28 replies to this topic

[BillD17]

I took my second LRGB image last night with an ASI1600-Mono.   When I import into Star Tools with the LRGB module, its as if I don't have any color frames at all.  I took 25 x 2 min L frames and 6 each x 5 min RGB frames.  Combining all the frames gives an image that appears to be completely monochrome.  The L channel is completely overwhelming the color channels.

 

 

How can I combine my L with RGB and still retain some color?

[BillD17]

 

For comparison, if I load only the RGB channels I get this image.  It has some color, but I'm throwing away most of my photons.

[bobzeq25]

Have you tried increasing the saturation in the COLOR tool?  I've been known to use 1000%.  Or, increasing the individual color ratios in the LRGB tool?  Just because the output from the default optioned LRGB tool looks mono, doesn't mean you have no path forward.

Edited by bobzeq25, 14 March 2017 - 02:02 PM.

[scopenitout]

+1 as per Bob's advice.

It does take a lot of fiddling around, but the ratio sliders in the LRGB tool are quite effective.

[BillD17]

I must be doing something wrong in the stacking.  I reprocessed this setting all color ratios to 10 and setting the saturation to 1000% and it still looks black and white.  

 

I put the four stacked FIT files at https://www.dropbox....0K_37g0M_a?dl=0

 

If anyone can point out what I'm doing wrong, I'd appreciate it.

[Jon Rista]

Bill...is the image in your first post what you get when you combine the L with the RGB in StarTools? Just curious if that's what you are showing, because if it is, then it looks as though you have weighted your L channel to 100% and your RGB channels to 0% in the combination. I don't know exactly how StarTools works, but most LRGB combination tools work by combining L and RGB with weighting. In PI for example, LRGBCombination blends L and Chroma 50/50 initially, however you can attenuate either the L or Chroma to control exactly how the data blends. It is possible to fully attenuate the chroma, leaving you with just the L, which would probably look something like the image in your original post. 

[BillD17]

In StarTools there are three sliders for weighting the R, G, and B channels.  In my last attempt I set them all the way to the right - a 10x weighting.  My original exposures were 50min of L and 30min each of R, G, and B - each of which is about 1/3 the intensity, so 5x should have made them about even.

[Jon Rista]

Hmm. There is no blending or weighting factor for L?

[BillD17]

I don't see a blending factor for L in StarTools.  Maybe its time to move up to PI.

 

Ideally, the import should do a color space conversion where the color (Cr,Cb) is determined entirely from the RGB channels and the luminance (Y) is determined from a combination of all four channels.  Then the (Y,Cr,Cb) should be converted back to (R,G,B).  I get the impression that the luminance is getting summed into all three RGB channels completely desaturating the color.

[bobzeq25]

Somewhat curt, need to go to bed.

 

10 minute hack job in StarTools.  Not great, the point was to get a color image.

 

Mostly used defaults.  Workflow.  LRGB, defaults except turn channel combination off.  Minor crop to remove edge effects.  Wipe (big light pollution).  Develop.  Color, saturation to 1000%, background sat zero.  Big crop, Wipe needed some attention, didn't get it. Turn off tracking and denoise.  Contrast, minor = 15%, light pollution was still bad.

 

PI might have been better, would have taken way more time.  Maybe tomorrow, but I'll be sleeping in.

 

There's color in them thar hills.  <grin>

 

Edited by bobzeq25, 15 March 2017 - 02:56 AM.

[cyber]

This is a bug in the LRGB module, happened a few times to me as well.

It's easy to fix though!

 

After you load each file (L, R, G, B), let StarTools finish processing the file (the spinning circle in the bottom right should not be visible anymore) before you load the next one.

 

Here's a very quick (5 minutes) look at the data after loading the LRGB set.

 

 

LRGB (RGB ratios were default (1.00), but I disabled channel interpolation and used 2 px RGB blur)

Bin 25%

Crop edges

Wipe (color cast only)

Develop

Crop light pollution

Wipe (gradient)

Develop

Color calibration (200%, dark saturation 4.00) 

Edited by cyber, 15 March 2017 - 04:56 AM.

[ChrisWhite]

I took a look at your data this morning.  Nor'easter here in Vermont so not much going on except snowblowing. 

 

Couple of comments:  Your details look nice, and you are off to a great start.  More integration time would naturally clean this up a ton.  Your amp glow is terrible in these stacks.  I suggest you calibrate with darks to remove these.  Post processing is a PIA with these strong glows.

 

I did a RGB combine with your color channels and added the L to it with a slightly custom weighting (not far from default, but to preserve color).

 

No other editing was done except a saturation boost and color balance.  Thought you would like a PI comparison.

 

[billdan]

Hi Bill,

 

I've been using Startools for 6 months now but only with a QHY12 OSC, so thank you for the opportunity to try my hand at LRGB. 

 

Were these FITS calibrated (Darks and Flats), the reason I ask they were very noisy and I had great difficulty removing the gradients.

 

I had to use the software 2X bin feature to try and remove some of the noise, and I cropped the image to help with gradient removal.

 

Anyway this my first effort at LRGB with Startools, other more experienced users may have better results.

 

Cheers

Bill

 

Attached Thumbnails

• 

[BillD17]

This is a bug in the LRGB module, happened a few times to me as well.

It's easy to fix though!

 

After you load each file (L, R, G, B), let StarTools finish processing the file (the spinning circle in the bottom right should not be visible anymore) before you load the next one.

 

 

That explains it.  I'm pretty impatient, so I'm loading the next file before it finishes.  When I get some time I'll try again and wait for it to finish.

[BillD17]

 

Were these FITS calibrated (Darks and Flats), the reason I ask they were very noisy and I had great difficulty removing the gradients.

 

Yes, these FITS were calibrated with Darks, Flats, and Offset frames.  However, this was shot with nearly a full moon from a "red zone", and M101 was low in the sky over the most light polluted part of the sky - right over downtown San Jose.  The background noise is not "amp glow".  There is very little dark current in this sensor at -20C.  Its light pollution.

[BillD17]

Your amp glow is terrible in these stacks.  I suggest you calibrate with darks to remove these.  Post processing is a PIA with these strong glows.

 

 

That's not an amp glow.  The frames are calibrated.  Its light pollution.

[ChrisWhite]

Bill,

The glows match up perfectly with ampglow for this camera. Are you certain that your darks are a good match, and that your calibration software handled it correctly? It does not look like lp gradients.

[bigjy989]

It is not required but I usually process each channel separately (stretch/denoise) before combining in LRGB.  Using layer you can also add the weighted RGB to LUM to further maximize your photons.   Fully Process L for detail, process RGB to maximize saturation and minimize noise (excessive denoise).  Merge all the files using LRGB module (still click linear and see if HDR, life, color, and denoise further improves your results.  Each file type would use different denoise weights for brightness and or color detail loss.   With the ASI 1600 I usually use a smoothness of 25 and correlation of 2 for best results.

 

I think Ivo posted a preferred workflow a few years back on CN.   If you haven't already also check the Startools forums, several useful references there.

[George Simon]

Like billdan, I use StarTools, but have had no occasion to do LRGB processing with it since I image with a OSC camera (a DSLR). So, like him, I jumped at the chance to use your data to try out StarTools' LRGB capabilities. I did not process each color channel separately, as suggested by bigjy989, but the result is still pretty pleasing, to my eye at least.

 

 

As others have noted, there was an awful lot of noise. However, applying the Isolate algorithm in the Life module, along with a mask over the galaxy, did a nice job of eliminating the noise in the background sky, and when Tracking was turned off, the noise in the galaxy was nicely handled.

 

Thanks for sharing your data.

[bobzeq25]

 

 

Were these FITS calibrated (Darks and Flats), the reason I ask they were very noisy and I had great difficulty removing the gradients.

 

Yes, these FITS were calibrated with Darks, Flats, and Offset frames.  However, this was shot with nearly a full moon from a "red zone", and M101 was low in the sky over the most light polluted part of the sky - right over downtown San Jose.  The background noise is not "amp glow".  There is very little dark current in this sensor at -20C.  Its light pollution.

 

This explains a lot.  My problem also was dealing with gradients and noise.  I couldn't get PixInsights DynamicBackgoundElimination tool to work well at all, and WIPE was clearly struggling.  In order to cope, I wound up with the rather dim image in #10.  Brighter was just noisier with a weird background.

 

You'll never be able to take good images of dim stuff like galaxies under those conditions.  M101 is low surface brightness.

 

My backyard is a red zone, Bortle 7, mag per arc sec 19-18.5.  I don't image below 45 degrees elevation, and I don't image galaxies/nebula in moonlight.  Maybe clusters.

 

Some people even say you can't image from a red zone.  I say you can, but you need to be careful about the situation.  AP is miraculous, but not supernatural.  <smile>

Edited by bobzeq25, 15 March 2017 - 03:18 PM.

[BillD17]

Bill,

The glows match up perfectly with ampglow for this camera. Are you certain that your darks are a good match, and that your calibration software handled it correctly? It does not look like lp gradients.

Its possible I'm doing it wrong, but I shot two sets of darks, one at -20C and 2min and one at -20C and 5min and stacked each set of lights in DSS with the appropriate dark master (as well as flat master and offset master).  Here is the file list for the L channel.

 

DSS file list
CHECKED TYPE FILE
1 light C:\M101_031317\Light\L\M101_0120_L_0001.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0001-1.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0002.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0003.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0004.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0005.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0006.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0007.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0008.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0009.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0010.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0011.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0012.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0013.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0014.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0015.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0016.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0017.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0018.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0019.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0020.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0021.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0022.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0023.fit
1 reflight C:\M101_031317\Light\L\M101_0120_L_0024.fit
1 light C:\M101_031317\Light\L\M101_0120_L_0025.fit
1 dark C:\M101_031317\MasterDark_ISO0_120s.tif
1 flat C:\M101_031317\MasterFlat_ISO0.tif
1 offset C:\M101_031317\MasterOffset_ISO0.tif
#WS#Software\DeepSkyStacker\FitsDDP|BayerPattern=4
#WS#Software\DeepSkyStacker\FitsDDP|BlueScale=1.0000
#WS#Software\DeepSkyStacker\FitsDDP|Brighness=1.0000
#WS#Software\DeepSkyStacker\FitsDDP|DSLR=
#WS#Software\DeepSkyStacker\FitsDDP|FITSisRAW=0
#WS#Software\DeepSkyStacker\FitsDDP|ForceUnsigned=0
#WS#Software\DeepSkyStacker\FitsDDP|Interpolation=Bilinear
#WS#Software\DeepSkyStacker\FitsDDP|RedScale=1.0000
#WS#Software\DeepSkyStacker\RawDDP|AHD=0
#WS#Software\DeepSkyStacker\RawDDP|AutoWB=0
#WS#Software\DeepSkyStacker\RawDDP|BlackPointTo0=1
#WS#Software\DeepSkyStacker\RawDDP|BlueScale=1.0000
#WS#Software\DeepSkyStacker\RawDDP|Brighness=1.0000
#WS#Software\DeepSkyStacker\RawDDP|CameraWB=0
#WS#Software\DeepSkyStacker\RawDDP|Interpolation=Bilinear
#WS#Software\DeepSkyStacker\RawDDP|RawBayer=0
#WS#Software\DeepSkyStacker\RawDDP|RedScale=1.0000
#WS#Software\DeepSkyStacker\RawDDP|SuperPixels=0
#WS#Software\DeepSkyStacker\Register|ApplyMedianFilter=1
#WS#Software\DeepSkyStacker\Register|DetectHotPixels=1
#WS#Software\DeepSkyStacker\Register|DetectionThreshold=15
#WS#Software\DeepSkyStacker\Register|PercentStack=80
#WS#Software\DeepSkyStacker\Register|StackAfter=0
#WS#Software\DeepSkyStacker\Stacking|AlignChannels=0
#WS#Software\DeepSkyStacker\Stacking|AlignmentTransformation=0
#WS#Software\DeepSkyStacker\Stacking|ApplyFilterToCometImages=1
#WS#Software\DeepSkyStacker\Stacking|BackgroundCalibration=0
#WS#Software\DeepSkyStacker\Stacking|BackgroundCalibrationInterpolation=1
#WS#Software\DeepSkyStacker\Stacking|BadLinesDetection=0
#WS#Software\DeepSkyStacker\Stacking|CometStackingMode=0
#WS#Software\DeepSkyStacker\Stacking|CreateIntermediates=0
#WS#Software\DeepSkyStacker\Stacking|DarkFactor=1.0000
#WS#Software\DeepSkyStacker\Stacking|DarkOptimization=1
#WS#Software\DeepSkyStacker\Stacking|Dark_Iteration=5
#WS#Software\DeepSkyStacker\Stacking|Dark_Kappa=2.0000
#WS#Software\DeepSkyStacker\Stacking|Dark_Method=7
#WS#Software\DeepSkyStacker\Stacking|Debloom=0
#WS#Software\DeepSkyStacker\Stacking|Flat_Iteration=5
#WS#Software\DeepSkyStacker\Stacking|Flat_Kappa=2.0000
#WS#Software\DeepSkyStacker\Stacking|Flat_Method=7
#WS#Software\DeepSkyStacker\Stacking|HotPixelsDetection=1
#WS#Software\DeepSkyStacker\Stacking|IntermediateFileFormat=2
#WS#Software\DeepSkyStacker\Stacking|Light_Iteration=5
#WS#Software\DeepSkyStacker\Stacking|Light_Kappa=1.5000
#WS#Software\DeepSkyStacker\Stacking|Light_Method=4
#WS#Software\DeepSkyStacker\Stacking|LockCorners=1
#WS#Software\DeepSkyStacker\Stacking|Mosaic=0
#WS#Software\DeepSkyStacker\Stacking|Offset_Iteration=5
#WS#Software\DeepSkyStacker\Stacking|Offset_Kappa=2.0000
#WS#Software\DeepSkyStacker\Stacking|Offset_Method=7
#WS#Software\DeepSkyStacker\Stacking|PCS_ColdDetection=500
#WS#Software\DeepSkyStacker\Stacking|PCS_ColdFilter=1
#WS#Software\DeepSkyStacker\Stacking|PCS_DetectCleanCold=0
#WS#Software\DeepSkyStacker\Stacking|PCS_DetectCleanHot=0
#WS#Software\DeepSkyStacker\Stacking|PCS_HotDetection=500
#WS#Software\DeepSkyStacker\Stacking|PCS_HotFilter=1
#WS#Software\DeepSkyStacker\Stacking|PCS_ReplaceMethod=1
#WS#Software\DeepSkyStacker\Stacking|PCS_SaveDeltaImage=0
#WS#Software\DeepSkyStacker\Stacking|PerChannelBackgroundCalibration=1
#WS#Software\DeepSkyStacker\Stacking|PixelSizeMultiplier=1
#WS#Software\DeepSkyStacker\Stacking|RGBBackgroundCalibrationMethod=2
#WS#Software\DeepSkyStacker\Stacking|SaveCalibrated=0
#WS#Software\DeepSkyStacker\Stacking|SaveCalibratedDebayered=0
#WS#Software\DeepSkyStacker\Stacking|SaveCometImages=0
#WS#Software\DeepSkyStacker\Stacking|UseDarkFactor=0

[billdan]

Bill, I don't use DSS for stacking (I use AA5), but can you mix FIT files and TIF files in DSS?  Would be less confusing if you save your master files as FITS I think.

 

Also once you have a master dark and a master flat, you should not have to use the master offset anymore.

 

The master dark should have the offset in it and that gets subtracted from the lights. You get a double subtraction of the read noise (that's what the master offset is) if you use the master offset as well . This will drag your usable signal down into the noise floor.

 

In AA5 the only time I ever use the master offset is to calibrate my flat files to create a master flat.

 

My dark files are simply median stacked with no master offset used, to create a master dark.

Edited by billdan, 15 March 2017 - 06:58 PM.

[ChrisWhite]

I'm not too familiar with DSS, so I can't be helpful other than as an armchair quarterback.  I have been wrong before... many many times.

[KenS]

I'm pretty sure, but not certain, that DSS uses the master offset in lieu of a dark flat.

[bobzeq25]

I'm pretty sure, but not certain, that DSS uses the master offset in lieu of a dark flat.

DSS can use either offset (bias) or dark flats.  Probably will input both together, although I don't see that as a good idea.