Jump to content


Photo

Raw Data available for Tarantula Nebula 17 hours

  • Please log in to reply
47 replies to this topic

#1 Terry R

Terry R

    Sputnik

  • -----
  • Posts: 48
  • Joined: 31 Jan 2013
  • Loc: Australia

Posted 17 April 2013 - 05:40 AM

Hey guys,

I submitted the image, Tarantula Nebula under Urban Skies - 17 Hour exposure in the Feb image/sketching contest. Well, why not let others play with the data. I am making it available under the condition that you show the results and let us know what you did. I would like to know as I am trying to learn this image processing stuff.

I know there is coma and image tilt – and it kills me. I’m still working that stuff out. Several nights the scope was buffeted by winds, but in the end, there is a lot of data available. This image started out as an experiment. I think it took seven nights when stuff actually worked. On those nights, the drill was dragging the mount out and setting the gear up from scratch, polar aligning, imaging, and then tearing it all down again after a nights imaging. It seems that the norm here is for the clouds to roll in after setting up and polar aligning. That’s harsh. Well, at least I mastered the whole polar alignment in the city with no pole star (Southern Hemisphere).

Location:
Melbourne Australia. (15 km from CBD)

Gear:
Vixen R200SS with comma corrector (f4 8 inch newt) – stock focuser
Astro-Physics AP900 mount.
First generation ST-8300 mono running at -15C
Baader Narrowband Filters: SII 8nm, H-alpha 7 nm, and OIII 8.5nm

Exposure times:
SII 17 x 20 min mapped to Red
Ha 18 x 20 min mapped to Green
OIII 16 x 20 min mapped to Blue

Pre Processing
CCDStack2 was used to calibrate, register, normalize, perform data rejection, and final stacking. The three FITs images are not sharpened in anyway. They are just as they came out of the camera – ready for processing – so I’m not hiding any mistakes here. This is meant to be a learning exercise.

Data in the raw is available at…

https://www.dropbox....y0j8/sIpaxhHYDI

The three images are in Fits format – 32 meg each. One for SII, Ha, and OIII.

Have fun; let’s see what you can do.

Terry Robison

Attached Files



#2 Mark72

Mark72

    Viking 1

  • -----
  • Posts: 670
  • Joined: 19 Sep 2012
  • Loc: Munich, Germany

Posted 17 April 2013 - 06:19 AM

Thanks for sharing Terry!

Do you have the stars/RGB by any chance?

Mark

#3 Terry R

Terry R

    Sputnik

  • -----
  • Posts: 48
  • Joined: 31 Jan 2013
  • Loc: Australia

Posted 17 April 2013 - 06:31 AM

Sorry, I live in the city. It's narrowband or nothing. But I can offer you narrowband. :) You could try I Bi-Colour image. I think that's done with Ha and OIII to give a RGB like image.

Terry

#4 Mark72

Mark72

    Viking 1

  • -----
  • Posts: 670
  • Joined: 19 Sep 2012
  • Loc: Munich, Germany

Posted 17 April 2013 - 06:52 AM

OK, Thanks Terry.

I was just asking for star colours - 1 min subs would do;)

Mark

#5 microstar

microstar

    Apollo

  • -----
  • Posts: 1223
  • Joined: 05 Jan 2008
  • Loc: Canada

Posted 17 April 2013 - 10:49 AM

Thanks Terry. I'll give it a shot, although work pressures will keep me from doing so for the next week or so. I'd like to see it in the natural color narrowband palette I use. It's a stunning object that I've only ever viewed once in a 2007 visit to your area so it will be a treat to work on it.
...Keith

#6 David Ault

David Ault

    Messenger

  • -----
  • Posts: 420
  • Joined: 25 Sep 2010
  • Loc: Georgetown, TX

Posted 17 April 2013 - 12:44 PM

Terry,

Thanks for posting this. I would love to see this in person, but this is almost as good as being there and better in a lot of others, especially when the data is so fantastic! Below is my version which I processed in PixInsight, blending SII, Ha & OIII into an RGB approximation. I made some HDR tweaks to try and make the core more visible. I also tried some deconvolution which had some good and bad results. Even with masking it was difficult (at least with my current knowledge) to get improvement in most of the image without causing some artificial appearance in the core. It was fun working with this data and I really appreciate you putting it out there for others to muck with ;)

Regards,
David

Attached Files



#7 Scotty H

Scotty H

    Vostok 1

  • -----
  • Posts: 147
  • Joined: 15 Nov 2010
  • Loc: Surrey,UK

Posted 17 April 2013 - 04:18 PM

Hi Terry

Thanks for the data

Used Pixinsight using histogram stretch, HDR Transformation then combined using pixelmath SII for R/K, Ha for Green, OIII for Blue to create RGB image, extracted luminance from RGB, another Histogram stretch on luminance then recombined to RGB image, saved as 16bit Tiff. Transferred to CS6 for selective colouring, soft colour gradient removal using a mask then played with saturation and unsharp mask. Saved as JPEG for here. Phew :p

Scott

Attached Files



#8 Terry R

Terry R

    Sputnik

  • -----
  • Posts: 48
  • Joined: 31 Jan 2013
  • Loc: Australia

Posted 17 April 2013 - 08:13 PM

Looks cool. I like all that extra faint detail you guys are pulling out of the data. In the orientation presented, the bottom of the image is moving into the Large Magellanic Cloud, and the top of the image is just on the rim of the Large Magellanic Cloud. At first I thought I had a large gradient in the image.

Have a look at http://apod.nasa.gov...d/ap090916.html and you will see what I mean. The image is almost in the same orientation. It is a very bright object with a bright bits all over.

Terry

#9 kbev

kbev

    Vostok 1

  • -----
  • Posts: 183
  • Joined: 29 Dec 2010
  • Loc: Far, far east Mesa

Posted 18 April 2013 - 03:02 AM

Wow, what a great learning experience! I tried for a more normal color balance so I looked around and found a thread on the PI forum about combining these channels for a psuedo-RGB, what I used was pixelmath for the combine with the following settings:
Red = SII + 0.78*Ha
Green = 0.15*Ha + OIII
Blue = OIII

After that it was a pretty standard workflow, final touchup was done in CS5. Here it is:
Posted Image

#10 steveyzh

steveyzh

    Vostok 1

  • -----
  • Posts: 116
  • Joined: 20 Feb 2011

Posted 19 April 2013 - 07:29 AM

Thanks Terry for the great raw data! :)
Posted Image

Processing steps:
1. Ha MDL DDP(FFT lowpass 10%cutoff)
OIII MDL DDP(FFT lowpass 15%cutoff)
SII MDL DDP(FFT lowpass 17%cutoff)
2. J-P Tone mapping method for Ha,Oiii,Sii(after DDP)
3. DBE in Pixinsight LE 1.0 for Sii tonemap layer;
4. Creat Luminance from Ha+Oiii(lighten 10%)+Sii(lighten 10%) in Photoshop CS3
5. Creat SHO Tone mapping Color in Photoshop CS3, Hubble Palette selective color;
6. Luminance+SHO in Photoshop CS3

Steve

#11 snommisbor

snommisbor

    Viking 1

  • *****
  • Posts: 890
  • Joined: 15 Jun 2009
  • Loc: Cedar Park, TX

Posted 19 April 2013 - 10:00 AM

Those are all awesome but I must say I am partial to the Hubble Palette of Steve's, but Terry yours actually shows off the gas and dust a little better than any of them. The more muted colors makes it look more mysterious which is a very cool look. :bow:

#12 Miles Zhou

Miles Zhou

    Lift Off

  • -----
  • Posts: 1
  • Joined: 20 Apr 2013
  • Loc: MA, USA

Posted 20 April 2013 - 09:15 AM

Thanks for the raw files. Here is mine:
http://orianhealth.c...e:Tarantula.jpg

Process:
MDL: DDP+Gamma(0.7)
PI: DBE
PS: Balance RGB=SHO; L=H(100%)|S(15%); Highlight Overlay; and final touch (Brightness/Contrast)

#13 microstar

microstar

    Apollo

  • -----
  • Posts: 1223
  • Joined: 05 Jan 2008
  • Loc: Canada

Posted 20 April 2013 - 09:47 PM

Here's my effort in a slightly different palette. The Ha, OIII and SII stacks were given a mild DDP as 32bit files before conversion to 16bit and importing into Photoshop CS5. The lights were layered as a natural color narrowband version with R=Ha+30%SII; G=OIII, B=OIII+24%Ha, and 30% SII brought in as a Yellow channel. All color channels were generated using clipping masks. I find that layering the SII as a Yellow channel balances the color of the stars better and highlights the SII regions in orange to distinguish the SII regions better. The large and medium features were sharpened using Topaz Labs Detail. Noise in the fainter region was reduced using a Noise Ninja layer with an inverse mask. Carboni's Tools were used to Increase Star Color, and star size was reduced using Annie's Actions Reduce Star Size. A final color balance was applied using Levels, with a final tweaking in Topaz Detail, then the background darkened using a Soft Light layer with an inverse mask.

Thanks for putting your data files up Terry. It was fun and very interesting to see the different versions; I love the work people are able to do with PixInsight, but I'm just not up for re-inventing my processing workflow and getting past PI's steep learning curve. This wasn't the same as processing my own lights somehow, but it was great to get a chance to work on a region that I'm unlikely to get the opportunity to collect for myself.
...Keith

Attached Files



#14 Stevous

Stevous

    Mariner 2

  • -----
  • Posts: 258
  • Joined: 01 Sep 2009
  • Loc: Melbourne, Victoria, Australia

Posted 21 April 2013 - 02:15 AM

Well done to everyone the pictures all look great.

Steve

#15 Paramount

Paramount

    Ranger 4

  • -----
  • Posts: 320
  • Joined: 03 Mar 2008
  • Loc: Hereford UK

Posted 21 April 2013 - 04:04 AM

Hi Terry
Thanks for the data, here is my go at processing it, its a shame but I don't get to see this nebula from where I live
Best wishes
Gordon

Processing
Stretched in Maxim DL using gamma stretch with value of 0.1 at maximum pixel value and 16 bit, then saved to 16 bit tiff with a linear screen stretch
Equalise histograms in CS5 then colour combine to HST palette
combination of levels and curves with some minor contrast curves adjustments
Selective colour adjustments
High pass filter

Attached Files



#16 Driven1

Driven1

    Vostok 1

  • -----
  • Posts: 167
  • Joined: 19 Aug 2010
  • Loc: N. Central PA near CSSP

Posted 21 April 2013 - 07:48 AM

^ Wow Paramount! Nice! Pulled some amazing detail. Great job!

#17 Paramount

Paramount

    Ranger 4

  • -----
  • Posts: 320
  • Joined: 03 Mar 2008
  • Loc: Hereford UK

Posted 21 April 2013 - 04:21 PM

Hi
I have enclosed a link to the full size version of my effort at processing Terry's data. Thanks again to Terry for providing the data
http://www.pbase.com...796215/original
Best wishes
Gordon

#18 Paramount

Paramount

    Ranger 4

  • -----
  • Posts: 320
  • Joined: 03 Mar 2008
  • Loc: Hereford UK

Posted 21 April 2013 - 04:21 PM

Hi
I have enclosed a link to the full size version of my effort at processing Terry's data. Thanks again to Terry for providing the data
http://www.pbase.com...796215/original
Best wishes
Gordon

#19 Terry R

Terry R

    Sputnik

  • -----
  • Posts: 48
  • Joined: 31 Jan 2013
  • Loc: Australia

Posted 22 April 2013 - 04:22 AM

It’s really amazing how processing can yield such different results from the same data. I really have to thank everyone for sharing their processing steps, it’s been fantastic. I’m still struggling to get any blue out of the data. Looks like a lot of people are using PixInSight in their processing.

Terry

#20 jgw12936

jgw12936

    Apollo

  • *****
  • Posts: 1091
  • Joined: 28 Mar 2007
  • Loc: The Woodlands, TX

Posted 22 April 2013 - 07:09 AM

Thanks Terry for the data! Really fun to process.

I did mine in CCDStack and PS3


Jim

Attached Files



#21 Warhen

Warhen

    Vendor (IP4AP)

  • *****
  • Vendors
  • Posts: 1071
  • Joined: 12 Nov 2005
  • Loc: WV, USA

Posted 24 April 2013 - 02:52 PM

Very impressive processing all. I'm not attracted to most NB images (love grayscale Ha), but really interesting to see your great results w/ differing methods.

#22 bluedandelion

bluedandelion

    Surveyor 1

  • *****
  • Posts: 1860
  • Joined: 17 Aug 2007
  • Loc: Hazy Hollow, Western WA

Posted 25 April 2013 - 01:10 AM

This data is too good to pass up. So I spent several hours on this. Terry, you are most generous :bow:

Two blends in parallel follow. Will post additional processing details tomorrow.

This blend follows Juan Conjero's (lead developer of PI) suggestion for a more "natural" palette.
R: 0.5*S2 + 0.5*Ha
G: 0.15*Ha + 0.85*O3
B: O3
Posted Image
Large version on Flickr
- Ajay

#23 bluedandelion

bluedandelion

    Surveyor 1

  • *****
  • Posts: 1860
  • Joined: 17 Aug 2007
  • Loc: Hazy Hollow, Western WA

Posted 25 April 2013 - 01:14 AM

Here is the traditional Hubble Palette.
R: S2
G: Ha
B: O3
Posted Image
Large version on flickr.

- Ajay

#24 bluedandelion

bluedandelion

    Surveyor 1

  • *****
  • Posts: 1860
  • Joined: 17 Aug 2007
  • Loc: Hazy Hollow, Western WA

Posted 25 April 2013 - 01:18 AM

A combination of the above: 60 percent Juan Conjero palette + 40 percent Hubble palette.

Posted Image
Large version on Flickr.

- Ajay

#25 Mark72

Mark72

    Viking 1

  • -----
  • Posts: 670
  • Joined: 19 Sep 2012
  • Loc: Munich, Germany

Posted 25 April 2013 - 06:48 AM

The last one is very good Ajay,

Mark






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics