#1  
Old 10-11-2014, 10:38 PM
alistairsam's Avatar
alistairsam
Registered User

alistairsam is offline
 
Join Date: Nov 2009
Location: Box Hill North, Vic
Posts: 1,837
star colours

hi,
I'm using the KAF8300 chip with Baader 36mm LRGB filters and am finding it very difficult to get decent colour on stars, they're all mostly white.
I check individual RGB subs in maxim and they're all above 65k which is why I'm guessing they're white, but in images taken by others, the stars have very strong blue, yellow or orange colours.
I downloaded a few images taken by Mike and others and split them into RGB, and they've got very low values like 300.
The intensity is around 15k, whereas mine is an extremely high value, even with short subs.

Does this have to do with filters, exposure length of RGB, or the chip?

I get good colours in nebulae and galaxies and did a G2V measurement with CCDAP to get a ratio of 1.32:1:1.19
even if I use that when combining, I hardly get any colours with stars.
The Astrodon filters claim to be matched to the 8300 chip and the Astronomik filters claim true colour representation or 1:1:1 ratios. would this make a difference?
what am I missing?
anyone had similar experience with the 8300 chip and Baader LRGB?

Cheers
Alistair
Reply With Quote
  #2  
Old 10-11-2014, 11:32 PM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
It's a good question Alistair.

as I see it, stars are white for two reasons:
1. if they are saturated, the cores will have equal values of R,G and B (65k), so they will be white.
2. when you stretch data to bring out dim stuff, you must also compress the bright end of the data, so that the star cores all end up with similar values - eg a star may have started out with R,G,B of 20k, 30k, 40k, but after stretching you might have 62k, 63k, 64k (ie you compress the channel ratios and the result is again almost white).

I assume that images with deep coloured stars have been modified to overcome the above by taking some star data with short subs to get around saturation and/or by applying different stretching to the stars to keep a better representation of the original RGB ratios (or selectively saturating them post-stretching)

Last edited by Shiraz; 11-11-2014 at 12:46 AM.
Reply With Quote
  #3  
Old 11-11-2014, 12:56 AM
alistairsam's Avatar
alistairsam
Registered User

alistairsam is offline
 
Join Date: Nov 2009
Location: Box Hill North, Vic
Posts: 1,837
Hi Ray,

I thought as much but looking at your meathook image, what would the values of the yellow and blue stars at the bottom be?
Did you mask them at all?

Wonder if Mike can chime in, he uses the 694 like yourself but not sure if stars are masked during processing to preserve colour.
Same with Rolf's images.
Is it possible for the filter to influence this at all? Astronomiks mention equal combine ratios, but if that is the only advantage, I could apply g2v calibration values with my baader for close to equal values.
I measured the pixel values in maxim for calibrated rgb subs before processing and they were very high for 90sec sub's from dark skies. So don't really understsnd whats happening.

This image was a 10min lum and 90sec single rgb bin 1, with most stars saturated in unstretched subs.
http://www.astrobin.com/134118/B/

I'll measure them again.

As for well depth, does this influence it at all? The 694 has a smaller well depth than the 8300 yet star colors seem to be preserved. Of course it varies a lot based on processing.

Cheers
Alistair
Reply With Quote
  #4  
Old 11-11-2014, 12:28 PM
strongmanmike's Avatar
strongmanmike (Michael)
Highest Observatory in Oz

strongmanmike is offline
 
Join Date: May 2006
Location: Canberra
Posts: 17,150
Sometimes if I have bright stars in my field where some are lookin saturated in my usual 5min RGB exposures, I will take some shorter (usually 2 or 3min) RGB exposures to use just for the star colours.

Yes I add or blend in the colour with a less stretched version of the stars.

I still don't get it 100% everytime but this approach usually gets it looking acceptable

Mike
Reply With Quote
  #5  
Old 11-11-2014, 06:42 PM
Geoff45's Avatar
Geoff45 (Geoff)
PI rules

Geoff45 is offline
 
Join Date: Jul 2006
Location: Sydney
Posts: 2,631
Sometimes they may not be as white as they appear. I find that protecting the background with a mask and then cranking up the saturation will reveal colours you didn't know were there.
Geoff
Reply With Quote
  #6  
Old 11-11-2014, 07:13 PM
RickS's Avatar
RickS (Rick)
PI cult recruiter

RickS is offline
 
Join Date: Apr 2010
Location: Brisbane
Posts: 10,584
Quote:
Originally Posted by ghsmith45 View Post
Sometimes they may not be as white as they appear. I find that protecting the background with a mask and then cranking up the saturation will reveal colours you didn't know were there.
Good point, Geoff. This helps counter the compression that Shiraz mentioned earlier. A little star masked blurring of chrominance beforehand can help push a little more colour towards the saturated centres too.
Reply With Quote
  #7  
Old 12-11-2014, 09:17 AM
Paul Haese's Avatar
Paul Haese
Registered User

Paul Haese is offline
 
Join Date: Jan 2009
Location: Adelaide
Posts: 9,937
I am using a KAF8300 on one of my rigs. I do long subs (minimum is 20 minutes for RGB) for every filter. I don't appear to get saturated stars. Some of this must revolve around what you are doing in post processing, to bring out the star colour.

You can do a star shrink to help bring back some of the colour.
Reply With Quote
  #8  
Old 12-11-2014, 10:31 AM
pluto's Avatar
pluto (Hugh)
Astro Noob

pluto is offline
 
Join Date: Dec 2011
Location: Sydney
Posts: 1,982
Quote:
Originally Posted by Paul Haese View Post
I am using a KAF8300 on one of my rigs. I do long subs (minimum is 20 minutes for RGB) for every filter. I don't appear to get saturated stars. Some of this must revolve around what you are doing in post processing, to bring out the star colour.
I get saturated stars with 10 minute L subs and 5 minute bin2x2 R, G, and B subs with my setup (TSA120 + STT8300).

And I was just looking at a 15 second L bin2x2 image I had used for platesolving and it has a couple of saturated stars.

This sounds different to your experience Paul, any thoughts?
Reply With Quote
  #9  
Old 12-11-2014, 10:39 AM
SkyViking's Avatar
SkyViking (Rolf)
Registered User

SkyViking is offline
 
Join Date: Aug 2009
Location: Waitakere Ranges, New Zealand
Posts: 2,260
I like to emphasize star colour in my images, but haven't had to do anything in particular to capture such colours. I take 5min subs and naturally the brighter stars will be saturated in their cores, but all stars are recorded as Gaussian profiles so regardless of saturation they still contain colour information within a certain radius from the centre.
When I stretch my data I take care not to saturate the brightest areas, including the stars. This allows for colour to be maintained during processing.
Reply With Quote
  #10  
Old 12-11-2014, 11:58 AM
Paul Haese's Avatar
Paul Haese
Registered User

Paul Haese is offline
 
Join Date: Jan 2009
Location: Adelaide
Posts: 9,937
Quote:
Originally Posted by SkyViking View Post
I like to emphasize star colour in my images, but haven't had to do anything in particular to capture such colours. I take 5min subs and naturally the brighter stars will be saturated in their cores, but all stars are recorded as Gaussian profiles so regardless of saturation they still contain colour information within a certain radius from the centre.
When I stretch my data I take care not to saturate the brightest areas, including the stars. This allows for colour to be maintained during processing.

Yes I should have made that point too. I get core saturation but not total saturation, and that means I have colour around the outer edges of the stars. Only really bright stars are over saturated.
Reply With Quote
  #11  
Old 12-11-2014, 12:03 PM
Paul Haese's Avatar
Paul Haese
Registered User

Paul Haese is offline
 
Join Date: Jan 2009
Location: Adelaide
Posts: 9,937
Quote:
Originally Posted by pluto View Post
I get saturated stars with 10 minute L subs and 5 minute bin2x2 R, G, and B subs with my setup (TSA120 + STT8300).

And I was just looking at a 15 second L bin2x2 image I had used for platesolving and it has a couple of saturated stars.

This sounds different to your experience Paul, any thoughts?
As I noted above I get cores saturated but the outer edges are not saturated in all but exceptionally bright stars. I don't know why you would be getting different results.
Reply With Quote
  #12  
Old 12-11-2014, 12:33 PM
pluto's Avatar
pluto (Hugh)
Astro Noob

pluto is offline
 
Join Date: Dec 2011
Location: Sydney
Posts: 1,982
Quote:
Originally Posted by Paul Haese View Post
As I noted above I get cores saturated but the outer edges are not saturated in all but exceptionally bright stars. I don't know why you would be getting different results.
Ah thanks for clearing that up, that makes more sense. I have saturated cores but plenty of colour info around the cores.
Reply With Quote
  #13  
Old 15-11-2014, 10:51 PM
alistairsam's Avatar
alistairsam
Registered User

alistairsam is offline
 
Join Date: Nov 2009
Location: Box Hill North, Vic
Posts: 1,837
Thanks for the replies, looks like consensus is to control stretching during processing. so are filters unlikely to influence reproduction of star colour even if they have differing drop off's or bandwidth across different brands?

Cheers
Alistair
Reply With Quote
  #14  
Old 16-11-2014, 07:58 AM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Quote:
Originally Posted by alistairsam View Post
Thanks for the replies, looks like consensus is to control stretching during processing. so are filters unlikely to influence reproduction of star colour even if they have differing drop off's or bandwidth across different brands?

Cheers
Alistair
I think that filter band shape is not such a big deal on most objects - after all, DSLRs can produce wonderful colour images of all sorts of scenes and yet their Bayer filters generally have really poorly defined bandshapes when compared to the interference filters that we use for LRGB imaging. What software/settings do you use for stretching?

Last edited by Shiraz; 16-11-2014 at 08:27 AM.
Reply With Quote
  #15  
Old 16-11-2014, 11:09 AM
alistairsam's Avatar
alistairsam
Registered User

alistairsam is offline
 
Join Date: Nov 2009
Location: Box Hill North, Vic
Posts: 1,837
Quote:
Originally Posted by Shiraz View Post
I think that filter band shape is not such a big deal on most objects - after all, DSLRs can produce wonderful colour images of all sorts of scenes and yet their Bayer filters generally have really poorly defined bandshapes when compared to the interference filters that we use for LRGB imaging. What software/settings do you use for stretching?
hi Ray,

I use Maxim for stacking and startools for stretching and post processing.
Maxim has the option of CIELAB or RGB or HSL when combining LRGB.
I occasionally use CCDStack, results are very similar.
I haven't explored options in startools yet, might have to ask Ivo about how to maintain star colour during processing.

I've uploaded an unprocessed, uncalibrated 5min bin1x1 B sub here
https://www.dropbox.com/s/5c58rt1q5q...005_B.fit?dl=0
are you able to check if the stars have typical values.


Cheers
Alistair
Reply With Quote
  #16  
Old 16-11-2014, 02:41 PM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Quote:
Originally Posted by alistairsam View Post
hi Ray,

I use Maxim for stacking and startools for stretching and post processing.
Maxim has the option of CIELAB or RGB or HSL when combining LRGB.
I occasionally use CCDStack, results are very similar.
I haven't explored options in startools yet, might have to ask Ivo about how to maintain star colour during processing.

I've uploaded an unprocessed, uncalibrated 5min bin1x1 B sub here
https://www.dropbox.com/s/5c58rt1q5q...005_B.fit?dl=0
are you able to check if the stars have typical values.


Cheers

Name
Alistair
just a point - when stacking, make sure that you use either average or median - not sum.

Your image looks normal to me, with a few saturated stars. Ran Startools digital development over it (96%) and it behaved as expected, with compressed bright values, but no additional saturation. As Geoff has suggested, select the stars with a mask and boost the saturation and you should be able to restore some of the colour lost by compression in the stretching process. Also try Rick's idea of blurring the chrominance a bit to bring some colour into the saturated regions.

All the best, Ray

Last edited by Shiraz; 16-11-2014 at 04:11 PM.
Reply With Quote
  #17  
Old 16-11-2014, 07:27 PM
ericwbenson (Eric)
Registered User

ericwbenson is offline
 
Join Date: Sep 2009
Location: Adelaide, Australia
Posts: 209
Quote:
Originally Posted by Shiraz View Post
just a point - when stacking, make sure that you use either average or median - not sum.
In Maxim using sum has no negative effect since it keeps the image buffers in floating point (32bit). It will ask if you want to stretch the image before saving it into a 16 or 8 bit file format, FITS can handle 32 bit natively of course. However using sum/avg to stack images is usually a last resort course of action in any case!

EB
Reply With Quote
  #18  
Old 17-11-2014, 05:56 AM
irwjager's Avatar
irwjager (Ivo)
Registered User

irwjager is offline
 
Join Date: Apr 2010
Location: Melbourne
Posts: 532
Quote:
Originally Posted by alistairsam View Post
Thanks for the replies, looks like consensus is to control stretching during processing. so are filters unlikely to influence reproduction of star colour even if they have differing drop off's or bandwidth across different brands?

Cheers
Alistair
Hi Alistair,

The conclusion above would be true for all software (that I know of) *except* StarTools.

StarTools (if you let it) recovers true RGB ratios by using the Color module which, thanks to the whole Tracking thing, knows how to 'unstretch' all your data, grab the right colors and apply them after stretching everything back. This is the reason why it is highly recommended to do your final colour calibration *after* you've brought out all your detail (something that would be pretty much nonsensical in other software as colour calibration normally requires linear data).

The benefits of this are pretty massive. As it actually allows for colours that are wholly independent of who shot it, independent of exposure times, independent of your luminance processing and independent of the characteristics of your gear. This is why it's called 'Scientific' mode; it is based on the assumption that colour *ratios* out there are constant, don't change and therefore should not change in your image no matter its luminance characteristics.

Then there is the procedure of colour calibration. In StarTools there a no less than 6 ways to calibrate your image.
  1. using a white reference and simply clicking on it
  2. using a suitable galaxy as a white reference in a mask
  3. using a wide enough star field as a white reference (in a mask)
  4. using a G2V star as a white reference
  5. using the aggregate of any diffraction spikes in a mask or by eyeballing
  6. using reasoned (by knowing about the composition of an object) calibration combined with the MaxRGB mode (which even lets you calibrate using non-calibrated screens)
The latter is (afaik) a new technique where (reasoned) channel dominance in known areas are used to calibrate the image. In the case of your NGC6744 image for example, we know that the core should be dominated by yellow (older stars) whereas the outer rim should be blueish (dominated by younger more luminous stars). The disk will be peppered by purplish HII areas (predominantly caused by a mixture of red Ha and cyan-ish Hb emissions). Lastly we know that green-dominant features are very rare (with some notable OIII-rich exceptions, such as M42's core or the Tarantula nebula). If green is dominant, 9 times out of 10 you should be backing off on the green ratio. A little green sprinkled around in your image is fine as long as it fits the colour noise signature of your image.



Finally, we also know that in a wide enough field all star temperatures should be represented equally, e.g. there should equally as many red, orange, yellow, white and blue stars scattered around the image which can serve as a final sanity check for your colour balance.

Hope this helps somewhat!
Reply With Quote
  #19  
Old 17-11-2014, 07:28 AM
gregbradley's Avatar
gregbradley
Registered User

gregbradley is offline
 
Join Date: Feb 2006
Location: Sydney
Posts: 17,877
On another point here - your M42 image shows that you have your colour filters mislabelled. Take your camera off, have a torch, set it to red, do a 30 sec exposure and shine the torch on the resulting filter selection. Is it red or something else? Do the same for the rest and make sure your software has them correctly labelled. M42 is reddish not blue/magenta like that.

Greg.
Reply With Quote
  #20  
Old 17-11-2014, 03:04 PM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Quote:
Originally Posted by irwjager View Post
Hi Alistair,

The conclusion above would be true for all software (that I know of) *except* StarTools.

StarTools (if you let it) recovers true RGB ratios by using the Color module which, thanks to the whole Tracking thing, knows how to 'unstretch' all your data, grab the right colors and apply them after stretching everything back. This is the reason why it is highly recommended to do your final colour calibration *after* you've brought out all your detail (something that would be pretty much nonsensical in other software as colour calibration normally requires linear data).

The benefits of this are pretty massive. As it actually allows for colours that are wholly independent of who shot it, independent of exposure times, independent of your luminance processing and independent of the characteristics of your gear. This is why it's called 'Scientific' mode; it is based on the assumption that colour *ratios* out there are constant, don't change and therefore should not change in your image no matter its luminance characteristics.

Then there is the procedure of colour calibration. In StarTools there a no less than 6 ways to calibrate your image.
  1. using a white reference and simply clicking on it
  2. using a suitable galaxy as a white reference in a mask
  3. using a wide enough star field as a white reference (in a mask)
  4. using a G2V star as a white reference
  5. using the aggregate of any diffraction spikes in a mask or by eyeballing
  6. using reasoned (by knowing about the composition of an object) calibration combined with the MaxRGB mode (which even lets you calibrate using non-calibrated screens)
The latter is (afaik) a new technique where (reasoned) channel dominance in known areas are used to calibrate the image. In the case of your NGC6744 image for example, we know that the core should be dominated by yellow (older stars) whereas the outer rim should be blueish (dominated by younger more luminous stars). The disk will be peppered by purplish HII areas (predominantly caused by a mixture of red Ha and cyan-ish Hb emissions). Lastly we know that green-dominant features are very rare (with some notable OIII-rich exceptions, such as M42's core or the Tarantula nebula). If green is dominant, 9 times out of 10 you should be backing off on the green ratio. A little green sprinkled around in your image is fine as long as it fits the colour noise signature of your image.



Finally, we also know that in a wide enough field all star temperatures should be represented equally, e.g. there should equally as many red, orange, yellow, white and blue stars scattered around the image which can serve as a final sanity check for your colour balance.

Hope this helps somewhat!
I am a bit confused here and would be grateful for some feedback from Ivo - hope you will allow this question Alistair, but I think that it fits with your original post.

Thanks for the very clear explanation Ivo, but my basic problem in understanding it is this: if colour is processed separately from brightness, then surely dynamic range must be affected - I can't see how you can simultaneously have full dynamic range and deep saturation. For example, if a part of an image is saturated red, that means that the blue and green emitters in the screen are turned off. Ignoring what the eye does, this means that fully saturated colours can never produce more than 1/3 the brightness of white regions (where all 3 emitters are on). Once you get above 1/3 max scene brightness, you cannot display deeply saturated colours and the brightest parts of an image can really only be white if you want to use the full dynamic range of a display. Have I got this right and if so, what does it mean for colour in astro images? I think that it means that the brightest parts of an image (eg star cores) should not be coloured, but am I missing something here?

Last edited by Shiraz; 17-11-2014 at 04:12 PM.
Reply With Quote
Reply

Bookmarks

Thread Tools
Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +10. The time is now 12:44 AM.

Powered by vBulletin Version 3.8.7 | Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Advertisement
Bintel
Advertisement
Testar
Advertisement