›

Color "gradient" in OSC images shot with narrowband filters

Wesley Mitchell
in PixInsight
Hi Adam,

I'm getting the following when trying to use one of the duo-band or tri-band filters with an OSC.
The image I'm enclosing is stacked 30s lights of M42, integrated in WBPP, and autostretched in PI.
Notice the left side has a magenta cast, while the right side has a green cast. I can (more or less) calibrate it out with DBE, but I was surprised that having flats that I thought were good didn't remove this "gradient". 

I shot the same target with the same settings, same equipment, but no filter at all, and got perfectly good integrations with no appreciable cast.
I shot a separate set of flats for both the filter (IDAS NB1), and the no filter setup, and used those for the calibrations.

First question: do you know what this ugliness is?
Second: is there a way to avoid it?
Third: what is the best way to process to get rid of it?

I actually got much better results (detail, range) with no filter. I live on the coast in San Diego, and at least the sky pointing west is about Bortle 6.

Equipment used:
Explore Scientific 102mm f/7 scope
ASI 2600MC camera
ZWO AM5 mount
ASI078MM guide camera
WilliamsOptics 30mm f/4 guide scope
ASIAIR+

Thanks for your help
Wes

Comments

  • Wesley Mitchell
    PS I get similar issues with Antlia tri-band filter, and Optolong L-Extreme
  • BlockHeadBlockHead
    I think the red/green gradient isn't the important point. It is the fact there is an overcorrection in general. This means that either the flats you are creating do not match the lights in some way- or there is a calibration error with regards to the flat. Since narrowband images do not have an appreciable sky it might be the error appears larger here.

    1. Are you certain your unfiltered has a UNIFORM result after flat field correction.. no dark center correct?
    2. If so... you need to show a screenshot of (I assume you used) your WBPP setup with raw images (not the masters).

    -the Blockhead

    P.S. I find the double filtering to be... I am not a fan of this I guess. 
  • Wesley Mitchell
    No, not double filtering. It means I tried three different filters, singly, and got similar results.

    I'm enclosing screen shots of the wbpp setup, the autostretched integration image, and the (unstretched) master flat. I don't think there is a dark center to the light image, but it's kind of hard to tell because of the bright center of M42 and the dark cloud halfway around it.

    There are two sets of darks, the 1s exposures I am using as Dark Flats. There were no bias frames.

    Would it be helpful to set up a folder on DropBox, with files from both the filter and non-filtered sets?

    Thanks for your help.
    master-light-stretched and master-flat.jpg
    1727 x 617 - 433K
    WBPP-1.jpg
    1412 x 947 - 309K
    WBPP-2.jpg
    1412 x 947 - 289K
  • BlockHeadBlockHead
    Your NB filters are in front of your OSC filters... this is double filtering (in my way of thinking). 

    Regarding the calibration, you are calibrating your flat with a dark that is a different exposure time. Short as they may be... some sensors are weird if you do not match times. It shouldn't matter likely for your sensor... but you are already tweaking convention.

    In addition, you are using a EXPTIME keyword? WBPP already takes care of matching for calibration data...or you can force it manually. However, I do not know why you are using this keyword. 

    But the bottom line is that using a dark that is LONGER than the flat to calibrate it is very unconventional..I do not like the logic. :)

    If you want to start to work on this, perhaps we can get on the same page with regards to this?

    -the Blockhead
     

  • Wesley Mitchell
    Hi,

    I was using EXPTIME because I was processing 30, 60, and 120 second exposures, all in one run. I had darks of the same exposure times for the lights.

    Are you saying to not use any filters at all with an OSC? If so, is there any use for these so-called nebulosity boosters? I have read comments to the effect that this happens with other ZWO OSCs. I can confirm it also happens with a Player One Poseidon-C.

    I can certainly take some dark frames of the same exposure times as the flats, and try calibrating the flats against those. However, I have some data which already has dark flats which match the exposure time of the flats, of a different target, but still exhibits the same behavior.

    I set up a separate directory in dropbox with dark exposures of 1.6 and 600 seconds, flats of 1.6 seconds, and a light frame of 600 seconds, so the times on the darks match the appropriate calibration target. I also included the master flat, master darks, and master light, and the calibrated and debayered outputs for one of the light frames. It's interesting that the single debayered light frame already has a strong side-to-side cast.

    This is a different target, M81, but all other factors that I control held constant. The filter used is an IDAS NB1.


    I was using a filter drawer, simply because it was handy. I experimented with covering the junctions in the imaging train with black electrical tape. This had no effect that I could see. I just removed the filter from the filter drawer, and inserted the empty drawer, to take the no filter images, which don't have any of this cast.




    I included the log file from the run. Is there anything else I missed here?

    Thanks
    Wes

  • Wesley Mitchell
    Per our discussion, I put a folder with the results of shooting without a filter here. There is a very slight gradient in the light frames, I think it's in the opposite direction of the gradient with the filter on.

    Thanks
  • BlockHeadBlockHead
    I believe you have precisely the same issue in your unfiltered data that you do in your narrowband (filtered) imagery. This is what I suspected when we talked and I think you should proceed with the suggestions I made.  (See the attached image)



    -the Blockhead
    Capture.JPG
    1030 x 712 - 338K
  • Wesley Mitchell
    edited March 2023
    I now believe that the problem is with the camera itself. I took full sets of photos using two different cameras, the ASI 2600 MC and the ASI 533 MC. The master light for the 2600 shows a strong gradient from green to magenta, and also the dust donuts were not calibrated out. The 533 set shows no color gradient, and all the dust donuts calibrated out.

    I took complete sets of bias, darks, flats, and lights for each camera. Same temperature, exposure time. Gain was a bit different.

    The imaging train was constant, except for the camera connector. No filters were used, just spacers for backfocus.

    If there is a light leak in the 2600, how do I find it? It does have a tilt plate, could light leak through there? How can I test?

    Thanks
  • BlockHeadBlockHead
    Test for light leak by taking calibration data in total darkness. 
    Take light frame data away from lights without moon in the sky.

    -the Blockhead
Sign In or Register to comment.