›

AmpGlow does not get eliminated using Darks

Nihal Mirza
in PixInsight
Hi Adam, 

Hope you are well, 

I've recently imaged M101 LRGB and decided to add Ha to the mix. Using WBPP (2.4.1) I get to the image registration stage after which I Use NSG (non subscribed version) for final Master Light. I've always let WBPP do the final image integration just to have something to compare against NSG output.

However, I noticed something that is amp-glow (ZWO ASI294MM Pro) wasn't calibrated out completely using darks in my Ha master subframe (LRGB was OK, no trace of amp-glow).

So I took the following steps to debug:
  1. The first thing I did was calibrated and stacked Ha subframes in isolation via WBPP, as still get the amp glow.
  2. Remove flats and flatdarks and calibrated and stacked using Ha Light and Dark frames, still get amp glow
  3. Redo my Darks, ensure Temp, Exposure and Gain are all what they should be, then WBPP'd darks and Ha light subframe, still get amp-glow
  4. Notice that Local Normalisation is checked by default, uncheck it and calibrated and stacked darks and Ha light subframe .....Voila...no Amp-glow in Master light Image
  5. Repeat with Flats and Flat darks (Local Normalisation unchecked), no amp-glow
  6. Repeat with old Darks (Local Normalisation unchecked), no ampglow

Therefore my conclusion is Local Normalisation is the cause, maybe the initial values are not adequate.

Given that I use NSG, I then proceeded to process the files from Step 5 to obtain a final image.

Here I fail again, I cannot reproduce the result of WBPP i.e. Step 5, Unless I change sigma low to ~ 1.20 and the final image seems rather washed out.

Any advice?

FYI I did try image integration using the ESD settings that WBPP uses, however couldn't figure out how to set esd_low =1, in anycase wasn't able to get same results as WBPP

Btw shouldn't the Darks completely remove amp glow, this is the first time I've some across this issue with any NB images +Amp glow + WBPP or NSG.

Image 1
Below on left is WBPP output with Local Normalisation unchecked (reject low is quite different to the one on the right),
On right is WBPP output with Local Normalisation checked by default.

Image 2
Last screenshot is for the Image integration settings for NSG

Image 3
Comparison of WBPP output (Local Normalisation unchecked) v NSG (NSG settings in Image 2)

image


image



image

Comments

  • BlockHeadBlockHead
    Hi Nihal,

    Would you mind reviewing this video?
    It would seem you need a pedestal. You can check for zeros...and confirm.
    -the Blockhead
  • Nihal Mirza
    Hi Adam,

    I've seen the video but it is a good suggestion to revisit and reassess.

    I am using Pedestal value (after I saw your video few months ago) for the Narrow band filter, may be this is a new filter 3nm (older was 7nm) therefore I need a different Pedestal value.

    See screenshot for WBPP set up, one thing that I have done differently is ask WBPP to select Registration reference frame wrt to Filter.



    image
    WBPP.JPG
    1750 x 828 - 215K
  • BlockHeadBlockHead
    I think my question is...in your calibrated data (with your new filter)... are there zeros? (especially in the location of the amp glow). This indicates either an oversubtraction (which means something is up with the dark frame) or the pedestal issue.

    -the Blockhead
  • Nihal Mirza
    edited May 2022
    So I had a look and most of the Image is mostly Red after I used the pixmath expression.

    Should I simply experiment by increasing the pedestal value?

    The video does show how to workout negative values (per pixel) but given that the frame has so many zero pixels, I think I'm better off just experimenting by starting with a high value.
  • BlockHeadBlockHead
    You need to make certain this isn't an oversubtraction of a bad dark. Verify the numbers make sense.
    If you are convinced the dark frame is good... you can use the auto pedestal of the WBPP or you can look at the value(s) of a pixel(s) in your raw light frame and you dark. You can easily see a place where the values become zero and see what value the pedestal should be. 

    I am still thinking this is a pedestal issue...but the larger the value needs to be... the more likely it is a bad dark issue. 

    -the Blockhead
  • Nihal Mirza
    I did use the auto pedestal in WBPP but that too didn't help. I'll do the following:

    1) Redo Darks
    2) Work out what the pedestal should be..

    Thanks Adam, will get back soon.

    Nihal.
  • Nihal Mirza
    To avoid any negative values I need a Pedestal of approx 900, 
    • Is this value a bit too high?
    • If I NSG these files with the settings show in my first message I still get amp glow, but then again it could be down to my settings in Image integration
    I'll redo the darks tomorrow.

    Btw my exposure was very low for M101 (as it wasn't planned at the time), exposure was 120s with Gain 130 (3nm Ha filter), I wonder if this might be to low for NB filter (with 7nm Ha filter I would use at least 300s exposure).


  • Nihal Mirza
    Hi Adam,

    I haven't been able to get to the bottom of this issue even after I redid the Darks. Any feedback on the Pedestal value from my previous comment.
  • BlockHeadBlockHead
    Hi Nihal,

    You will need to make the images available.
    I suggest 5 images of each type. If you do this...remember to make the folder have permissions that "anyone with link can access/view."

    -the Blockhead
  • Nihal Mirza
    Hi Adam,

    Hope this link works, let me know.

    Provided 20 images of each Flats, darks, lights.

    Also included are my results from WBPP integration with and without local normalisation, NSG integration.

Sign In or Register to comment.