StarTools

Hi Ivo,
Following on from our discussion on ISS about NGC6744's colours, I have had a go at NGC 253. Only 40 mins due to being fogged out. Again, a quick google shows lots of brown NGC 253's. This is the default colouring that came up with the colour module.

http://www.astrobin.com/full/109879/0/

You can get the .TIF out of DSS here: https://onedrive.live.com/redir?resid=6 ... file%2cTIF

Again, it shoudn't have any precolourbalances as I used DCRAW to convert to .tiff and then stacked them.
I did the second crop as there was a funny rectangular colour change in the background, so I just crop it out. I couldn't get WIPE to give me a clean background, I thought it looked better before the wipe.

My workflow:
File loaded [E:\Astrophotos\2014-07-25 Astrofest\Lights\20140724RC8\ngc253\ngc253.TIF].
---
--- Auto Develop
Parameter [Ignore Fine Detail <] set to [Off]
Parameter [Outside ROI Influence] set to [15 %]
--- Develop
Parameter [White Calibration] set to [Use Stars]
Parameter [Gamma] set to [1.00]
Parameter [Skyglow] set to [0 %]
Parameter [Digital Development] set to [90.29 %]
Parameter [Blue Luminance Contrib.] set to [100 %]
Parameter [Green Luminance Contrib.] set to [100 %]
Parameter [Red Luminance Contrib.] set to [100 %]
Parameter [Dark Anomaly Headroom] set to [5 %]
Parameter [Dark Anomaly Filter] set to [Off]
--- Contrast
Parameter [Expose Dark Areas] set to [No]
Parameter [Compensate Gamma] set to [No]
Parameter [Precision] set to [256 x 256 pixels]
Parameter [Dark Anomaly Filter] set to [1 pixels]
Parameter [Aggressiveness] set to [75 %]
Parameter [Dark Anomaly Headroom] set to [15 %]
--- Wavelet Sharpen
Parameter [Intelligent Enhance] set to [Yes]
Parameter [Scale 1] set to [100 %]
Parameter [Scale 2] set to [100 %]
Parameter [Scale 3] set to [100 %]
Parameter [Scale 4] set to [100 %]
Parameter [Scale 5] set to [100 %]
Parameter [Mask Fuzz] set to [8.0 pixels]
Parameter [Amount] set to [100 %]
Parameter [Small Detail Bias] set to [75 %]
--- Color
Parameter [Cap Green] set to [No]
Parameter [Bias Slider Mode] set to [Sliders Reduce Color Bias]
Parameter [Style] set to [Scientific (Color Constancy)]
Parameter [LRGB Method Emulation] set to [Straight CIELab Luminance Retention]
Parameter [Dark Saturation] set to [2.00]
Parameter [Bright Saturation] set to [Full]
Parameter [Saturation Amount] set to [200 %]
Parameter [Blue Bias Reduce] set to [1.00]
Parameter [Green Bias Reduce] set to [1.95]
Parameter [Red Bias Reduce] set to [1.67]
Parameter [Mask Fuzz] set to [1.0 pixels]
--- Life
Parameter [Detail Preservation] set to [Min Distance to 1/2 Unity]
Parameter [Compositing Algorithm] set to [Power of Inverse]
Parameter [Inherit Brightness, Color] set to [Off]
Parameter [Output Glow Only] set to [No]
Parameter [Airy Disk Sampling] set to [128 x 128 pixels]
Parameter [Airy Disk Radius] set to [8 pixels]
Parameter [Glow Threshold] set to [12 %]
Parameter [Detail Preservation Radius] set to [20.0 pixels]
Parameter [Saturation] set to [50 %]
Parameter [Strength] set to [100 %]
Parameter [Mask Fuzz] set to [1.0 pixels]
--- Contrast
Parameter [Expose Dark Areas] set to [No]
Parameter [Compensate Gamma] set to [No]
Parameter [Precision] set to [256 x 256 pixels]
Parameter [Dark Anomaly Filter] set to [1 pixels]
Parameter [Aggressiveness] set to [75 %]
Parameter [Dark Anomaly Headroom] set to [15 %]
--- Crop
Parameter [X1] set to [573 pixels]
Parameter [Y1] set to [611 pixels]
Parameter [X2] set to [4636 pixels (-566)]
Parameter [Y2] set to [2848 pixels (-617)]
--- Wavelet De-Noise
Parameter [Scale 1] set to [90 %]
Parameter [Scale 2] set to [90 %]
Parameter [Scale 3] set to [90 %]
Parameter [Scale 4] set to [90 %]
Parameter [Scale 5] set to [0 %]
Parameter [Mask Fuzz] set to [1.0 pixels]
Parameter [Scale Correlation] set to [3]
Parameter [Color Detail Loss] set to [12 %]
Parameter [Brightness Detail Loss] set to [12 %]
Parameter [Grain Size] set to [4.5 pixels]
Parameter [Read Noise Compensation] set to [2.77 %]
Parameter [Smoothness] set to [75 %]
--- Magic
Parameter [Mode] set to [Shrink]
Parameter [Mask Grow] set to [1 pixels]
Parameter [Iterations] set to [1 pixels]
---
Parameter [Quality] set to [Medium]
Parameter [New Darker Than Old] set to [No]
Parameter [Grow Mask] set to [0 pixels]
Parameter [Neighbourhood Samples] set to [0]
Parameter [New Must Be Darker Than] set to [Off]
Parameter [Neighbourhood Area] set to [200 pixels]
--- Repair
Parameter [Radial Samples] set to [32]
Parameter [Sub Sampling] set to [4x]
Parameter [Algorithm] set to [Warp]
Parameter [Grow Mask] set to [0 pixels]
File saved [E:\Astrophotos\2014-07-25 Astrofest\Lights\20140724RC8\ngc253\ngc253.tiff].

Hi Paul,

I will get back to you tomorrow on this image,
Some quick comments;
The coloring of NGC253 seems roughly consistent with what we know about galaxies (see here also on how these colours come about and what they mean), although it may be a little on the blue side. You can calibrate the image against the stars in the field for a 'second opinion'. Also make sure that Wipe did its job properly (as indicated on the IIS thread). Looking at a good randomly distributed continuum of star temperatures can tell you whether the colors in your image are plausible. Besides that, knowing a little about the history/makeup of the object can also go a long way in determining how your color calibration is going; the funny thing is that although NGC 253 is called a 'star burst galaxy', there is actually comparatively little star formation going on in the galaxy's actual outer rim and disc (it's all happening in/near the core, but that's obscured from vision). This in turn makes the disc not as blue as other galaxies where star formation is more prevalent (and hot young stars supply the bulk of the visible light). Adding to that, it contains huge amounts of dust, which tends to filter the light and makes it appear redder (or brownish) again than your typical spiral galaxy.

Another observation is that you're leaving your image resolution way too high. The image is severely oversampled (e.g. 1 unit of detail does not correspond to 1 pixel) and debayering is causing noise grain that is larger than one pixel (e.g. what was Poisson 'shot' noise as the CCD picked it up has been transformed into pattern noise by the debayering interpolation scheme, which is very hard to effectively remove). I would strongly recommend binning your image, as doing so will reduce noise, remove the pattern noise and will allow the application of deconvolution, allowing you to bring out more detail. The resolution as-is is going to waste as it is not resolving any detail at that scale.

EDIT:
Also, Wipe is a *necessity* and should not be left out. I think the trouble you've been having with Wipe stems from using data with dark anomalies in it and not countering them and/or increasing the 'Dark Anomaly Filter' parameter; you'll be left with remnants of a bias which can wreak havoc on your color balance (see also IIS response).

I couldn't help myself and had a look at the data (even though I should be doing other things )
This data too seems to be pre-color balanced

Can you confirm the RGB Channels Background Calibration and Per Channel Background Calibration are set to 'no' in DSS?
And can you confirm you use dcraw thus; Adding to the colour calibration notes above, there are three other means and things you can look at. One is diffraction spikes, which should show a good color range (though the color range will be skewed by star temperature). Another is G2V calibration (you'd need a G2V star in your field) and the last one is a special mode in StarTools' Color module called MaxRGB mode, which is activated by clicking on MaxRGB in the top right.
This mode shows which channel (red, green or blue) is dominant per pixel. Because not many things in space are predominantly green (strong OIII emissions excepted), reducing Green bias is usually a good idea if you see big contiguous areas of predominantly green pixels. I found this to be the case for the default calibration for the NGC253 data - it was way too green (cause by the fact that NGC253 is not a very good white balance calibration target). It's also a fantastic tool for those trying to color balance on a not-so-well calibrated monitor; it's an objective way to determine color balance without needing a perfect screen.

On NGC 253, I got reasonable results with using the starfield calibration method (Mask->Auto->FatStars, Keep, Sample, Mask, Clear, Invert), though the clipped colors caused a little skew (still too green). I therefore checked MaxRGB and, indeed, found a bias that was too green, causing inordinate amounts of green in the image (some green should be allowed for diffraction spikes and natural color noise). Gradually bringing up the Green Bias Reduce parameter starts to 'eat away at the large swathes of green' in MaxRGB mode and starts ceding dominance to other channels. I settled for red reduce 1.04, green reduce 1.42. Again, double check against the stars (all temperatures accounted for?) and (where possible) diffraction spikes. I think I'm even seeing hints of purple/pinkish HII areas! (EDIT: confirmed)

This is an example of what I came up with for the color balance;

Thanks Ivo. I can confirm DSS was set as you have shown, and DCRAW used as you have shown.
You have mentioned binning before and I totally forgot about that.
How can I know to use the Dark Anomaly filter in Wipe before seeing the colour results? And what do I set it to? Or is it try and error? If the colour is weird, change dark anomaly filter to 2 and redo colour, if still weird, change to 3 and redo etc?
I'm working on 6744 now with your information.

asterix2020 wrote:Thanks Ivo. I can confirm DSS was set as you have shown, and DCRAW used as you have shown.

Hmmmm... Don't mean to be a pain, but if you could share with me a single CR2 that would maybe enlighten things.

How can I know to use the Dark Anomaly filter in Wipe before seeing the colour results? And what do I set it to? Or is it try and error? If the colour is weird, change dark anomaly filter to 2 and redo colour, if still weird, change to 3 and redo etc?
I'm working on 6744 now with your information.

It is a little bit trial and error the first time, as it is very dependent on the quality of the data that you tend to process; noise and/or the lack of bias frames and/or the lack of dithering and/or the presence of dead pixels and/or anomalies in your flat frames all play a role. After a while you'll know the characteristics of the your data better.

'Untreated' dark anomalies have the effect of Wipe backing off locally - it's a subtractive process, so the only thing that happens when Wipe backs off is that it leaves in a remnant of the bias. If the remnant is circular, 9 times out of 10 you'll find a dead pixel in the middle of the circular remnant. If a remnant gets gradually worse towards an edge of the image, you have a stacking artifact (that is darker than the real background). In the case of a noisy background, all depends on the darkest pixel it can find in a certain area. A noisy background means that readings fluctuate around a mean value; some values are darker, some are brighter than the 'true' background value. Wipe will always lock on to the darkest pixels. It does this because removing more signal than the darkest pixel reads would start producing 'negative' values which would then (necessarily) have to be clipped to 0. This is where clipping starts to happen, which is effectively destruction of signal. Wipe (and StarTools in general) refuses to clip signal, unless you explicitly tell it to do so. By specifying a larger Dark Anomaly Filter kernel size, you're telling Wipe that it is okay to clip sometimes by telling it to look at the bigger picture, e.g. look at the mean of a bigger patch of pixels, instead of looking at a single pixel at a time.

Here's a single .CR2 of 6744: https://onedrive.live.com/redir?resid=6 ... hoto%2cCR2

I also checked my flats, they have dust bunnies, including the one to the right of 6744 that appears on the stretched versions of the stacked image. Not sure why the flats didn't get rid of it, as it seems to have gotten rid of the rest of them.

I'm in the middle of repro of 6744 and the default colour is coming up as in the IIS thread (slightly different numbers on the reduce bias sliders, but same idea for each colour). I did the 50% bin and wipe with dark anomaly filter to 3.

Can I use heal to hide the dust bunny? It's not that obvious as long as the image is not autodeveloped too much, I used a ROI around the in 50% of the galaxy to get the amount I wanted.

Also, the galaxy remained green when doing what you asked in the IIS thread. Screen shot attached.

asterix2020 wrote:Also, the galaxy remained green when doing what you asked in the IIS thread. Screen shot attached.

Eek! Thank you for showing me the problem!
I actually managed to replicate something similar by (accidentally) clicking randomly in the image. Does subsequently clicking 'Sample' fix the weird green output? Does the issue only happen when a popup dialog is clicked away? (for example the 'Parial mask in effect' dialog). If subsequently clicking Sample does not solve the problem and you simply cannot get rid of the green, then I cannot reproduce this odd default behavior on my machine - I'll spend tomorrow on some sleuthing...

Also thanks very much for uploading the CR2. The CR2 looks markedly different in terms of bias to the final stack (the latter has almost no bias, while the CR2 has a clear green bias that one would expect) - this is what puzzles me somewhat, but it could have something to do with the application of the flats? I'm now less worried about that though (note that a lot of this falls squarely in realm of nitpicking and perhaps even obsessiveness but I'm pretty passionate about getting the most virgin data possible). Also, for your camera you could try removing the -S 32767 option from the dcraw conversion stage, so that the stars overexpose 'properly' (per manufacturer's spec for your camera).
As for the slight star discoloration in the highlights, I can actually see the slight fringing in the CR2s as well - a distinct bluer bottom of the stars and a greener/yellower middle and top. Could an optical aberration be the cause of this perhaps? (I suppose you're more of an expert in this area than I am!)

asterix2020 wrote:Can I use heal to hide the dust bunny?

Yes, you can if it is in an area that is mostly background. You will have to do this either before or after engaging tracking, as the Heal module would otherwise throw off the noise tracking (all of a sudden pixels change radically under its nose, which it will interpret as a massive fluctuation in signal as opposed to other pixels in the image - and thus noise).

Check out this thread for a how-to. Afterwards use the Layer module, put the unhealed image into the foreground or background (Undo->Bg) and set the Blend Mode to Lighten. This layers back in any stars that may have been healed out but 'poked through' the original dust bunny.

Ivo, I tried sample with star masks, random clicking on stars, mask a random error and it all had a green bias, that when removed left the magenta. Not sure if that helps?

Also, it is possible that my RC8 is not in perfect collimation, I did notice those asymmetrical halos as well. But I wouldn't expect a colour change. Overall I'm not too fussed now I managed to get a good final image.


I managed to remove the dust bunnies with flood fill darker pixels and Heal, after noise reduction was done.