Author Topic: LRGB Combination...doing something wrong?  (Read 141 times)

Offline dmilner

  • Newcomer
  • Posts: 5
    • View Profile
LRGB Combination...doing something wrong?
« on: 2019 December 08 22:12:08 »
OK, I admit it, I'm probably being a little lazy.  But I'm just trying to get a color pic of the Andromeda Galaxy to show my wife the potential of all the equipment and software (in other words the money I've spent) that I have.  I'm not wanting a perfect pic yet, just something to give her an idea of what can be done.  That being said, I've got a C8 and a ZWO ASI1600mm Pro that I've taken over 500 pics total containing lights (LRGB), Darks, Bias, and Flats.  I've been utilizing Warren Keller's book "Inside PixInsight" to process the pics.  I've gotten through Chapter 6 "Image Integration" following his processes using my pics.  However, I'm really not wanting to take the time to learn how to do all the other post-processing just yet except to add the L and RGB together to make a color image.  So skipping about 100 pages to LRGB combination, I can't get PI to combine the pics due to (1) not being able to process gray scale or something similar and (2) it won't list the luminance pic as one of my choices to choose for the luminance field.  I've uploaded the Master LRGB files that have been fully pre-processed to Dropbox.  You should be able to download them from the following link: https://www.dropbox.com/sh/9v3n0fjv2l1lzea/AAD8eL3eRa0UC57q2IbhKrhGa?dl=0 .  The Master files are identified by the last letter in the name for each LRGB file.  If you have any words of wisdom about this, I'd appreciate it...even if it is "you dummy, you can't skip 100 pages".

Dennis

Online pfile

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 4537
    • View Profile
Re: LRGB Combination...doing something wrong?
« Reply #1 on: 2019 December 08 22:26:30 »
the pixel levels in your L image are exceedingly low... i had to use a 24-bit LUT to see anything. that's not normal... i don't know what went wrong to cause that as i've never seen that problem before.

i didn't look at the RGB masters.

one thing about the LRGB combination tool is that it should be used on stretched images. were you doing that?

the way to do this is to create an RGB image from the RGB masters, do-pre stretch processing (like DBE, linear noise reduction or color calibration), stretch the image (HT, curves) and then do any post-stretch processing you might want to do (non-linear noise reduction, additional saturation boost, etc.) then you process your L image in a similar fashion and stretch it to a similar brightness as the RGB image. finally you specify just the L image in the LRGB process and apply it to the RGB image, thus replacing the L component of the RGB image with your L image.

in short, maybe you need to read those intervening 100 pages :)

rob

Offline WillB

  • Newcomer
  • Posts: 19
    • View Profile
Re: LRGB Combination...doing something wrong?
« Reply #2 on: Today at 03:10 »
Hi Dennis

Besides Robs reply...

The L can't be added because it is already a RGB format image while the R + G + B files are all monochrome as expected.
It is not clear what was the mis-step that output the "L" as a RGB file.

As Rob has identified, the signal level in all the images is really low. The processed images don't really tell us much so we can't say why this is, perhaps something was wrong during the calibration and combination stage or perhaps the wrong exposure time, gain and offset settings were used during capture. You would need to put a single unprocessed frame of each of the LRGB captures as well as any corresponding calibration frames in a dropbox folder for us to take that further.

If you want a 'quick' tutorial of how to assemble a LRGB image with minimal processing then take a look at the two introductory videos on the PixInsight main website, under Resources, Video Tutorials. The source files are also available there to download and follow the demo live if you want. These tutorials were created with an older version of PixInsight so some of the icons look a little different but the processing steps are the same.

http://pixinsight.com/videos/index.html

Also, there is a more up-to-date video tutorial and a different set of practice data available on the "Easy PixInsight" website here (under Tutorials - Monochromatic L-R-G-B):

https://www.easypixinsight.com/

William.

Offline dmilner

  • Newcomer
  • Posts: 5
    • View Profile
Re: LRGB Combination...doing something wrong?
« Reply #3 on: Today at 08:15 »
So what you're saying is that I'm not anywhere near being an expert at astrophotography or PI yet  :) .  Anyway, thanks William and Rob for your info and advice.  I don't really know what is meant by low pixel levels or what caused that.  Here is again a Dropbox link to a sample of my original pics (LRGB) as well as a sample of my Calibration frames (Bias, Darks, and Flats):

https://www.dropbox.com/sh/jaajmvih03y3otw/AADqodmSW9lEK3RomTHHyJj2a?dl=0 

If you have a chance I would appreciate you looking at those and see if there is anything obvious I did wrong in taking the pics.  I'll look at those tutorials and I guess, yes, read those 100 pages.  Thanks, Rob, for the summary of how you add color to the pic.  I'll refer to that while I'm reading through the book again.

Dennis

Offline WillB

  • Newcomer
  • Posts: 19
    • View Profile
Re: LRGB Combination...doing something wrong?
« Reply #4 on: Today at 16:38 »
Hi Dennis.

In the single frames that you put in the Dropbox the individual target light frame subs contain virtually nothing usable, the signal level is much too low. The calibration flats, darks and bias are all fine.

I use a conventional CCD camera and not CMOS so don't claim to be expert but from what I have read in the past and looking at the information contained in the FITS headers for your light frames you need to reduce the offset value from 50 to 25.

The gain setting you used, 139, is not adequate for the length of exposure used for this data set, 30 seconds, and all other things being equal, i.e. good observing conditions, dew free optics, well focussed, an adequate mount etc, then four or five minutes per sub would have been the approximate exposure time at 139 gain setting.

Looking at your luminance sub for example the maximum pixel value given in PixInsight's statistics module is just 0.00302, that's not even 1% of maximum pixel saturation and shows that the image is underexposed. In contrast, your Autoflat is well exposed and has a median pixel value of 0.30665, that's around 30% peak pixel saturation value.
Your flat, bias and dark are all ok, its' just the light frames that are not right.

If the mount won't support a longer exposure time then increase the gain, the read noise of the ASI1600 is very low so the camera can use higher gain settings and still produce low-noise images. Ultimately the gain setting depends on the dynamic range of the target and the Andromeda galaxy dynamic range is not that high so a higher camera gain would have been more appropriate in this case if forced to use just 30 second exposures. 

Tip: When setting the camera gain value try to use even number multiples rather than odd, such as 139, as it leads to quantisation errors (rounding errors) when applied to the camera amplifiers, stick to even number values such as 60, 120, 180 or 80, 160, 240 etc. It's only a small point but it makes it easier later on when you want to accurately determine the characteristics of the camera.

Lastly, from the few stars I can see in the images it looks as though focus is a bit off and possible some tilt in the image train.

From the subs you have posted I would say there is no point you spending any more time with this particular data set, you won't be able to achieve anything to impress the wife just yet  :-\

Concentrate on the data capture side first and get that bit right rather than spending any more time post processing in PI with this data set.

If your capture software has a cursor position pixel value readout then as the image arrives after capture run the cursor over the stars and galaxy core.

For Andromeda the typical pixel value of the brightest part of the core will be something around 50% of maximum pixel saturation value. Much higher and the brighter stars will be over-exposed and blown out, much lower and the darker spiral arms will be down in the background noise. Adjust the gain and/or exposure times until you achieve that ~50% pixel saturation value and then capture a data set. you will need to increase the exposure time for the colour filters but stick to the same gain setting as for the luminance.

The cursor readout will either give a fractional value such as 0.2, 0.5 or a percentage such as 20%, 50% or a pixel count such as 20,259, 50,636 etc, just depends on the capture software you use.

Take a five minute exposure on luminance in the daytime to produce a saturated exposure and then run the cursor over the image in the capture program to determine what the 100% saturated pixel value is and then when under the stars set an exposure/gain that gives you roughly half that value in the core region of Andromeda.

FWIW Andromeda is a difficult first target and I struggled with that one too as a beginner, although it is quite big and very obvious in the sky it does not have much colour or surface brightness variability and post processing the first data set on this target can be quite uninspiring, you will find something like M42 much easier to capture and achieve a satisfactory first image, well that's my opinion anyway...

Sorry that I can't suggest anything else, hopefully the above will be of some use.
Maybe someone more experienced with CMOS cameras can help you out with more expert suggestions for image capture best practice.

William.

Offline dmilner

  • Newcomer
  • Posts: 5
    • View Profile
Re: LRGB Combination...doing something wrong?
« Reply #5 on: Today at 19:37 »
Thank you, William, for your helpful tips, suggestions, and evaluation of my images.  I knew PI was trying to tell me something; just didn't know what.  I have struggled with understanding and applying the concepts of offset, gain, and dynamic range and how they relate to exposure.  While your tips and suggestions will prove invaluable eventually, I think the most important takeaway from your post is that I need to go back and obtain a better understanding of the above concepts before I attempt another deep sky image.  I'm glad to know my calibration frames seem to be adequate and therefore don't have to spend more time trying to get a better understanding those also (at least not like the light frames).

FYI: My mount is a permanently mounted Paramount MX+ (so should allow 4 to 5 minute exposures) while I use TheSkyX to control the mount and Voyager to automate.

Thanks for your time...it is greatly appreciated.

Dennis

Online pfile

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 4537
    • View Profile
Re: LRGB Combination...doing something wrong?
« Reply #6 on: Today at 20:03 »
just make sure that you obtain new darks/bias if you end up changing gain. as long as you already have matching darks or bias for the flats you've taken you'll be OK as long as you still use those darks/bias when calibrating the flats.

rob