Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Orion Nebula in natural colour 4K HDR

  • Please log in to reply
7 replies to this topic

#1 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • topic starter
  • Posts: 2,919
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 25 February 2020 - 10:33 AM

Since buying a 4K HDR TV (LG OLED C9) I have struggled to find a way to display astro-images in HDR mode.  As far as I can tell, HDR is only triggered using a 10-bit 4K HDR video source.

 

Using (free) ffmpeg software I have now found a way to generate a 10-bit 4K HDR video, preserving both the natural colour and the natural range of relative intensities.  The latter point will make the image look underwhelming because we are used to images that are highly stretched.  However, from a technical viewpoint it is an interesting challenge.

 

The video (which contains a static image) has been uploaded to YouTube:  

https://www.youtube....h?v=e5ElRam1hZQ

The "2160p 4K" quality option will only be available if you have compatible equipment - my TV screen comes up with an HDR logo when it switches into HDR mode to view the video.

 

A preview is here, which does not do the HDR video justice:

 

orion_preview.jpg

 

 

Watched in HDR, the stars are intensely bright but the YouTube transcoding has created artifacts around the bright stars.  The original video can be downloaded here:

https://drive.google...IYK1IcQENia1WHh

Ignore the terribly washed out preview that Google Drive provides.

 

For those interested in the technical details, the main steps were as follows:

  • Stack the data (unmodified Canon 600D on Tak Epsilon, total exposure 30min)
  • Transform the linear data to the primaries of the Rec2020 colour space
  • Apply the Perceptual Quantizer (PQ) SMPTE-2084 transfer curve to the data instead of the usual Rec 2020 gamma
  •  Run ffmpeg on the resulting 16-bit TIF to generate the 10-bit 4K HDR video
  •  Copy it onto a USB stick for final display on the TV in the lounge!

Mark


Edited by sharkmelley, 25 February 2020 - 10:34 AM.

  • dmdouglass, Alen K, WarmWeatherGuy and 7 others like this

#2 bmhjr

bmhjr

    Apollo

  • ****-
  • Posts: 1,369
  • Joined: 02 Oct 2015
  • Loc: Colorado

Posted 25 February 2020 - 10:52 AM

Well, I played the video on my 4K TV and I have to say that was impressive.  Even the tiny stars and background seemed to show a lot of depth.  It looks so much better than any of my images I have tried to display on my tv (mine look even worse than usual).


  • sharkmelley likes this

#3 sunnyday

sunnyday

    Fly Me to the Moon

  • *****
  • Posts: 5,132
  • Joined: 23 Sep 2019
  • Loc: Ottawa,Canada

Posted 25 February 2020 - 10:59 AM

very interesting video.
thank you


  • sharkmelley likes this

#4 Astroman007

Astroman007

    Voyager 1

  • *****
  • Posts: 11,010
  • Joined: 27 Oct 2017
  • Loc: Northern Ontario, Canada

Posted 25 February 2020 - 11:32 AM

Love that image. Simply foggy with nebulosity.

 

Thank you for sharing!


  • sharkmelley likes this

#5 calypsob

calypsob

    Fly Me to the Moon

  • *****
  • Posts: 5,075
  • Joined: 20 Apr 2013

Posted 25 February 2020 - 02:18 PM

I need to try this on my tv

#6 Kevin Ross

Kevin Ross

    Viking 1

  • -----
  • Posts: 671
  • Joined: 28 Mar 2018
  • Loc: Traverse City, MI

Posted 25 February 2020 - 02:22 PM

Would you mind sharing your ffmpeg command line options?



#7 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • topic starter
  • Posts: 2,919
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 25 February 2020 - 07:08 PM

Would you mind sharing your ffmpeg command line options?

 

Sure, the command is here:

ffmpeg -loop 1 -i orion_rec2020_PQ.tif -r 1 -t 30 -pix_fmt yuv420p10le -c:v hevc -crf 15 -profile:v main10 -level 5.1 -an -preset slower -x265-params colorprim=bt2020:transfer=smpte2084:colormatrix=bt2020nc outfile.mp4

 

The above command takes a single 16-bit TIF image (with the PQ transfer curve applied) and repeats it at a rate of 1 frame/sec for a total time of 30sec.  I'm not saying this is necessarily the best way because I'm just a beginner with ffmeg.  But it does work.

 

The PQ transfer curve was applied in PixInsight using the following PixelMath expression:

c1=107/128;
c2=2413/128;
c3=2392/128;
m1=1305/8192;
m2=2523/32;
v=(c1+c2*exp(ln($T)*m1))/(1+c3*exp(ln($T)*m1));
exp(ln(v)*m2)

 

The variables c1, c2, c3, m1, m2, v must all be listed on the symbols tab of the PixelMath expression editor.

 

Here's a screen grab from PixInsight showing the very washed-out result after applying the PQ function to the linear data:

 

orion_pq_screengrab.jpg

 

 

I have verified the methodology by encoding my step wedge data:

https://www.cloudyni...ge-test-files/ 

 

It's one of the reasons I generated those step wedge test files.  I was able to photograph the TV screen displaying the resulting 4K HDR step wedge video and examine the camera's raw file to check that the TV was definitely displaying each bar with half the intensity of the neighbouring bar.

 

Mark


Edited by sharkmelley, 25 February 2020 - 07:27 PM.


#8 Kevin Ross

Kevin Ross

    Viking 1

  • -----
  • Posts: 671
  • Joined: 28 Mar 2018
  • Loc: Traverse City, MI

Posted 25 February 2020 - 07:38 PM

Thank you very much Mark! :)




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics