Advanced processing of Seestar S50 FITS output - help?

[Cosmic Geoff]

Briefly, I'll explain that the new ZWO Seestar S50 delivers two kinds of output files:

1) A high-resolution .jpeg delivered to the controlling smartphone, and maybe onwards via the Cloud to your image processing PC.  All so simple a child could do it.  An increase in the contrast brings out more detail.

2) A FITS file (accompanied by a low-res .jpg thumbnail) stored in the S50. This can readily be extracted to a PC or laptop via a file browser.  But what is one supposed to do with it?  Apparently the FITS file contains the stacked images, and needs to be stacked by a program e.g. Siril or Deep Sky Stacker, before being processed in one's program of choice (e.g. GIMP).

I have looked at Youtube tutorials, and downloaded various bits of software, but my small brain can't make sense of any of it. 

One  tutorial recommended using Siril using a particular script, which turned out not to be in the default installation, and which I could not find to download.  Fail 1

Sirilic was suggested, which seems to involve signing up to GitHub. And not finding anything to download. Fail 2  

I downloaded Deep Sky Stacker (which I managed to try out a couple of years ago on another PC)  but got nowhere with the S50 FITS. Fail 3

Isn't there a SIMPLE way of doing this?

[happy-kat]

is there a fits file we could access to play with? 

[ONIKKINEN]

Not too familiar with the seestar (have been just occasionally following the discussion) but am i right in assuming that these are all light frames? As in there are no calibration frames such as darks or flats.

If so, you cant actually use most of the scripts in Siril because they assume you have calibration frames, but not to worry it is actually very simple to stack without them since we dont have to worry about calibration. I will briefly go over the simplest way to stack them to an image below.

First set your home folder for Siril to be somewhere you want the temporary files and end result to be, using the button in the image below. So maybe you should create a folder named "siril processing" on a drive that has plenty of empty disk space and choose that.

Then import the .FITS files for siril to use next in the  "Conversion" tab shown below. Either click the + button and search for the files, or select all of them and drag and drop them to the empty field. You must also give this sequence a name and tick the Debayer button if we are working with colour images (assuming so).

Then at this point we have to register the images (star alignment). Go to the "Registration" tab pictured below and use the settings shown in the screen. The most important parts are highlighted. There is another, more robust way to do registration using the two-pass alignment but maybe dont worry about that now. To do that you would select the two-pass method and after that completes you would select the "apply existing registration" to run it again (needs to be done in that order).

This next step is optional, but might as well have a look at this point. The "Plot" tab contains an analysis of the input frames. Here you could de-select the worst images if you want to. If there are massive vertical spikes that are sticking out like a sore thumb, get rid of them here.

Then finally stacking itself. I recommend you use the settings i have below. After all this you have your stacked image.

1 hour ago, Cosmic Geoff said:

Isn't there a SIMPLE way of doing this?

DeepSkyStacker is much simpler, but it doesn't work with every dataset because its quite picky about quality.

[Cosmic Geoff]

2 hours ago, happy-kat said:

Is there a fits file we could access to play with? 

Here's one, for M57.

Header data follows:

SIMPLE  =                    T / file does conform to FITS standard
BITPIX  =                   16 / number of bits per data pixel
NAXIS   =                    3 / number of data axes
NAXIS1  =                 1080 / length of data axis 1
NAXIS2  =                 1920 / length of data axis 2
NAXIS3  =                    3 / length of data axis 3
EXTEND  =                    T / FITS dataset may contain extensions
COMMENT   FITS (Flexible Image Transport System) format is defined in 'Astronomy
COMMENT   and Astrophysics', volume 376, page 359; bibcode: 2001A&A...376..359H
BZERO   =                32768 / offset data range to that of unsigned short
BSCALE  =                    1 / default scaling factor
CREATOR = 'ZWO SeestarS50'     / Capture software
XORGSUBF=                    0 / Subframe X position in binned pixels
YORGSUBF=                    0 / Subframe Y position in binned pixels
FOCALLEN=                  250 / Focal length of telescope in mm
XBINNING=                    1 / Camera X Bin
YBINNING=                    1 / Camera Y Bin
CCDXBIN =                    1 / Camera X Bin
CCDYBIN =                    1 / Camera Y Bin
XPIXSZ  =     2.90000009536743 / pixel size in microns (with binning)
YPIXSZ  =     2.90000009536743 / pixel size in microns (with binning)
IMAGETYP= 'Light   '           / Type of image
STACKCNT=                   11 / Stack frames
EXPOSURE=                  10. / Exposure time in seconds
EXPTIME =                  10. / Exposure time in seconds
RA      =            300.15417 / Object Right Ascension in degrees
DEC     =            22.794444 / Object Declination in degrees
DATE-OBS= '2023-09-15T21:17:52.226397' / Image created time
FILTER  = 'IR      '           / Filter used when taking image
INSTRUME= 'ZWO ASI462MC'       / Camera model
BAYERPAT= 'GRBG    '           / Bayer pattern
GAIN    =                   80 / Gain Value
FOCUSPOS=                 1617 / Focuser position in steps
CTYPE1  = 'RA---TAN-SIP'       / TAN (gnomic) projection + SIP distortions
CTYPE2  = 'DEC--TAN-SIP'       / TAN (gnomic) projection + SIP distortions
CRVAL1  =        300.080461534 / RA  of reference point
CRVAL2  =        22.7485782229 / DEC of reference point
CRPIX1  =        287.141927719 / X reference pixel
CRPIX2  =        878.079833984 / Y reference pixel
CD1_1   =   -0.000640259548805 / Transformation matrix
CD1_2   =   -0.000155699886841 / no comment
CD2_1   =    0.000155330795079 / no comment
CD2_2   =   -0.000640007144967 / no comment
A_ORDER =                    2 / Polynomial order, axis 1
B_ORDER =                    2 / Polynomial order, axis 2
AP_ORDER=                    2 / Inv polynomial order, axis 1
BP_ORDER=                    2 / Inv polynomial order, axis 2
A_0_0   =                    0 / no comment
A_0_1   =                    0 / no comment
A_0_2   =   -2.72399212086E-07 / no comment
A_1_0   =                    0 / no comment
A_1_1   =    3.15089321972E-07 / no comment
A_2_0   =   -4.82895905823E-08 / no comment
B_0_0   =                    0 / no comment
B_0_1   =                    0 / no comment
B_0_2   =    1.01487242366E-07 / no comment
B_1_0   =                    0 / no comment
B_1_1   =    5.82881418145E-07 / no comment
B_2_0   =   -5.54562671461E-07 / no comment
AP_0_0  =    2.36061654639E-05 / no comment
AP_0_1  =    -5.6197243827E-08 / no comment
AP_0_2  =    2.72324015254E-07 / no comment
AP_1_0  =   -6.89411248193E-08 / no comment
AP_1_1  =    -3.1489293254E-07 / no comment
AP_2_0  =     4.8196817389E-08 / no comment
BP_0_0  =   -3.84631989892E-05 / no comment
BP_0_1  =   -1.12292551537E-07 / no comment
BP_0_2  =    -1.0135344279E-07 / no comment
BP_1_0  =    2.21043408709E-07 / no comment
BP_1_1  =   -5.82848681282E-07 / no comment
BP_2_0  =    5.54345404161E-07 / no comment
IMAGEW  =                 1080 / Image width,  in pixels.
IMAGEH  =                 1920 / Image height, in pixels.
END

Stacked_M 57_10.0s_20230915-221030.fit

Edited September 20, 2023 by Cosmic Geoff
Added FITS header

[Cosmic Geoff]

Onnikinen - I tried that, but your method seems to be for a set of fits files, not just one containing multiple images (11 in the M57 example I have posted here).

Edited September 20, 2023 by Cosmic Geoff

[ONIKKINEN]

1 hour ago, Cosmic Geoff said:

Onnikinen - I tried that, but your method seems to be for a set of fits files, not just one containing multiple images (11 in the M57 example I have posted here).

The file you linked does not match in its fits header to the fits header you pasted. The attached file is a stacked image and consists of 5 exposures, yet the fits header of the pasted text consists of 11?

Do you have many files like this or just the one? If many, you can still stack them even if they are already stacked, its just not ideal (ideally you would have the individual exposures, but not sure you can get them out of the seestar).

[MalcolmP]

2 hours ago, Cosmic Geoff said:

11 in the M57 example

I am no expert ( I just like playing with posted images )
I was about to post similar to @ONIKKINEN -  -  ASTAP opens it as a single colour image with a stack count of 5 and a total exp of 10sec.
So yes it is the result of a stack, not the individual subs to be stacked.
Gimp opens it as 3 layers, one named and the other 2 it calls 'background', no indication as to which are the R,G & B layers.
Nice M57,
 and ASTAP also picks out (only just visible) IC1296 close by !
 

Edited September 20, 2023 by MalcolmP
speelin !

[michael8554]

"Apparently the FITS file contains the stacked images, and needs to be stacked by a program e.g. Siril or Deep Sky Stacker, before being processed in one's program of choice (e.g. GIMP)."

Hopefully that should read  "contains the  (stacked)  images".

So that you have the ability to stack a number of unstacked images with Siril or DSS.

Michael

[Oddsocks]

15 hours ago, Cosmic Geoff said:

 

I have looked at Youtube tutorials, and downloaded various bits of software, but my small brain can't make sense of any of it. 

One  tutorial recommended using Siril using a particular script, which turned out not to be in the default installation, and which I could not find to download.  Fail 1

Sirilic was suggested, which seems to involve signing up to GitHub. And not finding anything to download. Fail 2  

I downloaded Deep Sky Stacker (which I managed to try out a couple of years ago on another PC)  but got nowhere with the S50 FITS. Fail 3

Isn't there a SIMPLE way of doing this?

I may have misinterpreted your opening topic but I think that you have two separate questions here.

1. Can I reach inside the SeeStar, or it's output .fits image, and grab the individual frames that SeeStar stacked internally so that I can build an image using more advanced techniques?

2. What I can I do with the standard output .fits image from the SeeStar to give more pleasing image than is possible with the SeeStar's basic .jpeg output image?

I do apologise if I've got that wrong and if so you can ignore my response below....

-------------

The SeeStar's jpeg format export option is a compressed image format and depending on its bit depth you could enhance that image using various filters and tools in many "Photoshop-type" general-purpose image processing applications.

The limiting factor of that approach will be the compressed bit-depth of colour jpeg images, there is only so much you can do before the image starts to look "pixelated" and noisy.

The .fits format image is uncompressed raw data and although for the SeeStar we don't know whether ZWO apply any kind of automatic dark current and bias subtraction to the output image nevertheless you can extract more information and produce a more pleasing image (I think) by post-processing the image in an application that is specialised (or optimised) for handling .fits images.

The output .fits image from the SeeStar is already multiple frames stacked internally, as others have already explained above, and AFAIK you can't disassemble that image any further. I'm not aware of any methods that allow you to reach inside the SeeStar itself to extract the individual frames that it used/uses to build its stacked output image.

You mention above that you have previous experience with Deep Sky Stacker.

For a single (internally stacked) output .fits image from the SeeStar then Deep Sky Stacker has no useful purpose, but if you were to collect multiple (internally stacked) SeeStar .fits images of the same target then you can register and stack those together in DSS to reduce noise further and allow you to stretch the resulting image more to reveal fainter detail.

With a .fits image, either a single frame that was internally-stacked in the SeeStar, or an externally registered and stacked combined image (DSS etc), created from multiple SeeStar internally stacked images, then you can do much more in post processing.

I've attached two images below using your raw .fits file, one of which was very quickly (less than five minutes) post-processed in PixInsight, colour calibrated, BlurXterminator and NoiseXterminator applied then stretched and colour saturation increased, finally Histogram transformation applied.

I've only applied a very basic process to your image, there are multiple tools available to you once you get to work on .fits images, it just depends how much time you want to invest in post processing and how much time you are willing to allocate to gathering as much raw data on a single target as possible.

The second image shows the raw .fits image straight from the SeeStar with the same level of stretch applied.

I hope that gives you some idea of what is possible in post processing the .fits image, and, as described above if you had multiple internally-stacked .fits images you could stack those together which would allow you go further in post processing.

HTH

William.

Processed .fits in Pixinsight.

 

Raw .fits from SeeStar.

[Cosmic Geoff]

I made a mistake posting the header text above. That one is for M27, and the M57 one actually indicates that it contains five images.  

Below I have posted the iphone .jpg for M57. This 1.4 MB 1080px1920px file has not been processed by me in any way.

[Cosmic Geoff]

I have found partial answers to my own question.  Earlier editions of the Seestar app and firmware, including those I used for my 'first light', only let you save stacked FITS images to the S50.  Later editions, now available, apparently let you save an array of FITS images (i.e. not stacked).

Various beta testers (who did not tell us what app and firmware versions they were using) evidently had advance copies of the non-stack versions, and posted Youtube videos explaining how to stack the files and then process them (usually with Siril).  If you have stacked FITS S50 images, you can just skip the part of the instructions where they tell you to stack the images.

The FITS images are dark, and require a considerable stretch (or auto-stretch) before you can see much.  A simpler workflow I discovered is to open and stretch the image in ASTAP, export as a TIFF and then process it in GIMP.

[Giles_B]

Thanks - just finally had my first light after a cloudy week, and this is just the question I was asking and the answers I was looking for! The basic guide to Siril is really helpful. Could anyone recommend a tutorial that I could follow if I wanted to do more advanced processing of s50 images? I much prefer a written tutorial to follow over a video.