should I determine a PSF myself before removing the stars?

    And, it gets a bit blurrier still after reading the BlurX documentation in PI (below). When he mentions "narrowband color balancing" is he referring to Linear Fit? Which could also determine where best to use BlurX.

    Sorry if I am getting words and phrases and other definitions/concepts all convoluted - and missing the point. Maybe I need WordX. 

    cheers - bob
    2.2.6 Narrowband images

    When processing narrowband images, if any channel mixing is planned (e.g., mixing Hα with [S II] for the red channel), perform this operation after deconvolving the image in distinct-channel form: one filter per color channel. For example, perform a straightforward SHO color combination initially, keeping the channel weights relatively equal for this step. Do any strong narrowband color balancing and/or channel mixing after applying BlurXTerminator. Mixing the channel data prior to deconvolution, or applying a very strong boost to one or more channels (e.g., [S II]) may alter the PSF enough to cause inconsistent results, particularly on bright stars.

  • Bob... on #7... WHAT did I BXT?
    That matters...

    "Is Russ suggesting.... "
    Yes, he is..and so am I. It is usually best to operate on the channels individually instead of a tri-color image. 
    When I made the Wizard Neb example- I do not think I was thinking carefully about it. However... if you know you are going to use one of the channels for the lightness... then you really only need to BXT *that one*.
    Otherwise, you could BXT the individual channels before hand (and then either use one of the channels with lightness or not). There is no single way to do this stuff.

    "How to determine PSF... " 
     You just find a reasonable PSF in your image. It does not need to be perfect. You already know the answer I suspect. When you focus... how many pixels in FHWM is your focused star on good nights? This is the value you use for manual if you want to have an ordering of BXT where you use it after star removal. There is no "step" it falls between. There are just some "rules" that give better results. 

    Yes, narrowband color balancing is effectively linear fit or normalization. However, the suggestion that Russ mentions here is not a strong rule... .it is a weaker one. Performing BXT on channels that have been "boosted" may actually give better results because you are also boosting the significance of the noise and this may (or may not) be something you need to see with BXT at this point. 

    Remember, when Russ wrote that...he had just released BXT. Keep in mind that developers and content creators cannot go through every possible variation for forsee every eventuality... especially early on. This is one reason what I often wait a bit to make a video. I need to literally run a process or technique many times to fully understand it. No short cuts.

    You are looking for a hard and fast rule. If you want that- do not follow me exactly. Just use BXT on the individual channels while they have stars. If you are willing to be a bit flexible. You can apply BXT to starless images and change the ordering of things as necessary.

    -the BLockhead
  • Thanks Adam!!

    Starting with your closing comment - basically, be flexible, no hard and fast rules. I get it. For 40 years I had an outdoor adventure company with over 100 employees - talk about moving targets. 

    1. "What did I BXT?"

    See attached screenshot.

    2. PSF - got it - had it.

    3. Like I said - WordX.

    Cheers - bob

    Screenshot NB 2.jpg
    1420 x 1894 - 384K
  • Regarding the picture...
    Hey... look at that! At least I am consistent in my teaching!
    great.... 
    -the Blockhead
Sign In or Register to comment.