Author Topic: Salmon Pink Ha  (Read 5481 times)

Offline Jules

  • PixInsight Guru
  • ****
  • Posts: 513
    • View Profile
Salmon Pink Ha
« on: 2011 June 18 06:33:16 »
Ha

I have been experimenting with some data I have of IC443. I use the ha as luminance and blend with RGB in PI (also tried the same in PS CS4). It all looks very nice until the final image and the Jelly fish is a nice salmon pink, I suspect this is not the correct colour. I have tried merging the Red with the Ha as per Harry's tutorials then the red is really overpowering.

Does anybody else suffer this problem?

Julian

Offline Juan Conejero

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 6540
    • View Profile
    • http://pixinsight.com/
Re: Salmon Pink Ha
« Reply #1 on: 2011 June 18 13:10:40 »
Hi Julian

This 'salmon pink' problem is typically caused by poorly adapted luminance and chrominance components before LRGB combination. The pink color results from too high luminance values, since the excess of luminance tends to desaturate the final image.

You can match the fluxes between RGB and L with the LinearFit tool (description excerpted from this thread, which I recommend you read for more detailed information on key topics):

- Apply the initial nonlinear histogram transformations to RGB and L. Adjust the L image first, to the desired brightness and contrast. Then try to match the overall illumination of L when you transform RGB. Do it roughly by eye using the CIE L* display mode (Shift+Ctrl+L, Shift+Cmd+L on the Mac). Don't try to do a particularly accurate work here; we'll do much better in the next steps.

- Extract the CIE L* component of RGB with the ChannelExtraction tool (select the CIE L*a*b* space, uncheck a* and b*, and apply to RGB).

- Open the LinearFit tool (ColorCalibration category) and select the L image as the reference image. Apply to the L* component of RGB that you have extracted in the previous step.

- Reinsert the fitted L* in the RGB image with the ChannelCombination tool.

Now your RGB and L images have been matched very accurately. Use the LRGBCombination tool with them. You shouldn't change the luminance transfer function, neither the channel weights, as LinearFit has already done the matching job much better than anything you could do manually.

An alternative way to match RGB and L accurately is using the ScreenTransferFunction tool (STF). You apply the automatic screen stretch function to both RGB and L (clicking the 'A' button on STF), then transfer the STF instance to HistogramTransformation (HT) by dragging the blue triangle from STF to HT's control bar. Apply HT. Since STF's automatic stretch is based on image statistics, in this way both images will be stretched in a similar way.

Let me know if this improves your results.
Juan Conejero
PixInsight Development Team
http://pixinsight.com/

Offline Jules

  • PixInsight Guru
  • ****
  • Posts: 513
    • View Profile
Re: Salmon Pink Ha
« Reply #2 on: 2011 June 18 14:43:00 »
Juan

Thank you very much. I will try what you have described.

Julian

Offline twade

  • PTeam Member
  • PixInsight Old Hand
  • ****
  • Posts: 445
    • View Profile
    • http://www.northwest-landscapes.com
Re: Salmon Pink Ha
« Reply #3 on: 2011 June 19 12:47:50 »
Julian,

It also has to do with H-beta.  There's typically 15-25 percent of H-beta (blue) associated with H-alpha.  To counteract the affects of H-beta, some imagers add some of the H-alpha signal to the blue channel.  H-beta is easily diminished by intervening dust so it varies considerably across the sky.  The easiest solution is to image in H-beta too, but this dramatically increases acquisition times.

Wade

Offline Jules

  • PixInsight Guru
  • ****
  • Posts: 513
    • View Profile
Re: Salmon Pink Ha
« Reply #4 on: 2011 June 19 12:51:07 »
Wade

Thanks for the info, I will have a think about that. I have just tried Juan's ideas and about to up load them.

Julian

Offline Jules

  • PixInsight Guru
  • ****
  • Posts: 513
    • View Profile
Re: Salmon Pink Ha
« Reply #5 on: 2011 June 19 12:58:01 »
Juan

I have tried your methodology and have found the STF to HT works better, but they are still pretty grim.

I have not been able to upload an image to show. Is this a OSX thing as I use Safari?

Julian

« Last Edit: 2011 June 19 14:12:40 by Jules »

Offline Juan Conejero

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 6540
    • View Profile
    • http://pixinsight.com/
Re: Salmon Pink Ha
« Reply #6 on: 2011 June 22 00:47:10 »
Hi Julian,

I'm writing you from Safari / OS X and it lets me attach images without problems. When you click on "Additional Options", don't you have an "Attach" section where you can upload files?

Juan Conejero
PixInsight Development Team
http://pixinsight.com/

Offline Jules

  • PixInsight Guru
  • ****
  • Posts: 513
    • View Profile
Re: Salmon Pink Ha
« Reply #7 on: 2011 June 22 02:31:49 »
Hi Juan

I have additional options and upload files, the maximum size is 512kb. The images go from 80mb to 256kb. I was not sure if you could get anything meaning full from that? Anyway here they are, and thank you for your time.


Julian
« Last Edit: 2011 June 22 02:42:02 by Jules »

Offline Jules

  • PixInsight Guru
  • ****
  • Posts: 513
    • View Profile
Re: Salmon Pink Ha
« Reply #8 on: 2011 June 22 02:49:39 »
Juan

The detail in the Ha seems to lost when I combine with the RGB. Please see attached Ha image.

Julian

Offline Juan Conejero

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 6540
    • View Profile
    • http://pixinsight.com/
Re: Salmon Pink Ha
« Reply #9 on: 2011 June 22 03:25:19 »
Hi Julian,

Looks like you have a weak (and hence noisy) chrominance and you have stretched your luminance to match the chrominance. This would explain why you're getting a so dim image.

Try doing the opposite. The third image you've posted (luminance) is IMO nicely stretched. Using it as a reference, try to stretch your RGB image to match it. Then perform the LRGB combination. What do you get this way? I guess your answer will be something like 'salmon pink' :)

If you don't achieve dense colors after BackgroundNeutralization + ColorCalibration (in structure detection mode to calibrate with many stars), then I think the reason is that you don't have enough chrominance data. You probably need to accumulate more exposure in RGB. Does this make sense?
Juan Conejero
PixInsight Development Team
http://pixinsight.com/

Offline Jules

  • PixInsight Guru
  • ****
  • Posts: 513
    • View Profile
Re: Salmon Pink Ha
« Reply #10 on: 2011 June 22 04:02:56 »
Juan

It certainly does make sense. I shall try the reverse process as you mention, but I have a little bell in my head ringing about doing this earlier, either that or I am still in bed and that is the alarm!

Thanks a lot.

Julian

Offline Jules

  • PixInsight Guru
  • ****
  • Posts: 513
    • View Profile
Re: Salmon Pink Ha
« Reply #11 on: 2011 June 22 04:34:16 »
Juan

You are exactly right, it turned out to be a nice salmon pink colour using the reverse process. Please consider that this was my first image and I should have got a lot more RGB.

I have 615 minutes of Ha, and 30 minutes each of RGB, initially I thought it was my processing methodology, you have shown me it was my image collection inexperience. Thanks very much and I hope you do not consider I have wasted your time.

Julian

Offline SteveJP

  • Newcomer
  • Posts: 10
    • View Profile
Re: Salmon Pink Ha
« Reply #12 on: 2018 December 28 23:09:05 »
Hi Julian

This 'salmon pink' problem is typically caused by poorly adapted luminance and chrominance components before LRGB combination. The pink color results from too high luminance values, since the excess of luminance tends to desaturate the final image.

You can match the fluxes between RGB and L with the LinearFit tool (description excerpted from this thread, which I recommend you read for more detailed information on key topics):

- Apply the initial nonlinear histogram transformations to RGB and L. Adjust the L image first, to the desired brightness and contrast. Then try to match the overall illumination of L when you transform RGB. Do it roughly by eye using the CIE L* display mode (Shift+Ctrl+L, Shift+Cmd+L on the Mac). Don't try to do a particularly accurate work here; we'll do much better in the next steps.

- Extract the CIE L* component of RGB with the ChannelExtraction tool (select the CIE L*a*b* space, uncheck a* and b*, and apply to RGB).

- Open the LinearFit tool (ColorCalibration category) and select the L image as the reference image. Apply to the L* component of RGB that you have extracted in the previous step.

- Reinsert the fitted L* in the RGB image with the ChannelCombination tool.

Now your RGB and L images have been matched very accurately. Use the LRGBCombination tool with them. You shouldn't change the luminance transfer function, neither the channel weights, as LinearFit has already done the matching job much better than anything you could do manually.

An alternative way to match RGB and L accurately is using the ScreenTransferFunction tool (STF). You apply the automatic screen stretch function to both RGB and L (clicking the 'A' button on STF), then transfer the STF instance to HistogramTransformation (HT) by dragging the blue triangle from STF to HT's control bar. Apply HT. Since STF's automatic stretch is based on image statistics, in this way both images will be stretched in a similar way.

Let me know if this improves your results.

Juan,
I'm getting the Salmon Pink problem too, but I'm doing the LRGBCombination on fully processed RGB and Ha images. Should I be doing the RGB & Ha combination in the linear stage? Am I right in assuming the method you give here assumes linear images?
Thanks
Steve

Offline SteveJP

  • Newcomer
  • Posts: 10
    • View Profile
Re: Salmon Pink Ha
« Reply #13 on: 2018 December 28 23:49:06 »
Why is "salmon pink" a problem?  What colour do you expect to achieve from a cloud of gas containing Hydrogen and some Oxygen and therefore emissions at the H-alpha, H-beta, H-gamma and OIII wavelengths?

Take a look at Roger Clark's discussion:
http://clarkvision.com/articles/color.of.nebulae.and.interstellar.dust/

Mark
Hi Mark,
That's and interesting article, and I'll certainly bear it in mind, however, I think the Salmon Pink "problem" is a recognised issue, and probably stems from my defective processing rather than representing the real colour. I want to try Juan's method, but I'm uncertain if it applies to linear or non-linear images. I'm processing NGC2070 and there is no lack of images available to use as a reference. Also, when I combine my Ha and RGB using NBRGBCombination, I do not get the salmon pink appearance.
Thanks
Steve