›

OSC Dual Narrow band processing

Mark Mayer
in PixInsight
This might seem an odd question ...

I've processed a lot of LRGB and SHO images and am reasonably proficient, but my experience with OSC is limited. I understand Debayering, and the benefit of keeping RGB channels separate with WBPP, and I understand how to use synthetic luminance for broad band filter.

I'm less clear how to proceed with OSC data acquired with dual narrow band filters, e.g. L-enhance with a broad band pass for Ha and Oiii.

Are their specific AB videos I should watch to get me up to speed for this (new to me) format.

Comments

  • BlockHeadBlockHead
    I am not clear myself on the effectiveness of dual-band filters in front of OSC filters. 
    Sorry, I do not have videos that cover this subject.
    -the Blockhead
  • Mark Mayer
    Adam, I'm surprised by this, and really hope you might address how to process OSC narrow band data in the future. At some point this convolves with traditional SHO processing, but where I'm not sure.

    There is a lot of OSC narrow band data available and your expertise in teaching people how to process such data would be really helpful. This is especially true at the stages of SPCC for stars extracted from OSC dual narrow band data, and also for (L)RGB combination of OSC narrow band images. 

    Not all of us have the luxury of great imaging conditions for SHO narrow band imaging with mono detectors; for those of us who rely on OSC images with dual narrow band, (i.e. Ha and Oiii) filters with different band pass that might include Hb, your insight would be really valuable.
  • David Stearn
    edited August 2023
    I use the 533MC Pro and IDAS NBZ UHS with my Hyperstar.

    I use DBE or ABE, Blur XT, Noise XT then and unlinked stretch using Bill Blanshan's script. Than Bill Blanshan's HOO script. After that I use Curves with Color Masks to to enhance the image. No SPCC should be applied prior to running HOO.
    Here is an example using this method hope it helps.
  • Luke Argent
    Mark, I have OSC and use either a 4nm Ha/Oiii or a 4nm Sii/Oiii
    Depending on that "Triband" filter, it may be much wider on the bands, especially Oiii and you'll be bringing in some LP/moon glow.
    Here is my process that I shared with the PI community over the weekend

    • WBPP - Remember to create a CosmeticCorrection process first, tick CFA and AutoDetect Hot Pixel 3
    • WBPP - Maximum Quality and Local Normalisation default settings
    • WBPP - if OSC you should be ticking Drizzle enabled, Scale 1, Drop Shrink 1 on 'Post-Calibration' tab
    • Dynamic Crop - remove those abberations from the edges, mainly from dithering (I do it myself rather than WBPP autocrop)
    • StarXTerminator generate stars image, do not tick 'Unscreen Stars' as that's for non-linear
    • On the starless, channelextraction R = Ha, G = G, B= B
    • PixelMath, 0.6*G + 0.4*B, new image named Oiii
    • DBE on each of HA and Oiii images whilst starless (great trick I learnt from Adam Block Studios)
    • BlurX on each of HA and Oiii images whilst starless (I used the tick Nonstellar then Stellar) --- I knew my stars were bad due to spacing most nights, I'm replacing them and normally when I have bad stars BlurX makes them worse
    • NoiseX on each of Ha and Oiii
    • Bill Blanshan's (AnotherAstroChannel on YT) HT_Stretch_Unlinked on each of the Ha and Oiii
    • Bill's GH_Stretch on the Stars image
    • Bill's ModifiedSCNR_v5 on the Stars image
    • HT and Curve small tweaks to Ha and Oiii, essentially contrast boosting
    • ChannelCombine R = Ha, G = Oiii, B = Oiii , rename to starless
    • Extract CIE *L from new image. Shawn Nielsen's great 'hack' I've continually used since finding https://www.youtube.com/watch?v=uzwVTZUYNZw&t=659s
    • Apply L created from above to starless
    • Bill's Colour Mask for Red, saturation boost and lightness lift
    • Bill's Colour Mask for Blue (removing other mask first), saturation boost and a stronger blue hue
    • DarkStructureEnhance script, at 0.3 instead of default
    • Bill and Mike Cranfield's script 'Screen Stars' to combine starless and stars
    • Bill and Mike's other script 'Reduce Stars'
  • Woodbine
    I have an OSC and a dual band Ha OIII filter. I am still learning how best to use them, especially post-processing. My thoughts so far; firstly I think that an important limitation is that the Bayer filters are not very selective, so the Ha “bleeds” into the G and B channels which can swamp the OIII if it is relatively weak (which it usually is). For my camera (ASI071) the red rejection is significantly better in the blue channel compared to green. This is visible in the camera specs and on the images. So I used only the blue channel for OIII and added that into both the G and B channels of the HOO image. Although this works pretty well IMO, there is a high price to pay in exposure time as it throws away half the data and two thirds of the weak OIII. This means a lot more exposure time needed to recover the SNR and to have enough frames to effectively drizzle (my system is undersampled). So it is possible to get a decent result, however it is quite tricky and inefficient - I can see why people move towards mono for NB.

    For me, I am considering a compromise by buying an OIII filter. This will allow me to use all the blue and green pixels to make my OIII capture much more efficient. My Ha will still be fairly inefficient, but that is usually a strong signal where efficiency is a bit less of a concern.

    https://www.astrobin.com/a6l6yj/



  • Mark Mayer
    Luke, Many thanks for sharing your work flow- makes sense and easy to follow. 
Sign In or Register to comment.