Astro Pixel Processor

Calibration frames with different ISO settings, how to process?

Viewing 15 posts - 1 through 15 (of 44 total)
  • Author
    Posts
  • #13190

    MauriceToet
    Participant
    posts: 51

    Usually I expose my lights (and darks) at a different ISO setting than my flats (ISO 1600 versus ISO 200). Therefore I have seperate (master) bias frames.

    What is the best practice / workflow to process the data? How do I make sure the correct master bias frame is applied to (calibration) frames that were exposed at the same ISO setting?

    #13195

    Haverkamp
    Participant
    posts: 641

    Hi Maurice,

    APP has the following rules for applying master calibration frames to frames that need calibration:

    First, for BIAS frames there are no rules, since they can’t be calbrated

    Darks: can only by calibrated by a master bias of the same ISO or gain + offset (but usually gain + offset are not even in the fits header of the software that does the capture)

    Flats:

    1) can be calibrated by a master bias of the same iso (gain+offset)

    2) can be calibrated by a master dark of the same iso (gain+offset) and the same exposure time

    3) can be calibrated by a Bad Pixel Map

    Lights:

    1) can be calibrated by a master bias of the same iso (gain+offset)

    2) can be calibrated by a master dark of the same iso (gain+offset) and the same exposure time

    3) can be calibrated by a master flat ( no restriction on ISO or gain/offset )

    4) can be calibrated by a Bad Pixel Map of the same instrument name (instument is fits convention for the used camera)

    So your masterbias of iso200 for your flats and your masterbias of iso1600 for your lights will automatically be applied to the correct frames (that have the same ISO), since APP has the strict rule that master bias frames must match the iso of the frame that needs calibration.

    Even better, you can visually check if calibration works like it should.

    When you have created your master calibration files, load your lights and your master calibration files.

    For each light there will be certain marks in the frame column in the bottom file list window, indicating this frame can be calibrated now with a certain master calibration frame.

    B for Bias

    D for Dark

    F for Flat

    BPM for BPM..

    Now above the image viewer window, there is a dropdownbox, which is set at linear by default.

    Click on one of your images in the filelist and it will be shown linearly.

    And you can also display your frame, fully calibrated with the masters. Set the dropdownbox to

    l-calibrated

    and the image will be shown linearly and calibrated with the master frames that are supplied. So you can visually check if the master flat is of good quality and also if the BPM for instance actually removes bad pixels.

    Let me know if this is clear 😉

    • 1 person likes this.
    #13212

    MauriceToet
    Participant
    posts: 51

    Excellent, Mabula. Very clever! I’ll have to make new (master) bias frames; which is not a problem. It didn’t work with converted PixInsight xisf to fits master bias files (probably due to missing gain info in the header).

    #13213

    Haverkamp
    Participant
    posts: 641

    Hi Maurice,

    Thanks, you can convert your old master calibration files from PI to APP, but it is not so easy. The main reason being:

    APP uses the whole sensor of your camera for the creation of the calibration frames. And a program like PI, that has a dependency on DCRAW doesn’t do this. So the calibration frames that PI makes have smaller image dimensions than those of APP.

    To use your old master calibration frames you need to figure out what offsets to add on all 4 borders of your PI calibration masters to get to the correct image dimensions for your whole sensor. I can help with this if you want to. I have done this conversion for instance on old calibration files of Scott  from Images Plus, which also has the dcraw dependency.

    You can use the batch modify tool, it can even uncrop and undebayer, something I needed to do on Scott’s data.

    But for now, it’s probably best to use new bias frames. That will be much easier for now.

    Let me know if you succeed in calibrating your lights properly.

    #13214

    MauriceToet
    Participant
    posts: 51

    Thanks for the further clarification and your proposal to convert the PI calibration files to APP files. I’ve just shot new sets op bias frames and APP is “chewing” on it as we speak 😉

    #13259

    Haverkamp
    Participant
    posts: 641

    Did everything work out fine @Mauricetoet?

    #13279

    MauriceToet
    Participant
    posts: 51

    Yes, @bula, everything worked out fine. Next question is… how to create a seamless mosaic? ? That’s something to discuss in a new topic.

    #13512

    Rosen
    Participant
    posts: 16

    Maurice, Mabula – When I was using a DSLR, I also liked to capture my lights at ISO 1600 and Flats at ISO 200 (longer exposure means that each flat has better SNR).  What I would do is to use Bias frames at ISO 1600, and flat darks at ISO 200 (and, of course, same exposure time as my flats). I had used Images Plus for calibration, and it would default to applying the flat darks to the flats.  I presume that IP only used the bias frames for dark scaling.

    Mabula – Would APP by default use such flat darks for calibrating the flats?    Would this be a good method for Maurice to use?

    #13604

    Haverkamp
    Participant
    posts: 641

    @srosenfraz, personally I don’t have experience with IP, but I would think bias frames can’t be used for dark scaling since a bias frame is not a representation of any significant dark current.

    Dark scaling is on my todo list, since APP can’t do this yet. But having said that, dark scaling should always be considered as a last resort. It’s always better to make good darks on the right temperature. But I know, especially for uncooled camera users, this can be a tricky business. Then dark scaling could help.

    (If you use dark scaling, don’t use darks with exposures of a few seconds. Always use a master dark with a real and significant amount of dark current. )

    Returning to your question ;-)…

    Yes, APP will use the flat darks by default in this situation.

    APP has strict rules on the master calibration files that can be applied to a frame that needs calibration:

    a dark can only be subtracted by a bias of the same iso / gain

    a flat can only be subtracted by

    – a masterbias of same iso /gain

    – a masterdark of same iso /gain and same exposure within a certain tolerance %

    or BPM corrected

    – by a BPM of the same image dimensions and instrument name.

    The same rules apply to the calibration of a light frame

    So if you have calibration frames of different iso values for the lights and the flats then APP will automatically apply the correct ones.

    And it is always advisable to visually check the calibration, before proceeding to star analysis, with the image viewer option l-calibrated (above the image viewer) instead of showing the image only linear.

    #13605

    Haverkamp
    Participant
    posts: 641

    I think @mauricetoet can use dark flats instead of bias, but I would only advise dark flats if the sensor requires it, that means: some sensors have a Fixed Pattern Noise (FPN) dark current which is already significant with short exposures. Then flat darks are the better option. If there is no FPN in the flat dark, bias are recommended since it should contribute less noise injection from the calibration.

    And If you have made a BPM, I advise to always apply the BPM in flat calibration as well. It will reduce noise from hot pixels in your master flat.

    #13622

    Rosen
    Participant
    posts: 16

    As far as IP using the bias for dark scaling – I believe the methodology is to subtract the bias from the dark so that only the dark signal is left.  Then the dark can be scaled.  So, let’s say I have a 5 minute dark and want to use it for a 10 minute light.  If I subtract the bias from the dark, I can scale the dark x 2 to match the (presumably linear) dark signal in the 10 minute light.  Then I can separately subtract the bias from the light.

    Personally, I’m wish you – I avoid these “short cuts” and much prefer to use properly matched calibration frames.

    #13658

    Haverkamp
    Participant
    posts: 641

    @srosenfraz, okay, that makes perfectly sense then ?

    Providing dark scaling is on my todo list, but indeed, I would advise everyone to only take “short cuts” for previews or if no other option is available…

    #14652

    KeesScherer
    Participant
    posts: 1236

    Now back to the practical side, i have many objects imaged with the Canon 6Da in my archive and want to combine those RGB’s with new Luminance data. So i have years of iso1600 RGB data with corresponding iso200 flat frames. But there is no way that i can (re)stack this data with APP because the flat calibration is not done for these datasets. So i have to do the (re)stacking in Pixinsight if there is no “override” option for this.

    #14654

    Haverkamp
    Participant
    posts: 641

    Hi @keesscherer,

    This is perfectly possible in APP with your data.

    (I calibrate, all my own data with flats of different iso or gain compared to the iso/gain of the light frames )

    Since you indicate that you still have the light frames and flat frames, you only need to make sure that the bias pedestal is removed from both light and flat frames using the correct bias/gain for those frames in the calibration proces.

    So using an iso 1600 masterbias/dark to remove the bias pedestal in iso200 flats, is fundamentally wrong. You need to apply an iso200 masterbias/dark to remove the bias pedestal in iso200 flats.

    This rule in the APP calibration engine actually prevents user errors and doesn’t pose limitations, since bias frames are easily made. Even years after you took the original data for most camera’s.

    What would need  to be “overridden” in your opinion?

    #14655

    Haverkamp
    Participant
    posts: 641

    To add:

    You can then save your calibrated rgb frames and split them directly in the r,g,b channels as an option in “save calibrated files” for further combination with your monochrome data of your new camera.

    • 1 person likes this.
Viewing 15 posts - 1 through 15 (of 44 total)

You need to log in or to reply to this topic.