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

ZWO ASI071MC-Cool Color Issues

astrophotography
  • Please log in to reply
26 replies to this topic

#1 tjones84

tjones84

    Sputnik

  • -----
  • topic starter
  • Posts: 29
  • Joined: 25 Jan 2012
  • Loc: St. Catharines, Ontario

Posted 20 March 2017 - 08:44 AM

Hi all,

 

I am very new to CCD-style imaging, and I am using the new ASI071MC-Cool (Color) camera with the Sony IMX071 sensor.

 

Coming from a DSLR background, terms like Gain, Offset, ADU and debayering are new to me.

 

The issue I am having is the color balance of my stacked images coming out of DeepSkyStacker.

 

I have set the RAW/FITS Settings to the bayer pattern listed by the camera (In SGP settings):  RGGB.

 

There was a previous thread about GBRG being the correct bayer pattern to use, but in my tests that was not the case.

 

The GBRG stack came out extremely purple and could not be corrected. (Skip to end of video)

 

The RGGB bayer pattern seems to work best, but the final image color is off.  It has a greenish cast and is generally pale.  Is not possilbe to balance the levels in PS afterwards either.

 

Notes: I am shooting in (red zone) with no filters. Darks only, no flats or bias.  (Still learning how to use the camera, and do not own filters)

 

I am using SGP for image acquisition using a various exposure lengths and Gain/Offset values.  The camera was not cooled - as I do not have an external power source for the camera yet, but was used in -1 degree C weather.

 

I understand that I can improve the data immensely by using a LP filter and adding flats, but I feel like the color issue is something different.

 

If anyone has any experience with the ASI071MC-Cool (Color) I would greatly appreciate your input!  I am happy to elaborate on any details if it helps diagnose the issue.

 

Attached is the best I could process out of the data aquired. 

 



#2 tjones84

tjones84

    Sputnik

  • -----
  • topic starter
  • Posts: 29
  • Joined: 25 Jan 2012
  • Loc: St. Catharines, Ontario

Posted 20 March 2017 - 08:50 AM

m81_m82_astrobackyard.jpg



#3 lylver

lylver

    Vostok 1

  • -----
  • Posts: 103
  • Joined: 15 Feb 2017
  • Loc: near Geneva but in France

Posted 20 March 2017 - 10:59 AM

I use SharpCap, I did not achieve yet the balance I want to.

 

Only daylight seems OK. The following pictures are done with Red=52 and Blue=76

http://atom.lylver.o...iel_0001-dn.jpg

http://atom.lylver.o...od-207ap-dn.jpg

It depends on your computer monitor too.

 

I do my night balance on the moon : use histogram and overlap all colors.



#4 Jon Rista

Jon Rista

    Hubble

  • *****
  • Posts: 16211
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 20 March 2017 - 01:09 PM

Hi all,
 
I am very new to CCD-style imaging, and I am using the new ASI071MC-Cool (Color) camera with the Sony IMX071 sensor.
 
Coming from a DSLR background, terms like Gain, Offset, ADU and debayering are new to me.


Well, first...as a DSLR user, you should be familiar with all of these things, except maybe offset. Gain is basically the same as ISO on a DSLR. ISO is really just the user-controllable way of changing gain with a DSLR. You can think of Gain on the ASI071 as a very finely-tunable ISO setting. Same basic deal.

You should also be very familiar with debayering. DSLR RAW images must be debayered before they can be used, however the same care must be taken with DSLR RAW as with ASI071 fits data. You MUST calibrate the data in either case before you debayer (also called demosaicing). Some tools may handle this for you, such as DSS. With DSS, you just need to set a bayer pattern when integrating FITS data for it to debayer for you.

DSLR cameras also have an offset. It is usually a fixed offset. In the case of Canon cameras, it is usually 512 ADU and on some newer Nikons it is 600 (ADU, Analog to Digital Units, the numeric range that the camera ADC outputs. Currently the vast majority of DSLRs and mirrorless cameras are 14-bit, so the numeric range is 0 - (214-1), or 0-16383. Many older Nikons don't use a bias offset, which results in clipped data. This is not a user-configurable option on a DSLR, however it is there. With the ASI071, it's just another user-configurable options much like ISO. To start, you should stick with the default ASCOM driver settings for gain and offset, and you won't really have to worry about either.
 

The issue I am having is the color balance of my stacked images coming out of DeepSkyStacker.
 
I have set the RAW/FITS Settings to the bayer pattern listed by the camera (In SGP settings):  RGGB.
 
There was a previous thread about GBRG being the correct bayer pattern to use, but in my tests that was not the case.
 
The GBRG stack came out extremely purple and could not be corrected. (Skip to end of video)
 
The RGGB bayer pattern seems to work best, but the final image color is off.  It has a greenish cast and is generally pale.  Is not possilbe to balance the levels in PS afterwards either.


The RGGB pattern is indeed correct. I, too, have a very strong green color cast. I think that is in part due to the LP (low and high pressure sodium and mercury lighting can lead to greenish gradients) and also in part to the natural sensitivity of the green pixels in this particular sensor (the sensitivity of the green pixel to a fair amount of the NIR range is quite high and broad, which unfiltered can cause an even stronger green color cast). The color cast is most definitely correctable, however. You must properly remove the gradient (in PI, you would use DynamicBackgroundExtraction), and then you must properly neutralize the background and calibrate the white point (you can use PS with some basic layers or curves adjustments to neutralize the background, and eXcalibrator to calibrate the white point to G2V; alternatively in PI you could use BackgroundNeutralization and ColorCalibration to calibrate the color, possibly followed up by SCNR to nuke any remaining color cast).
 

Notes: I am shooting in (red zone) with no filters. Darks only, no flats or bias.  (Still learning how to use the camera, and do not own filters)


I believe this camera is sensitive to the NIR range, so no filter at all could be problematic, both from a color standpoint as well as from a star bloat standpoint. While I am not a big fan of LP filters these days (I find they are all problematic in one way or another), you will likely need to use a UV cutoff filter or an L filter of some kind to prevent issues with IR light. You might want to consider a filter like this:

https://astronomy-im...-ir-cut-filter/

To make sure you are passing only the visible spectrum. Most LP filters will also suffice as a cutoff filter. I recommend the IDAS LPS-D1 for this camera if you want to use an LP filter.

 

I am using SGP for image acquisition using a various exposure lengths and Gain/Offset values.  The camera was not cooled - as I do not have an external power source for the camera yet, but was used in -1 degree C weather.
 
I understand that I can improve the data immensely by using a LP filter and adding flats, but I feel like the color issue is something different.
 
If anyone has any experience with the ASI071MC-Cool (Color) I would greatly appreciate your input!  I am happy to elaborate on any details if it helps diagnose the issue.
 
Attached is the best I could process out of the data aquired.


Since you are in a red zone, I would start out using the preset low gain (highest dynamic range) setting in the ASCOM driver for this camera. That will set the gain and offset properly, you won't have to worry about them, and it will give you maximum dynamic range. In a red zone, even at this setting, you'll likely only be able to expose for at most a couple of minutes, and by the time you do your read noise will be more than sufficiently swamped.

I have found that this camera does need flats, regardless of the lens or scope I use with it. It always seems to have a bit of vignetting, and I'm not exactly sure why, but flats definitely clean it up. Having a flat field is important to being able to properly calibrate the color and stretch the image without running into other issues.
 


  • tjones84 and rockstarbill like this

#5 lylver

lylver

    Vostok 1

  • -----
  • Posts: 103
  • Joined: 15 Feb 2017
  • Loc: near Geneva but in France

Posted 20 March 2017 - 02:34 PM

Yep

 

a) yes put a UV/IR cut filter before doing color balance, 2" is needed, 1"1/4 cause vignetting

b) use RGGB as debayer pattern.

*** normal modern cam. indeed ***

c) go to http://zwoug.org/viewforum.php?f=21

to be kept informed, you can ask questions.

 

Unity gain for ASI071 is

Gain 90, Offset 20
- Dynamic range = 12.5 stops

 

No need to search and set offset with SharpCap : it does automatically.


  • tjones84 likes this

#6 tjones84

tjones84

    Sputnik

  • -----
  • topic starter
  • Posts: 29
  • Joined: 25 Jan 2012
  • Loc: St. Catharines, Ontario

Posted 20 March 2017 - 04:39 PM

 

Hi all,
 
I am very new to CCD-style imaging, and I am using the new ASI071MC-Cool (Color) camera with the Sony IMX071 sensor.
 
Coming from a DSLR background, terms like Gain, Offset, ADU and debayering are new to me.


Well, first...as a DSLR user, you should be familiar with all of these things, except maybe offset. Gain is basically the same as ISO on a DSLR. ISO is really just the user-controllable way of changing gain with a DSLR. You can think of Gain on the ASI071 as a very finely-tunable ISO setting. Same basic deal.

You should also be very familiar with debayering. DSLR RAW images must be debayered before they can be used, however the same care must be taken with DSLR RAW as with ASI071 fits data. You MUST calibrate the data in either case before you debayer (also called demosaicing). Some tools may handle this for you, such as DSS. With DSS, you just need to set a bayer pattern when integrating FITS data for it to debayer for you.

DSLR cameras also have an offset. It is usually a fixed offset. In the case of Canon cameras, it is usually 512 ADU and on some newer Nikons it is 600 (ADU, Analog to Digital Units, the numeric range that the camera ADC outputs. Currently the vast majority of DSLRs and mirrorless cameras are 14-bit, so the numeric range is 0 - (214-1), or 0-16383. Many older Nikons don't use a bias offset, which results in clipped data. This is not a user-configurable option on a DSLR, however it is there. With the ASI071, it's just another user-configurable options much like ISO. To start, you should stick with the default ASCOM driver settings for gain and offset, and you won't really have to worry about either.
 

The issue I am having is the color balance of my stacked images coming out of DeepSkyStacker.
 
I have set the RAW/FITS Settings to the bayer pattern listed by the camera (In SGP settings):  RGGB.
 
There was a previous thread about GBRG being the correct bayer pattern to use, but in my tests that was not the case.
 
The GBRG stack came out extremely purple and could not be corrected. (Skip to end of video)
 
The RGGB bayer pattern seems to work best, but the final image color is off.  It has a greenish cast and is generally pale.  Is not possilbe to balance the levels in PS afterwards either.


The RGGB pattern is indeed correct. I, too, have a very strong green color cast. I think that is in part due to the LP (low and high pressure sodium and mercury lighting can lead to greenish gradients) and also in part to the natural sensitivity of the green pixels in this particular sensor (the sensitivity of the green pixel to a fair amount of the NIR range is quite high and broad, which unfiltered can cause an even stronger green color cast). The color cast is most definitely correctable, however. You must properly remove the gradient (in PI, you would use DynamicBackgroundExtraction), and then you must properly neutralize the background and calibrate the white point (you can use PS with some basic layers or curves adjustments to neutralize the background, and eXcalibrator to calibrate the white point to G2V; alternatively in PI you could use BackgroundNeutralization and ColorCalibration to calibrate the color, possibly followed up by SCNR to nuke any remaining color cast).
 

Notes: I am shooting in (red zone) with no filters. Darks only, no flats or bias.  (Still learning how to use the camera, and do not own filters)


I believe this camera is sensitive to the NIR range, so no filter at all could be problematic, both from a color standpoint as well as from a star bloat standpoint. While I am not a big fan of LP filters these days (I find they are all problematic in one way or another), you will likely need to use a UV cutoff filter or an L filter of some kind to prevent issues with IR light. You might want to consider a filter like this:

https://astronomy-im...-ir-cut-filter/

To make sure you are passing only the visible spectrum. Most LP filters will also suffice as a cutoff filter. I recommend the IDAS LPS-D1 for this camera if you want to use an LP filter.

 

I am using SGP for image acquisition using a various exposure lengths and Gain/Offset values.  The camera was not cooled - as I do not have an external power source for the camera yet, but was used in -1 degree C weather.
 
I understand that I can improve the data immensely by using a LP filter and adding flats, but I feel like the color issue is something different.
 
If anyone has any experience with the ASI071MC-Cool (Color) I would greatly appreciate your input!  I am happy to elaborate on any details if it helps diagnose the issue.
 
Attached is the best I could process out of the data aquired.


Since you are in a red zone, I would start out using the preset low gain (highest dynamic range) setting in the ASCOM driver for this camera. That will set the gain and offset properly, you won't have to worry about them, and it will give you maximum dynamic range. In a red zone, even at this setting, you'll likely only be able to expose for at most a couple of minutes, and by the time you do your read noise will be more than sufficiently swamped.

I have found that this camera does need flats, regardless of the lens or scope I use with it. It always seems to have a bit of vignetting, and I'm not exactly sure why, but flats definitely clean it up. Having a flat field is important to being able to properly calibrate the color and stretch the image without running into other issues.
 

 

I can't tell you how incredibly helpful this reply was. That clears things up A LOT for me.  I appreciate the time you took to answer each question in terms I could understand! 



#7 tjones84

tjones84

    Sputnik

  • -----
  • topic starter
  • Posts: 29
  • Joined: 25 Jan 2012
  • Loc: St. Catharines, Ontario

Posted 20 March 2017 - 04:40 PM

Yep

 

a) yes put a UV/IR cut filter before doing color balance, 2" is needed, 1"1/4 cause vignetting

b) use RGGB as debayer pattern.

*** normal modern cam. indeed ***

c) go to http://zwoug.org/viewforum.php?f=21

to be kept informed, you can ask questions.

 

Unity gain for ASI071 is

Gain 90, Offset 20
- Dynamic range = 12.5 stops

 

No need to search and set offset with SharpCap : it does automatically.

Thank you for the tips!!  Especially with the 2" filter.  Makes sense considering the size of the sensor on the ASI071.  Cheers!



#8 lylver

lylver

    Vostok 1

  • -----
  • Posts: 103
  • Joined: 15 Feb 2017
  • Loc: near Geneva but in France

Posted 22 March 2017 - 04:53 PM

I confirm some color issues with my ASI071MC.

First : night imaging without IR cut causes grass and bush to become violet.

IR cut is mandatory for the color balance.

Today :

I changed a bit my color balance Red=56 and Blue=80 with Sharpcap for evening artificial light setup.

I encounter this issue :

I don't know where it comes from but when imaging a red source (a color led), it appears that there is much green added to red pixels (should not be possible, it is a monochromatic red led)

I changed the green color with fitswork4 : green = -25% red, 100% green, 0% blue to adjust.

Perhaps not a definitive solution, but the rendering of red hues are better.

On many other checks : I couldn't achieve bright/deep red color, maybe a bayer matrix filtering issue.

Can you confirm the same issue ?

 

RGB_org.jpg

more realistic rendering

RGB_-25.jpg


Edited by lylver, 22 March 2017 - 04:58 PM.


#9 Jon Rista

Jon Rista

    Hubble

  • *****
  • Posts: 16211
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 22 March 2017 - 07:10 PM

I confirm some color issues with my ASI071MC.

First : night imaging without IR cut causes grass and bush to become violet.

IR cut is mandatory for the color balance.

Today :

I changed a bit my color balance Red=56 and Blue=80 with Sharpcap for evening artificial light setup.

I encounter this issue :

I don't know where it comes from but when imaging a red source (a color led), it appears that there is much green added to red pixels (should not be possible, it is a monochromatic red led)

This is one of the inherent color purity issues with using OSC cameras. I've covered this in mathematical detail in the past, but here is a chart that gives you a visual idea of what's happening:

 

https://astronomy-im...s/ASI071-QE.jpg

 

Notice that the green plot does not go below 10% in the red spectrum (750nm to 610nm), and is as high as 50% at the green side and 30% at the NIR side. There is definitely green light passing through the red pixel filters. Similarly, note that there is 5-15% red passing through both the green and blue pixel filters. Color purity is generally poor on OSC cameras. This can to a degree be fixed mathematically, if you know the proper matrix transformation to redistribute the necessary "energy" from "contaminated" channel back into the correct channel. This is, actually, what most DSLR RAW converters do...they apply unique matrix transformations for each type of camera (which is one of the reasons why you usually have to wait for a RAW editor to be upgraded with support for new DSLR/Mirrorless cameras after they come out, because there is definitely no "one size fits all" demosaicing converter for full color cameras.) 


  • bobzeq25 and lylver like this

#10 lylver

lylver

    Vostok 1

  • -----
  • Posts: 103
  • Joined: 15 Feb 2017
  • Loc: near Geneva but in France

Posted 22 March 2017 - 11:29 PM

Ok, I was surprised by the amount of contamination and it was the first time I checked color fidelity so deeply.

 

ASI185MC has a similar QE curves but it was enough corrected for what I did with it. I did not made extensive tests about it, perhaps I should try it on my new red/blue/green lamp led for personnal understanding and quantitizing a corrected matrix.

 

The tech support for the cam could explain what to do more precisely or fix this in driver I suppose, I put a message on ZWO forum. wink.gif

Thanks for the clean historical explanation.

 

Myriam



#11 Jon Rista

Jon Rista

    Hubble

  • *****
  • Posts: 16211
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 22 March 2017 - 11:55 PM

I actually made a slight mistake before. You said that some green light was contaminating your image. Technically speaking, it was actually RED light that was contaminating your green pixels. ;P That would add the color "green" to your image, even though all the light was from a narrow band red emission. So it's really the red plot on that chart you want to look at...and yes, the red definitely contaminates both the greens (note that at the 590nm line, where the red and green channels meet, the red transmission level is about 70%! It doesn't drop to 10% until about 565nm, and it remains between 5-10% throughout the rest of the "greens". It also adds some blue light as well, although definitely not as much. 

 

So yes...your pure-red narrow band LED lamp is indeed adding signal to the green pixels, which is where the unwanted green signal is coming from. Such is the sad reality of OSC cameras. Color purity can be a real problem. This is one of the reasons I was eager to get off my DSLRs. They are no better, and some are worse than the ASI071. 



#12 lylver

lylver

    Vostok 1

  • -----
  • Posts: 103
  • Joined: 15 Feb 2017
  • Loc: near Geneva but in France

Posted 23 March 2017 - 12:25 AM

yes Jon, red to green, no problem, just an opportunity test while thinking at the subject we started.

 

For the records, I hesitated about this in December. Between the two possibilities : ASI1600 mono + filter or the ASI071, i choosed ASI071 because of pixel sized matching with my future f/D=8 scope and the nikon historical use of this sensor.

And also to do nature and birding daylight imaging.

Anyway, after balancing pro and cons, I still prefer direct imaging, it is for pleasure before maximum quality.

Imaging with a mono cam and filter wheel is more sensitive to wheather conditions. As I have an unspecialized 80mm f/D=6, exposures are a bit long for DSO.

I will manage the issue with post-processing and/or wait for support.

Perhaps later with time and better scope and accessories, I will shift to the sharper method of imaging. Lets go step by step.



#13 Jon Rista

Jon Rista

    Hubble

  • *****
  • Posts: 16211
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 23 March 2017 - 01:00 AM

lylver, if you can figure out the proper matrix transformation for the Nikon D7100, you could probably apply it to the ASI071 as well. That would redistribute the color information mathematically (you would have to do it on the final integration, not the individual subs), and give you better results. A member on these forums, Sharkmelly (spends most of his time in the DSLR forum) did some research on this subject maybe a year ago. He has a thread in the DSLR forum that covers the math, and it can be applied in PixInsight with PixelMath pretty easily. You can probably find the right information on DXOMark's 7100D review. 



#14 Pauls72

Pauls72

    Surveyor 1

  • *****
  • Posts: 1647
  • Joined: 28 Oct 2007
  • Loc: LaPorte, IN USA

Posted 23 March 2017 - 09:00 AM

Is Deep Sky Stacker:

 

You can raise or lower the adjustments for the individual channels so when stacked all 3 end up in the same place. Something like this:

DSS-Color.jpg

 

 

Or after stacking you can move the sliders to get all 3 color channels aligned.

Misalligned colors with a green cast.

DSS-Stacked.jpg

 

Color channels aligned and green cast is gone.

DSS-Aligned_Color.jpg

 

 

 

 

 


  • tjones84 and lylver like this

#15 lylver

lylver

    Vostok 1

  • -----
  • Posts: 103
  • Joined: 15 Feb 2017
  • Loc: near Geneva but in France

Posted 23 March 2017 - 09:05 AM

Found this here https://www.dxomark....---Measurements

CIE DSO (daylight spectrum)
 Color response
Sensitivity metamerism index
ISO 17321 		79
 
White balance scales
R raw 	2.5
G raw 	1
B raw 	1.76

Color matrix as defined in ISO standard 17321
     R sRGB  G sRGB  B sRGB
R raw  1.49  -0.44  -0.05
G raw  -0.07  1.3  -0.23
B raw  0.04  -0.37  1.33
CIE-A (tungsten light bulb)
Color response
Sensitivity metamerism index
ISO 17321   76

White balance scales
R raw  1.44
G raw  1
B raw  2.96

Color matrix as defined in ISO standard 17321
      R sRGB  G sRGB  B sRGB
R raw  1.61  -0.51  -0.09
G raw  -0.1  1.3  -0.21
B raw  0.06  -0.49  1.43


#16 DonBoy

DonBoy

    Viking 1

  • *****
  • Posts: 899
  • Joined: 06 Sep 2007

Posted 23 March 2017 - 11:06 AM

 

To make sure you are passing only the visible spectrum. Most LP filters will also suffice as a cutoff filter. I recommend the IDAS LPS-D1 for this camera if you want to use an LP filter.

I found this not to be the case with my ASI1600MC-Cool and the ASI071.  This D1 filter was designed for modified DSLRs as per Hutech and they suppress the Red/IR starting at 650nm where the P2 model clips at 700nm.  I feel the Idas P2 is a better choice for these two ASI cameras.

 

Please see my post:

 

https://www.cloudyni...a/#entry7774858


Edited by DonBoy, 23 March 2017 - 11:30 AM.

  • lylver likes this

#17 Jon Rista

Jon Rista

    Hubble

  • *****
  • Posts: 16211
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 23 March 2017 - 03:21 PM

 

 

To make sure you are passing only the visible spectrum. Most LP filters will also suffice as a cutoff filter. I recommend the IDAS LPS-D1 for this camera if you want to use an LP filter.

I found this not to be the case with my ASI1600MC-Cool and the ASI071.  This D1 filter was designed for modified DSLRs as per Hutech and they suppress the Red/IR starting at 650nm where the P2 model clips at 700nm.  I feel the Idas P2 is a better choice for these two ASI cameras.

 

Please see my post:

 

https://www.cloudyni...a/#entry7774858

 

The slight suppression of the ultra deep reds is the point, though. These OSC cameras can be very sensitive to the reds (which is the case with this camera, if you look at it's filter bandpass chart: https://astronomy-im.../ASI071-QE.jpg; Note how broad the read pass is.) The thing about an IDAS LPS-D1 over an IDAS LPS-P2 is that it DOES cut off a little bit of the deep reds (however, it still passes Ha, NII and SII, so you'll still get the key red emissions from emission nebula), so they don't become so overpowering of the other colors.

 

Personally, I actually liked imaging with my unmodded 5D III, as it helped me preserve the other colors in my nebula images. I often acquired uniquely colorful images of common objects, that usually end up overpoweringly red with modded DSLRs or OSC cameras.



#18 StarMike8SE

StarMike8SE

    Viking 1

  • -----
  • Posts: 844
  • Joined: 26 Mar 2014
  • Loc: Florida

Posted 24 March 2017 - 08:24 AM

It is my understanding (and that sometimes is wroing) that most OSC color cameras have the bayer of RGGB.  Remember these chips were not made for astrophotography, but for cameras used mostly in daylight, where the RGGB (double green) is needed.  Not so much for night time photography.  That is why you often get the green cast on OSC cameras.  But this should be fairly easy to get out while processing later



#19 lylver

lylver

    Vostok 1

  • -----
  • Posts: 103
  • Joined: 15 Feb 2017
  • Loc: near Geneva but in France

Posted 24 March 2017 - 09:27 AM

I was surprised last night, I took "terrestrial" pictures with a Wollensak 1/2 inch fl CS-mount lens, reducing image size to 2400x1600 (to remove vignetting).

Without adjusting the color balance (50/50) and without IR-cut, I still get good color palette range. They are always odd things when spotting vegetation (violet ...) but for the remaining color fidelity was nice.



#20 Jon Rista

Jon Rista

    Hubble

  • *****
  • Posts: 16211
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 24 March 2017 - 12:23 PM

I was surprised last night, I took "terrestrial" pictures with a Wollensak 1/2 inch fl CS-mount lens, reducing image size to 2400x1600 (to remove vignetting).

Without adjusting the color balance (50/50) and without IR-cut, I still get good color palette range. They are always odd things when spotting vegetation (violet ...) but for the remaining color fidelity was nice.

For AP, you would indeed want to use a UV/IR cutoff filter of some kind, as otherwise you'll never be able to properly focus all the light from stars, and the IR would bloat them. You need to make sure you are cutting off the unnecessary UV and IR light to get good focus when doing AP images. 



#21 lylver

lylver

    Vostok 1

  • -----
  • Posts: 103
  • Joined: 15 Feb 2017
  • Loc: near Geneva but in France

Posted 24 March 2017 - 03:28 PM

I know Jon, just trying to expose some evidence. Maybe a issue with the color balance due to the IR cut frequency @700nm ?



#22 Jon Rista

Jon Rista

    Hubble

  • *****
  • Posts: 16211
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 24 March 2017 - 03:51 PM

I know Jon, just trying to expose some evidence. Maybe a issue with the color balance due to the IR cut frequency @700nm ?

I get the feeling that you are applying daytime photography concepts to AP, which is really just confusing matters. Daytime photography doesn't really have much of anything to do with AP, the two generally play by totally different rulebooks. 



#23 smolony

smolony

    Explorer 1

  • *****
  • Posts: 87
  • Joined: 15 Dec 2016
  • Loc: Baton Rouge, La

Posted 24 March 2017 - 04:14 PM

Thanks Trevor for posting this question and John for your answer.  I have been fighting with the 071 as well.  I was pleasantly surprised to find this string today.  I tried again this week after talking to you and this is what I got out of it.  

 

I'm going to try and apply some of those changes John listed here in DSS.

Attached Thumbnails

  • BODES GALAXY & CIGAR ALL IMGS-3.jpg
  • BODES GALAXY & CIGAR ALL IMGS ALIGN DSS.jpg

Edited by smolony, 24 March 2017 - 04:15 PM.


#24 lylver

lylver

    Vostok 1

  • -----
  • Posts: 103
  • Joined: 15 Feb 2017
  • Loc: near Geneva but in France

Posted 25 March 2017 - 05:58 PM

Yes Jon,

I was speaking about daylight white balance. Trying first to understand how it works.

Next I will do my color balance on moon for AP. Perhaps it is not the good way to do but, for my eyes, it gives good colors.

I just want to go further with your information about the built in matrix.

 

I tested this "descrambling color" matrix (with fitswork4) on my previously daylight balanced pictures and I found they have better (natural) colors

It is based on overlapping surface (not accurate) from the QE curves between B (390-490nm) G (490-600nm) R(600-720nm)

corrective-matrix-asi071MC.JPG

Radjusted= 122%Rraw -19%Graw -3%Braw

Gadj= -28%R +143%G -15%B

Badj= -9%R -37%G +146%B

 

http://atom.lylver.o...rama_Annecy.jpg

http://atom.lylver.o...a_Annecy_2t.jpg

red and blue back color for far object instead of unnatural green

http://atom.lylver.o...iel_0001-dn.jpg

http://atom.lylver.o..._0001-dn_2t.jpg

Look at the house up right : blue yellow, brown, red-brown

http://atom.lylver.o...olline3dn10.jpg

http://atom.lylver.o...ine3dn10_2t.jpg

The seconds appears more vivid : the sun was shining. First picture appeared flat and grey shaded, in second one stone with brown earth, green trees, global shinny appearance etc.

 

Matrix is not mandatory, but now I understand better the color impurity due to those integrated bayer matrix.

------------------------------

A last terrestrial image with 80mm apo + Kepler 0.75x + UV-IR-cut with SharpCap, 3ms exposure

R=64, B=71, unity gain

I used the matrix in fitswork

http://atom.lylver.o...UT WB XFORM.jpg


Edited by lylver, 26 March 2017 - 11:12 AM.


#25 lylver

lylver

    Vostok 1

  • -----
  • Posts: 103
  • Joined: 15 Feb 2017
  • Loc: near Geneva but in France

Posted 08 April 2017 - 03:32 AM

For general info I got very good hue perception (according to this test) http://atom.lylver.o...or-result-0.jpg

-----------------

I went on with this work for another cam. This is intermediary result that I need to refine for ASI071MC

For ASI185MC, I use Ascom driver in Sharpcap and a set of calibrated color LED from Thorlabs. (+/-10nm @50% intensity)

I took pictures, verifying histogram (main illumination between 70% and 90%), I was badly surprised by native color rendition.

First line is native cam rendering, second line is after a corrective matrix I refined.

DSCR-balance_Ascom-ASI185MC.PNG

For check, Fraunhofer rainbow :

Raies%20de%20Fraunhofer.jpg

Lack of patience I know, but applying this preliminary result to ASI071MC too ! I was able NOW to remove backgroung "orange" (light pollution) color from the next image. I obtain darker background, more realistic saturated star color and less extended but more sharp nebula limit

This is a sort of color shift formula to obtain the right yellow (main goal), cyan (does'n agree with fraunhofer chart but visually not far for me as a sort of deep green/blue), violet/deep red hues that is so joyfull to eye.

Uncorrected : http://atom.lylver.o...-13-cropped.jpg

Corrected sky background & color shift full (big) : http://atom.lylver.o...ents/M45cor.png

Corrected preview (under, use the link up to zoom) : enjoy the central gold/blue double and surrounding, also Atlas & Pleione colors, Alcyone surrounding triplet subtle hues, Taygata ... flowerred.gifM45cor.jpg


Edited by lylver, 08 April 2017 - 03:57 AM.



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.






Also tagged with one or more of these keywords: astrophotography



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics