Jump to content


Photo

Achieving color in galaxy shots

  • Please log in to reply
48 replies to this topic

#1 akulapanam

akulapanam

    Mariner 2

  • -----
  • Posts: 239
  • Joined: 26 Aug 2012

Posted 03 February 2013 - 12:40 PM

I was wondering what the key was to getting the bluish color in galaxy shots. The shot below is from Scott Rosen's site and is 45 second exposures.

Scott Rosen's M81

I have attached my image that I did over the course of a couple of nights. My shot was done with 40 seconds instead of 45 seconds and the total time is about 30 minutes less but as you can tell it is just lacking the vibrance of Scott's shot.

Attached Files



#2 mmalik

mmalik

    Fly Me to the Moon

  • *****
  • Posts: 5359
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 03 February 2013 - 01:06 PM

Scott over processes...just kidding :)

Achieving color saturation during processing, galaxy or otherwise, is a gradual process. Read step #38, #78, and appendix C (Block Method) in the instructions here...

Also read this..., preferably the whole thread....

Sorry, no simple answer, but you'll understand as you read through the above resources.

If you need to practice, unprocessed FITs are here.... Thx

#3 mmalik

mmalik

    Fly Me to the Moon

  • *****
  • Posts: 5359
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 03 February 2013 - 01:10 PM

You could also post your unprocessed FIT for folks to try. Thx

#4 akulapanam

akulapanam

    Mariner 2

  • -----
  • Posts: 239
  • Joined: 26 Aug 2012

Posted 03 February 2013 - 02:12 PM

Here is a link to the cropped Tiff.

http://sdrv.ms/YuT7F1

#5 RedLionNJ

RedLionNJ

    Viking 1

  • -----
  • Posts: 893
  • Joined: 29 Dec 2009
  • Loc: Red Lion, NJ, USA

Posted 03 February 2013 - 02:18 PM

And your image was also through an 8" at f5.5 with pixels the size of the 20Da? If you don't have the signal in the first place, there's going to be insufficient color to saturate.

Grant

#6 akulapanam

akulapanam

    Mariner 2

  • -----
  • Posts: 239
  • Joined: 26 Aug 2012

Posted 03 February 2013 - 02:51 PM

Mine was actually through a C9.25 w/ focal reducer with a Canon T3 but it should be pretty close.

#7 Footbag

Footbag

    Fly Me to the Moon

  • *****
  • Posts: 5841
  • Joined: 13 Apr 2009
  • Loc: Scranton, PA

Posted 03 February 2013 - 03:29 PM

Dark skies and lots of subs to start. Then you can use different saturation methods. Keep in mind that saturating also brings out noise, so it's better to start with a low noise image.

The easiest way to saturate is to duplicate the layer, set the layer to color mode and then boost the saturation. If you've gone too far, then reduce the opacity of the layer.

A few questions about your image... Did you use darks? What was your sub length?

#8 RedLionNJ

RedLionNJ

    Viking 1

  • -----
  • Posts: 893
  • Joined: 29 Dec 2009
  • Loc: Red Lion, NJ, USA

Posted 03 February 2013 - 06:16 PM

I can pull out the blue in the outer arms from your image, but there is a LOT of noise lurking in the background. With this noise present, it takes a lot of work to create an image like Scott's without blowing out the core in extreme contrast. Likewise, I could pull out just the yellow core, but at the expense of losing the outer arms in the noise.

Unfortunately, the preferred first step to a solution is likely "more data" to both cancel out more of the noise and maybe to go just a tad deeper. The more sensitive sensor in the T3 should definitely put you on a par with the 20Da, despite the 20Da's larger pixel size. The 'a' should not have much effect for something like M81. The f5.5 vs maybe f7 (you didn't say what you measured your f-ratio as) _might_ have a difference in intensity at the sensor, too.

I do think the image is on it's way - just a little more data.

Grant

#9 akulapanam

akulapanam

    Mariner 2

  • -----
  • Posts: 239
  • Joined: 26 Aug 2012

Posted 03 February 2013 - 08:50 PM

I did use darks and flats. The darks were 40 seconds each and I took 30 of them over the course of both nights. I did notice the F5.5 on Scott's image but as far as I know there is no option to achieve that with any current reducer. I think Scott is also using that reducer by the looks of his equipment page. I also use the Celestron F6.3 and that is about what I get with my current setup.

I noticed the noise issue and I have always had that problem. This image had bad vertical bands which was odd.

#10 srosenfraz

srosenfraz

    Skylab

  • *****
  • Posts: 4051
  • Joined: 06 Mar 2011
  • Loc: United States

Posted 03 February 2013 - 11:58 PM

Your setup sounds pretty similar to what I used to capture my M81 image. The f/5.5 focal reducer was an old Lumicon telecompressor. It would make that imaging train about 30% faster than your f/6.3; so, all things being equal, I'd have a better SNR than you. However, I can assure you that your T3 is substantially less noisy than my old 20Da. As such, I would say there's generally good reason to expect your SNR to be better than mine. Again - all things being equal.

That having been said, I have a feeling that our skies aren't equal. Do you know how dark your skies are? For me, I'm imaging from a blue/green zone (typically Bortle 4 - SQM 21.2 - 21.5). If your skies are significantly more light polluted, that's going to have a huge impact on SNR.

I gave a try at processing your data. I posted my .psd files (91MB) as well in case you want to look at them. This was a straight RGB processing (not a DSLR-LLRGB method of processing). My processing wasn't terribly careful - your image looks better to me (in terms of detail, etc.). I was mostly trying to bring the image to the point that I could look at the color issue.

It was rather a challenge - especially trying to pull out some blue. The first thing I had trouble with is that it was saved as a 32 bit tiff. While there's nothing wrong with that (presumably there's something good about it), I had to convert it in Photoshop to a 16 bit Tiff. When I did that, Photoshop wanted to do an HDR Toning to it. I've used HDR Toning, but only to enhance images that are well along the way in terms of stretching and detail enhancement. So, I'm not sure if I did any damage to the data while doing the 32 bit to 16 bit conversion. My gut is that I did not hurt it, but I really can't be sure.

The image had a significant gradient, as it was much brighter on the left side. Gradient Xterminator worked well to eliminate the gradient.

As you mentioned, you had some vertical banding in the image. Carboni's Vertical Banding Noise Reduction handled that problem quite well. But, after that, I was left with a circular gradient. I was able to eliminate this using an artificial flat.

At this point, I was ready to increase saturation. When I did so, I found a lot of red coming out, but no blue. This is very similar to the image you posted. To offset this, I pushed a little bit of blue into the galaxy with a slight blue curve, and then proceeded to do saturation adjustments using a variety of techniques.

I'm not terribly pleased with the result, particularly since I had to push some blue into the image. I have used color curves to enhance the color in images, but I usually don't have to push color into an image where its not there. I have my suspicions as to what may be the trouble here, but I'd be interested to hear your answers to my above questions (re:LP, modding, and CWB) before I throw out my speculations.

#11 bluedandelion

bluedandelion

    Surveyor 1

  • *****
  • Posts: 1874
  • Joined: 17 Aug 2007
  • Loc: Hazy Hollow, Western WA

Posted 04 February 2013 - 01:29 AM

The first step to getting good color is to do color correction. Saturating an uncorrected image will not help.

Color correction involves neutralizing the sky color to a less wine-colored (to a neutral color), making white stars actually look white using some type of reference and then subduing some of the extra green you may get. I use Pixinsight for processing these days. If you want PM me and I can send you some reference links on how to use PI for doing this. If you are using Photoshop perhaps others may be of help.

Once you do these corrections, saturating the image will give it that oomph you are looking for.

Ajay

#12 akulapanam

akulapanam

    Mariner 2

  • -----
  • Posts: 239
  • Joined: 26 Aug 2012

Posted 05 February 2013 - 10:52 PM

First off thanks for taking the time to reply.

You are correct that my skies are far worse than yours as I live in North Dallas. I believe my skies are in the red zone. I have a cls clip in filter but I have been holding off on using it until I get my autoguider setup so I can take shots longer than 45 seconds to a minute. My camera is unmodified. I set my white balance is set to daylight.

Sorry about the 32 bit tiff. I stopped using the 16 bit save option in dss after I realized that dss had bug that was resulting in some data being quantized to 8 bits. I use exposure and gamma to convert the 32 bit to 16 bit and I thought i had done that with the crop for the purpose of testing out startools.

#13 srosenfraz

srosenfraz

    Skylab

  • *****
  • Posts: 4051
  • Joined: 06 Mar 2011
  • Loc: United States

Posted 06 February 2013 - 12:54 AM

First off thanks for taking the time to reply.


Glad to try to help.

You are correct that my skies are far worse than yours as I live in North Dallas. I believe my skies are in the red zone.


So, that would account for the difference in SNR in our (otherwise comparable) images. Your SQM would probably be on the order of 19, whereas I'm at 21.5. So, my sky is probably 2 1/2 magnitudes darker than yours. If you review Samir's web page discussing this topic, you can see that my sky would be 2.5^2.5 = 10 times darker than yours (give or take). So, you would tend to need 10 times the exposure to achieve comparable SNR. Or conversely, we can say that your noise is Sqrt(10) = 3.2 times more than my image at a similar total exposure length.

Now, its probably not practical for you to capture 10 times as long a total exposure. Using a LP filter will help, as it'll reject some of the unwanted signal and give you a functionally darker background. That is probably the practical way to improve your SNR, along with putting in as much total exposure time as you reasonably can.

I have a cls clip in filter but I have been holding off on using it until I get my autoguider setup so I can take shots longer than 45 seconds to a minute.


Seems like there's always another upgrade to do!

My camera is unmodified. I set my white balance is set to daylight.


Perfect.


Sorry about the 32 bit tiff. I stopped using the 16 bit save option in dss after I realized that dss had bug that was resulting in some data being quantized to 8 bits. I use exposure and gamma to convert the 32 bit to 16 bit and I thought i had done that with the crop for the purpose of testing out startools.


I'm at a loss as to why your image was so lacking in blue - it definitely wasn't there. I'd be curious to retry your data from scratch. If you're interested, would you mind zipping up about 5 each of your lights, darks, flats, and bias frames and posting them? Obviously, I can't process a high quality image from this limited data, but it may help to understand what's happening to your color.

#14 pfile

pfile

    Gemini

  • -----
  • Posts: 3164
  • Joined: 14 Jun 2009

Posted 06 February 2013 - 01:43 AM

there is some really gnarly pattern noise in your tiff. how were the subs calibrated (how many darks, bias, flats) and how were the flats calibrated?

#15 akulapanam

akulapanam

    Mariner 2

  • -----
  • Posts: 239
  • Joined: 26 Aug 2012

Posted 06 February 2013 - 11:26 AM

30 darks and 15 flats. I don't use bias. I suspect part of the noise issue is the week that passed between the first imaging session and the second one.

Darks
http://sdrv.ms/14E966R

Flats
http://sdrv.ms/14E9aDG

Lights
http://sdrv.ms/14E9fHm

I have included 5 darks (a couple from each night), 5 flats, and 10 lights (5 from each night)

#16 pfile

pfile

    Gemini

  • -----
  • Posts: 3164
  • Joined: 14 Jun 2009

Posted 06 February 2013 - 12:40 PM

so i have moved on to CCD and have not been motivated to work on this with my 50D, but there is some evidence that you need a master bias made from a ridiculous number of bias subs to overcome the pattern noise in some canon cameras. if that's true of your camera then it would be better to make such a "superbias" and use it against your darks and lights, rather than relying on the bias pattern that's in the master dark. 30 is far too few for the bias issue - i'm talking about 100 or 150 bias frames.

darks and bias should be relatively stable over the course of a week, so that should not be a problem. but the quality of the light subs one week to the next could be.

#17 Mike Clemens

Mike Clemens

    Frozen to Eyepiece

  • *****
  • Posts: 7577
  • Joined: 26 Nov 2005
  • Loc: Alaska, USA

Posted 06 February 2013 - 01:09 PM

As well as careful calibration, I have had luck with really grossly dithering too. I shoot objects from four orientations (90 degrees rotation before and after the meridian) and pattern noise is very low.

#18 pfile

pfile

    Gemini

  • -----
  • Posts: 3164
  • Joined: 14 Jun 2009

Posted 06 February 2013 - 01:13 PM

good point. i've had the most problems with the banding in widefield images where typically the amount of dithering is really just a couple of pixels. i suppose that can be tweaked in PhD at least, though.

#19 srosenfraz

srosenfraz

    Skylab

  • *****
  • Posts: 4051
  • Joined: 06 Mar 2011
  • Loc: United States

Posted 06 February 2013 - 07:50 PM

30 darks and 15 flats. I don't use bias.




Its not the biggest issue you'll have, but bias frames will help to make your flats more accurate. Your flat frames (and all images captured with your camera) have bias noise in them (well, to be accurate - bias signal). By having bias frames, the bias is subtracted from your flat before your light is calibrated by dividing by the flat. A flat is a well exposed image, so the bias is not a huge component of the signal, but it does represent an inaccuracy in your calibration to not include bias. Bias frames are quick and easy to take, so I'd encourage you to take the extra minute or two that it takes to capture 15 to 30 bias frames.



I suspect part of the noise issue is the week that passed between the first imaging session and the second one.



I wouldn't think so. I regularly combine subs from nights that are a month or two apart. I've even made images that have data captured more than a year apart. Adding a group of subs from one night to a group taken another night should improve your SNR - not increase noise. You do want to be sure that you're using proper calibration frames for each set of lights - new flats if you've changed the camera to optics orientation in any way, and darks that are properly matched for the temperature on each of the nights you capture.

If you use a single set of darks for the different nights and the temperatures on each of the nights are not close, then your dark frames can increase the noise your image.




Darks
http://sdrv.ms/14E966R

Flats
http://sdrv.ms/14E9aDG

Lights
http://sdrv.ms/14E9fHm

I have included 5 darks (a couple from each night), 5 flats, and 10 lights (5 from each night)


I tried to download your data, but I think you have the permissions set so that only you can download it. I believe you need to change the permissions to public (although I'm not familiar with using Skydrive).

Perhaps you can let us know when you've corrected this?

Thanks.

#20 avarakin

avarakin

    Surveyor 1

  • *****
  • Posts: 1884
  • Joined: 13 Jul 2009
  • Loc: Parsippany NJ, USA

Posted 06 February 2013 - 11:01 PM

Were the darks taken at the same temperature, exposure and iso as the images? Just checking ;)8 Unfortunately shooting galaxies from red zone is really tough. I shoot them only if there are no better targets ( which have Ha signal in them). Modifying the camera and getting Ha filter IMO is the only way to do meaningful AP from red zone.
Alex

#21 akulapanam

akulapanam

    Mariner 2

  • -----
  • Posts: 239
  • Joined: 26 Aug 2012

Posted 07 February 2013 - 11:25 AM

Darks were taken on the same nights (15 each night) at the same temp as the light frames.

The links seem to work ok for me. Once you see the file in skydrive right click on them and click the download from the menu.

#22 pfile

pfile

    Gemini

  • -----
  • Posts: 3164
  • Joined: 14 Jun 2009

Posted 07 February 2013 - 04:08 PM

by the way, i just noticed that the subexposure length is 40 seconds. i think you're at f/6.3, so you're not going to gather too much signal. i think that's the cause for the pattern noise i was seeing. you have to stretch the data so far that the bias pattern noise comes right to the front.

i'd suggest longer subexposures, unless you think you really had cleared the read noise. i can't download your links without creating a microsoft account, so i can't really tell.

#23 avarakin

avarakin

    Surveyor 1

  • *****
  • Posts: 1884
  • Joined: 13 Jul 2009
  • Loc: Parsippany NJ, USA

Posted 07 February 2013 - 06:42 PM

I normally shoot from red zone 30" at ISO1600 and F5.6 without filters and getting into desired area of histogram, so his exposure sounds right.
Alex

#24 srosenfraz

srosenfraz

    Skylab

  • *****
  • Posts: 4051
  • Joined: 06 Mar 2011
  • Loc: United States

Posted 07 February 2013 - 06:52 PM

The links seem to work ok for me. Once you see the file in skydrive right click on them and click the download from the menu.


John -

It allows me to download the darks and the flats. But, when I try to download the Lights, it wants me to log in.

It probably works for you because you're already logged in. Try logging out (so that it sees you as a "public" person) and then try downloading. I don't think it'll work for you when you're not logged in - its not working for me.

#25 akulapanam

akulapanam

    Mariner 2

  • -----
  • Posts: 239
  • Joined: 26 Aug 2012

Posted 07 February 2013 - 09:22 PM


The links seem to work ok for me. Once you see the file in skydrive right click on them and click the download from the menu.


John -

It allows me to download the darks and the flats. But, when I try to download the Lights, it wants me to log in.

It probably works for you because you're already logged in. Try logging out (so that it sees you as a "public" person) and then try downloading. I don't think it'll work for you when you're not logged in - its not working for me.


The issue seems to be that I can't share a file that large with someone without a skydrive account. I'll try and make it two smaller ones.






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics