Return to the Cloudy Nights Telescope Reviews home pageAstronomics discounts for Cloudy Nights members
· Get a Cloudy Nights T-Shirt · Submit a Review / Article

Click here if you are having trouble logging into the forums

Privacy Policy | Please read our Terms of Service | Signup and Troubleshooting FAQ | Problems? PM a Red or a Green Gu… uh, User

Astrophotography and Sketching >> DSLR & Digital Camera Astro Imaging & Processing

Pages: 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | (show all)
mmalik
Postmaster
*****

Reged: 01/13/12

Loc: USA
Re: Learning PixInsight new [Re: mmalik]
      #5746842 - 03/21/13 02:56 AM Attachment (17 downloads)

Auto DD in ImagesPlus of PixInsight Integration.FIT looks like this...

Post Extras: Print Post   Remind Me!   Notify Moderator  
mmalik
Postmaster
*****

Reged: 01/13/12

Loc: USA
Re: Learning PixInsight new [Re: mmalik]
      #5746843 - 03/21/13 02:56 AM Attachment (11 downloads)

Auto DD in ImagesPlus of ImagesPlus CombineFilesAVG.FIT looks like this...

Post Extras: Print Post   Remind Me!   Notify Moderator  
mmalik
Postmaster
*****

Reged: 01/13/12

Loc: USA
Re: Learning PixInsight new [Re: mmalik]
      #5746845 - 03/21/13 02:57 AM Attachment (17 downloads)

Histogram of PixInsight Integration.FIT looks like this...

Post Extras: Print Post   Remind Me!   Notify Moderator  
mmalik
Postmaster
*****

Reged: 01/13/12

Loc: USA
Re: Learning PixInsight new [Re: mmalik]
      #5746846 - 03/21/13 02:57 AM Attachment (13 downloads)

Histogram of ImagesPlus CombineFilesAvg.FIT looks like this...

Post Extras: Print Post   Remind Me!   Notify Moderator  
mmalik
Postmaster
*****

Reged: 01/13/12

Loc: USA
Re: Learning PixInsight new [Re: mmalik]
      #5746859 - 03/21/13 03:09 AM

Regardless of the greenish looking auto-stretched Integration.FIT in PixInsight, results of my testing thus far tell me that PixInsight's "natively" converted/aligned/integrated file will be almost impossible to process based on DD and histogram findings above. Your feedback will be welcome if you suggest otherwise? Regards

Post Extras: Print Post   Remind Me!   Notify Moderator  
Falcon-
Post Laureate
*****

Reged: 09/11/09

Loc: Gambier Island, BC, Canada
Re: Learning PixInsight new [Re: mmalik]
      #5746866 - 03/21/13 03:18 AM

That looks fairly expected. PixInsight's histogram may *look* odd at first but it is displaying data that originated in a 12bit or 14bit space in a 32bit space - once the black and white points are set correctly it will look very much more like the one from ImagesPlus.

The other issue is the fact that PixInsight has *at this state* done no colour balance corrections AT ALL, not even to account for the fact that there are 2x the number of green pixels.

From here you get into post processing including colour calibration. Check out things such as Harry's video tutorials. He has examples with background gradient removal (DBE) and colour calibration.


Post Extras: Print Post   Remind Me!   Notify Moderator  
Falcon-
Post Laureate
*****

Reged: 09/11/09

Loc: Gambier Island, BC, Canada
Re: Learning PixInsight new [Re: Falcon-]
      #5746868 - 03/21/13 03:20 AM

BTW - if you can upload that 32bit floating point FITS to some place and link it I would be happy to give your green image a try in PixInsight here....

Post Extras: Print Post   Remind Me!   Notify Moderator  
mmalik
Postmaster
*****

Reged: 01/13/12

Loc: USA
Re: Learning PixInsight new [Re: Falcon-]
      #5746880 - 03/21/13 04:24 AM Attachment (9 downloads)

Quote:

if you can upload that 32bit floating point FITS to some place and link it I would be happy to give your green image a try in PixInsight




Sure; PixInsight integration.fit uploaded here....
EDIT: Download the latest integration.fit in case you had downloaded an earlier version.

If you can, please elaborate on granular tasks you may perform so folks could follow along for learning sake as well as any future documentation I might create. Regards


Note: ImagesPlus CombinedFilesAvg.FIT can be found here....


While we are on the subject, if I try opening an ImagesPlus converted/aligned/combined file (CombinedFilesAvg.FIT) in PixInsight I get following error; please see if you or someone can help. Thx


Post Extras: Print Post   Remind Me!   Notify Moderator  
bluedandelion
Carpal Tunnel
*****

Reged: 08/17/07

Loc: Hazy Hollow, Western WA
Re: Learning PixInsight new [Re: mmalik]
      #5747297 - 03/21/13 10:39 AM

Try this. Its part of my workflow and works with images acquired with my 350D.

Ignore the greenish cast. Trim edges of the the stack and do a DBE correction right away. Most of the color balance will be restored. An auto screen stretch should look mostly neutral.

Do a further color correction via the Star method since there are no large galaxies in the field. If the greens still persist use SCNR.

Ajay


Post Extras: Print Post   Remind Me!   Notify Moderator  
jsines
sage
*****

Reged: 09/06/11

Loc: Berkley. Michigan
Re: Learning PixInsight new [Re: pfile]
      #5747456 - 03/21/13 12:02 PM

Quote:


in the RGB image, the red pixels are on the red plane, green on the green and blue on the blue. on a given plane, wherever there would be pixels of a different color, there are black pixels on that plane. the CFA image is sort of more like the sensor itself - for a canon camera, the red pixel is next to a green pixel and above the other green pixel. the blue pixel is diagonally opposed to the red pixel.
at some point in PixInsight history the Debayer process could only handle one type of image... now i can't remember which. but now Debayer can handle CFA or RGB bayer images so it does not matter which one you use.





thanks, pfile. I appreciate your help.


Quote:


After running Batch DeBayer, Star Alignment and Image Integration, Auto Stretch looks like this...





That is in part because you have the channels linked. Uncheck the icon at the top left, the one with the chains, and you will unlink your channels, then click the orange/black circle to auto-stf. It's the first thing I do when I open Screen Transfer Function...always.


Quote:


Sure; PixInsight integration.fit uploaded here....
EDIT: Download the latest integration.fit in case you had downloaded an earlier version.
If you can, please elaborate on granular tasks you may perform so folks could follow along for learning sake as well as any future documentation I might create. Regards




I'll also download your integrated file tonight when I get to my home computer with PI on it and see what I get. If possible, could you upload the original CR2 files that have in-camera noise reduction? I could run them through the whole process to see what I end up with.

Hope this helps.


Post Extras: Print Post   Remind Me!   Notify Moderator  
pfile
Post Laureate


Reged: 06/14/09

Re: Learning PixInsight new [Re: jsines]
      #5747515 - 03/21/13 12:34 PM

cant download the fits files without a microsoft account which i don't have/don't want... can you host the files somewhere that does not need a login?

Post Extras: Print Post   Remind Me!   Notify Moderator  
mmalik
Postmaster
*****

Reged: 01/13/12

Loc: USA
Re: Learning PixInsight new [Re: pfile]
      #5747608 - 03/21/13 01:10 PM

Quote:

cant download the fits files without a microsoft account which i don't have/don't want... can you host the files somewhere that does not need a login?




I see the problem; let me try fixing if I could. Thx


Post Extras: Print Post   Remind Me!   Notify Moderator  
Ugmul
super member


Reged: 05/17/11

Loc: Tucson AZ
Re: Learning PixInsight new [Re: mmalik]
      #5747635 - 03/21/13 01:20 PM

You can always run a background neutralization tool first to remove the greens. Follow with DBE tool to remove gradients from your image. After that run a color calibration to bring the color channels back in line.

Remember to do these on the linear fit. Using the screen transfer to preview.

Colors will always look odd until you do these steps. Green or red backgrounds are very common with DZLR.


Post Extras: Print Post   Remind Me!   Notify Moderator  
guyroch
Vendor (BackyardEOS)
*****

Reged: 01/22/08

Loc: Under the clouds!
Re: Learning PixInsight new [Re: Ugmul]
      #5747722 - 03/21/13 02:19 PM

Being on the PixInsight learning wagon myself I'm loving this thread. Keep it alive!

Thanks to all who are posting tips.

Guylain


Post Extras: Print Post   Remind Me!   Notify Moderator  
pfile
Post Laureate


Reged: 06/14/09

Re: Learning PixInsight new [Re: guyroch]
      #5747795 - 03/21/13 03:00 PM

also it may not make sense to open fits files from one application in another. the fits standard is poorly defined, so lots of representations are legal.

pixinsight saves fits files as 32-bit (or 64-bit) floating point numbers in the range [0..1]. other apps (like maxim) save fits files in the range [0.0...65535.0].

the fits reader in pixinsight does have a rescaling function so if you know how the data is represented you can handle it. images plus, i don't know. it might be interpreting the PI data differently and that's why it looks so dark. theoretically though all the data should still be in there.


Post Extras: Print Post   Remind Me!   Notify Moderator  
mmalik
Postmaster
*****

Reged: 01/13/12

Loc: USA
Re: Learning PixInsight [Re: bluedandelion]
      #5747971 - 03/21/13 04:20 PM Attachment (8 downloads)

Quote:

Ignore the greenish cast. Trim edges of the stack and do a DBE correction right away. Most of the color balance will be restored. An auto screen stretch should look mostly neutral.




Thanks for the tip; I tried running DBE on Integration.FIT and results look better...no more green cast upon Screen Auto Stretch. See pic below...


Quote:

Most of my DSLR images were captured with Nebulosity and I saved the images as compressed FITS. PixInsight's stacking routine has problems with this. So with these images I do calibration and stacking with the new batch process facility in Neb and then move to PI for registration and final stacking.





Thing I am not so sure about if DBE is the "right" way to fix the green cast of 'conversion/alignment/integration' processes or is this just masking the problem? Ideally, wouldn't one want to have a normal looking integration.fit upon screen Auto Stretch like in other programs? At least I would like to see an 'OK' looking conclusion to a core step in the process instead of hoping the next will take care of the previous.


Would it be a correct presumption that 'conversion/alignment/integration' processes are NOT optimized in PixInsight? NOT a critique of PixInsight in any way, just would like an honest assessment of the 'conversion/alignment/integration' problem if there is one and if one should skip these processes in PixInsight if they may be deficient?


If true, this complicates matters further. I am NOT able to process CR2s converted/aligned/combined in ImagesPlus for e.g., DBE processing in PixInsight?

Quote:

...if I try opening an ImagesPlus converted/aligned/combined file (CombinedFilesAvg.FIT) in PixInsight I get following error; please see if you or someone can help.







Post Extras: Print Post   Remind Me!   Notify Moderator  
harry page 1
sage


Reged: 07/25/09

Re: Learning PixInsight [Re: mmalik]
      #5747996 - 03/21/13 04:42 PM




Thing I am not sure about if DBE is the "right" way to fix the green cast of 'conversion/alignment/integration' processes or is this just masking the problem? Ideally, wouldn't one want to have a normal looking integration.fit upon screen Auto Stretch like in other programs? At least I would like to see an 'OK' looking conclusion to a core step in the process instead of hoping the next will take care of the previous.
--------------------------------------------------------------------------------



The problems with other programs is that they carry out a automatic white balance ( sometimes ok sometimes not)
Pi will only do what you tell it to do , gives you more control which you will come to love

Harry

Edited by harry page 1 (03/21/13 04:43 PM)


Post Extras: Print Post   Remind Me!   Notify Moderator  
Falcon-
Post Laureate
*****

Reged: 09/11/09

Loc: Gambier Island, BC, Canada
Re: Learning PixInsight [Re: harry page 1]
      #5748055 - 03/21/13 05:13 PM Attachment (23 downloads)

Here is my quick go at your data. The green cast is actually not all that strong, the auto-STF makes it look worse then it is.

In any case this is what I did:

1) cropped out the black borders left over from alignment
2) used BackgroundNeutralization
3) ran SCNR on the green channel
4) ran ColorCalibtration (but it did not change things very much)
5) used HistogramTransformation to do initial rough-stretch
6) ran ACDNR to reduce noise a bit
7) used Curves a couple of times to make smaller changes to intensity and saturation
8) Created a StarMask and ran MorphologicalTransformation to reduce star sizes a bit (was not really needed here, but I want to see what difference it made - answer was not much )
9) one final very small HistogramTransformation tweak.


Post Extras: Print Post   Remind Me!   Notify Moderator  
pfile
Post Laureate


Reged: 06/14/09

Re: Learning PixInsight [Re: Falcon-]
      #5748080 - 03/21/13 05:21 PM

the "right" way to color calibrate images is to do color calibration. however, even from the darkest skies almost every astrophotograph is going to have gradients, so DBE is pretty much always something you are going to do. by default DBE does not normalize the image (which is why the color cast goes away) but if you check 'normalize' you'll see that your color issues come back.

anyway, you are kidding, right? the calibration, registration and alignment in pixinsight is the best there is, hands down.

you're jumping way in over your head and just going wild all over the place, which is fine, because that's a good way to learn. but don't come to wrong conclusions about the software because of user error.


Post Extras: Print Post   Remind Me!   Notify Moderator  
Falcon-
Post Laureate
*****

Reged: 09/11/09

Loc: Gambier Island, BC, Canada
Re: Learning PixInsight [Re: Falcon-]
      #5748094 - 03/21/13 05:24 PM

Regarding the processing I did - I did not try and do anything to preserve or accentuate the details in the horse-head or the reflection nebulosity - a bit of time spent there would likely improve things quite a bit. Also I did not use DBE since there really was no clean background to work with to do background extraction. With a narrow FOV like this entirely full of nebulosity DBE is likely to remove some of your nebula colour unintentionally so I used BackgroundNeutralization instead.



An important tip for ColorCalbiration and BackgroundNeutralization processes is to make sure you set the "Upper Limit" for the background level. It defaults to 0.1 while in your image the background was closer to .2. If you do not do this the process will have a hard time keeping the background a neutral/balanced colour.

The way I figure out what level to set the Upper Limit to is by using the HistogramNeutralization process window. In the lower-left of the process window is a checkmark. Enabeling that will cause the histogram displayed to always be whatever image View you are currently working on. If you hover your mouse over the histogram graph it will show you the coordinates of the mouse pointer/crosshair in text just below the histogram graph. Put the crosshair just to the left of the base of the histogram's sharp background peak and you have your upper limit value. In this case I *think* it said 0.18something so I used 0.2 in the BackgroundNutralization process.


Post Extras: Print Post   Remind Me!   Notify Moderator  
Pages: 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | (show all)


Extra information
5 registered and 22 anonymous users are browsing this forum.

Moderator:  Dave M, WOBentley, tecmage 

Print Thread

Forum Permissions
      You cannot start new topics
      You cannot reply to topics
      HTML is disabled
      UBBCode is enabled


Thread views: 11175

Jump to

CN Forums Home


Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics