Deep Sky Stacker doesn't get the colors right...WHY!!

[wols...@gmail.com]

Earlier in 2020 the folks at Deep Sky Stacker changed DSS so that it looks for the BAYERPAT keyword in the FITs header and will use it to decide whether a FITs file is a RAW color or greyscale image.  If DSS finds a BAYERPAT keyword it will intepret the pixel values as RAW color values with the colors assigned as declared in text string associated with the BAYERPAT keyword.  Some examples of the text string are RGGB, BGGR, GRBG etc. CaLIGHTs has been programmed to support the BAYERPAT keyword and will include this keyword in calibrated color LIGHT frames.

There are many astrocams programs/drivers etc. that do not support the BAYERPAT keyword.  For those astrocams, DSS has included a checkbox called "Monochrome FITS Files are RAW files created by a DSLR or a color CCD camera".  This is an unfortunate choice of wording because while your camera may be a DSLR or a color CCD camera you should NOT have this box checked if you are stacking FITS files created by CaLIGHTs.  I would have preferred that this checkbox be called "Assume that all FITS files contain RAW color data that needs to be debayered using the CFA pattern declared below".

This is really messy...bear with me.  When this box is checked DSS will not bother looking for the BAYERPAT keyword and will assume that the CFA pattern is declared correctly using the Camera: drop-down selector.  The bottom line is that ONLY users NOT using CaLIGHTs OR astrocam users who want to have DSS calibrate their LIGHT frames  AND they know that the BAYPAT keyword is not being written to their FITS files should have this box checked.

If you have been using CaLIGHTs and you have this box checked then the Camera: selection being made here will override the BAYERPAT keyword added by CaLIGHTs.  If you don't have this box checked then a CaLIGHTs user can benefit from two results:

1)Any color LIGHT frames will be displayed with the correct colors in DSS.

2)If a CaLIGHTs user decides to mono bin a color LIGHT frame(yes...this is possible with CaLIGHTs), CaLIGHTs will not include the BAYERPAT keyword which will cause DSS to  understand that the calibrated LIGHT frame is to be interpreted as a monochrome image.

When I originally wrote CaLIGHTs the folks at DSS did not support the BAYERPAT keyword so the only way to have DSS interpret a FITS file was to have this box checked and declare what the correct CFA pattern was.  The CaLIGHTs help file was written with this initial understanding.  Obviously it needs to be modified to reflect this new normal.  The help file will be corrected for the next version of CaLIGHTs.

Peter

[wols...@gmail.com]

As of December 19th, 2020 at 3:30PM I modified the installer link to include the updated version of the help file. I did not bother to update the version number.

Peter

[Scott Kuchma]

Hello Peter ,

Well , hold up a minute . Do you know which release of DSS reflects these changes ? I'm assuming that since I'm not seeing these issues , at least I think I'm not , then the DSS issue was not in v4.2.3 ?

Then is this now a combination upgrade effort ? We should Upgrade to CaLIGHTs v3.10.92  *AND* have the latest DSS , v4.2.5 , installed too ?

I assume the new Help File will now tell us to deselect the entry that we had to have have selected before ? Does this account for the CFA Pattern being seen as not RGGB ?

You know what this means don't you ? You're going to have to add Stacking to some future release of CaLIGHTs . LOL ! Just kidding . We don't want to burn you out .

Scott..............

[wols...@gmail.com]

Scott,

This issue is confusing.  The support for the BAYERPAT FITS keyword in DSS was introduced in version 4.2.4 beta 5.  This is also the version where they updated to the latest version of LibRAW.  I had started experimenting with the BAYERPAT keyword way back and I actually removed the logic for including BAYERPAT in the calibrated LIGHT frame in v3.1.0.50 because it seemed too confusing. In v3.1.0.80 I re-introduced BAYERPAT support and it has been supported since then.

You should upgrade to v3.1.0.92 and upgrade DSS to V4.2.5.  Trying to use older versions is too confusing and not supported by anyone writing software. The new help file has been corrected and the correct CFA Pattern will be seen.  Did I mention that this CFA pattern issue only affects a very small handful of DSLRs.  So far only John's camera is in this category.  Scott...you can leave this box checked or not checked because your CFA Pattern does not change...your masters are not affected either but I would ask that you regenerate them because I don't want to wait six months to find out that there is an issue.  I need to know now...while all the modifications I made are fresh in my mind.

"Does this account for the CFA Pattern being seen as not RGGB ? "  Do you mean to say that your camera is now reporting a different CFA pattern?

Peter 

[wols...@gmail.com]

Scott,

While it is true that, in your case, it does not matter whether the box is checked or not...their is one specific case where you should have the box NOT checked.  I would suggest, and the help file also suggests that you NOT have this box checked.  This specific case is when you wish to experiment with LRGB processing and you want to create a luminance image using CaLIGHTs.  CaLIGHTs allows you to do this by selecting a mono binning.  When you do this CaLIGHTs will generate a monochrome image and when it is written to disk the BAYERPAT keyword is not included.  If you have the DSS box NOT checked then you will see that DSS will interpret your calibrated mono LIGHT frame as a greyscale image and it will be displayed and stacked as a greyscale image.  It's an advanced topic but it is available with CaLIGHTs.  Basically...when the DSS box is NOT checked, DSS will interpret a FITS file containing a BAYERPAT keyword as a color image that needs to be debayered.  If the FITS file does not contain a BAYERPAT keyword then DSS will interpret the image as being a greyscale image that does not require debayering.

Peter

[Scott Kuchma]

Hi Peter ,

My comment regarding the CFA Pattern was in reference to John M's Post where he was now getting a GRBG when he thought it was supposed to be RGGB . Maybe I'm still missing the point . No worries as long as it works .

Anyway , I will Install the 92 version of CaLIGHTs and the 4.2.5 version of DSS , redo all my Cal Masters and then retry a set of Lights to compare with what I've done already this week .

Might take a while but I'll let you know when I'm done .

Scott...........

[Scott Kuchma]

Hi Peter ,

OK , so I'm on CaLIGHTs v3.10.92 and I've started redoing all my Masters and I'm getting something odd .

I don't see any Camera Info or File Details and the File name is odd showing MasterFLAT ISO0 .MSTR .

I assume that's because there are no Details being read ?

Previous Version pulled the ISO and Shutter Speed from the CR2 File and added that to the Master name...

Have I missed a new Setting or do I just need to hand enter the details in the Filename Window ?I have tried all the other RAW CR2 Files I have and got the same thing .

I'm on hold for now until I hear from you .

Scott................

[Scott Kuchma]

Also noticed that on Startup , the temporary Window shows the current version of CaLIGHTs but there is now no version given for the LibRAW . Is it not at 2.0.xx ?

Scott..........

[Scott Kuchma]

One more item to the list . I simply tried importing a few Light frames and using an old FLAT Master and got this...

So it seems my CR2 Files are not being read by this version of CaLIGHTs .

Scott............

[wols...@gmail.com]

Scott you must be using a 32b version of windows.  64b computers are the standard these years.  CaLIGHTs v3.1.0.92 uses a 64b version of LibRAW and I know that trying to use a 64b program on a 32b operating system will give you these symptoms.  That's why the splash screen that appears when CaLIGHTs first run does not report the version of LibRAW being used.  There are two solutions for you. 

a) Close CaLIGHTs  and navigate to the "c:\Program Files (x86)\CaLIGHTs\CR3" Support folder and double click the program called VC_redist.x64.exe  This will try to install the C+ 64b libraries required to support a 64b DLL.  This may, or may not work. There are some 64b computers that are running a 32b version of windows.

b) Close CaLIGHTs and navigate to the  c:\Program Files (x86)\CaLIGHTs\Backup Copies folder.  Copy the two files [libraw.dll and PROCRAW.exe] into the main CaLIGHTs folder [  c:\Program Files (x86)\CaLIGHTs ]  This will overwrite the 64b versions of these files with their 32b versions. This will work fine for your camera.  If you get a camera like John's then you will need a 64b computer. I will maintain the Backup Copies folder for the foreseeable future in any updates.

One of my tasks for this winter is to port CaLIGHTs to 64 bits and no longer support 32b windows.  I might be able to rebuild these updated LibRAW libraries in 32b.  The LibRAW folks were only supporting 64b at the time.

Peter

[Scott Kuchma]

Hi Peter ,

I can't see why this particular version would suddenly stop working with my System which is 64-bit !

Any other ideas ?

Scott..............

[wols...@gmail.com]

Scott,

That is very good to hear that you are running a 64 bit machine.  I really want to port CaLIGHTs to 64b.  I don't think you can buy a 32b computer anymore.  Have you tried solution a)?  Your computer may not have the redistributables for 64b C++.  My 64b computer initially didn't have them installed.  You may also want to try and download the installer again. You may have downloaded a corrupt copy.  You should also check that your antivirus software isn't generating alarm messages.  Let me know what happens.

Peter 

[wols...@gmail.com]

I see you are running window 7.  You may want to consider upgrading to windows 10.  I upgraded from window 8 to windows 10 and it went very well.

Peter

[Scott Kuchma]

Hi Peter ,

Running the VC_redist.x64.exe seems to have resolved today's issues . I have CaLIGHTs 3.10.92 showing LibRAW 0.20.0-WorkInProgress on startup and all the Details showing for each CR2 File again .

I still don't understand why every single Update before this version installed without me having to do this step .

Even though this Laptop is showing the "Upgrade to Win10" Option , the Microsoft Site says the Hardware is not compatible for Win10 . I'll have to wait until April to see if I'm getting anything back from Taxes before any Upgrade will happen .

Too many other areas requiring financial attention around the House right now . LOL .

I may go back and retry DSS 4.2.5 and see if this has stopped it from crashing as well . Never ends does it ?

Thanks for all the help .

Scott.............

[wols...@gmail.com]

Every single Update before this version was installing a 32b version of LibRAW.  V3.1.0.92 is the first version to use the 64b version of LibRAW by default.  I think that DSS 4.2.5 was crashing for the same reasons that CaLIGHTs had.  Your computer lacked the C++ support files for 64b compiled libraries (DLLs).

Glad you are back in business.

Peter

[Scott Kuchma]

Just a DSS 4.2.5 update . I migrated to this latest version and downloaded the 64-bit option . Still can't get it to run any processes . It crashes when I try to Register or Stack . So I went back and tried the 32-bit version of DSS 4.2.5 . This works fine .

However there is an issue with the Comet Stacking Procedure . If I select Align with Comet or Align with Stars it will work and of course blur the not aligned objects . If I choose Align with Comet and Stars I get a nice comet nucleus and several stars dotted next to each other . Not quite a smear or Star Trail but not what I wanted . I've gone back to the 32-bit DSS 4.2.3 and it works perfectly in that version . Can't find anything on this specific topic but I don't know the DSS Forums very well .

Scott............

[wols...@gmail.com]

Scott,

Is there an error message generated by DSS 4.2.5?  Can you get a screen capture of it.  I just checked using some of the .cr2 files you have provided and DSS 4.2.5 didn't seem to have a problem.  The issues with DSS 4.2.5 won't be related to LibRAW because your calibrated light frames from CaLIGHTs are FITS files.  I thought it might be the 64b support files.  You might want to try loading your .CR2 files directly into DSS and see if it likes them better.  I'm no expert in the comet stacking feature but I did try it this year and I remember having to identify the comet in every light frame.  There has been alot of chatter on this forum regarding comet stacking but I didn't pay much attention to it.  The overall chatter has settled down so I assume that any issues with comet tracking were addressed.

Peter

[Scott Kuchma]

Hi Peter ,

I don't have a screen capture of the error window but it was simply stating that "DSS has encountered a problem processing the Data" and then I had to shut it down . Right now , I'm going to leave it as I still have a few early Image Groups I'd like to run through CaLIGHTs 3.10.92 and if DSS 4.2.3 works then I can move on . This seems to be the month of Software/Firmware issues around here . Had a ton of "pressure imbalance" issues showing up on my thermostat . Reflashed the Firmware and , fingers crossed , that seems to be resolved . I couldn't find any hardware issues as the cause .Then the car started showing "Please turn off the Ignition" when it was in Accessory mode listening to the radio . Went to Ford today and they had to reset the Battery Monitor Program in the Sync 3 Software . I really need things to run a little more smoothly for the rest of the year . LOL .

Scott......

[wols...@gmail.com]

Scott,

Wow...Ghosts in the machine!  Technology is suppose to serve man...not the other way around.

Peter