Ngc 2997

[graham.hobart]

Hi All, this is the fruit of three nights between clouds. 4hrs 40mins of five minute subs. Fully calibrated with darks, flats, dark flats and bias frames and I still can't get rid of the gradient.
Any tips greatly appreciated- have tried WIPE in Star Tools already!
Orion Eon refractor f6 110mm
QHY 8 OSC (old style)
Cheers
Graz.

[Shiraz]

Hi Graham. looks like the gradient is messing up an otherwise useful image. What to do depends on what software you have. I'm no expert, but suggest you could try:

as a first step, maybe just wind down the bottom end of the histogram using curves to make the background darker - could help make the gradients less objectionable.

Since you have tried all the calibration techniques as well as WIPE in StarTools, you might (in desperation) try a mask that protects the stars and galaxies and then reduce the colour saturation in the intervening sky, possibly followed by darkening of the sky. The only way I could find to generate the required mask in StarTools was using an automatic star mask plus "flood fill lighter pixels" to highlight the galaxies (or you could do it manually) - then invert the mask to protect the stars and galaxies and wind back the sky saturation using the colour module. Tried this on an image and it worked OK.

In Pixinsight, the dynamic background extraction tool should tidy up the problems automatically. There is also a very capable gradient removal tool in IRIS, but working out how to use it is somewhat challenging.

Any idea what causes the problem? - its always easier to tidy this sort of thing up at the source if you possibly can

[graham.hobart]

Thanks Ray good advice.
I suspect because I was shooting this galaxy above my neighbours house and also that my Wife had a light on in the house has contributed to the gradient, plus shooting over in the NE is right above the Hobart city light dome. When I autodev in Star tools it comes up a horrid greeny yellow colour so I assume that is background LP. I used to use a CLS filter in my DSLR so am thinking about a filter?
In the mean time I shot another 3 hours last night of the same target in really good seeing so I will add these and try some of your recommendations.
I tried DBE in PI, but the area with the worst problem comes up as a nil weight so placing a pixel marker there goes red and won't register it (manual or moving an auto generated marker)
Cheers
Graham

[Shiraz]

hi Graham. FWIW, I found that increasing the tolerance, increasing the shadows relaxation and decreasing the minimum sample weight all seem to help/enable DBE to generate/accept samples in difficult images (might need big changes, eg I used 0.95, 8 and 0.1 on a recent problem image). The radius parameter also has an effect. I don't know enough to be sure that it would help, but it might be worth changing a few values if you have not already done so. regards Ray

[irwjager]

Hi Graham, if you could send over the data, I could have a look to see what Wipe can do. There are also some other tricks that you can use to ameliorate an uneven background and still keep the object of interest.

Cheers,

[graham.hobart]

Hi!
I have more date that I will add tonight (I hope) but if I could send you the 16bit TIF that comes out of DSS that would be good.
Cheers
Graham.

[irwjager]

As I already PM'ed, you have some small problems in your data that cause Wipe to lose track of the 'real' background (so called 'dark anomalies' - read the help in Wipe for more info).
These problems (some dead pixels and faulty columns) cause pixels in your image that dip below the real interstellar background. This confuses Wipe and it will assume that there is a sharp gradient that needs to be removed.

They are easily dealt with by either masking them out (so that Wipe doesn't see them to begin with), or, if they're small, by upping the Dark Anomaly Filter. It turns out the latter was enough for this image.

I processed it in 1.3.5.279;

--- Auto Develop
To see what we got.
We got a green/yellow cast, some stacking artifacts and fairly bloated stars.
Oversampling isn't too bad and data quality seems reasonable.
--- Crop
To better frame the galaxy (and get rid of stacking artifacts).
Parameter [X1] set to [838 pixels]
Parameter [Y1] set to [478 pixels]
Parameter [X2] set to [2148 pixels (-890)]
Parameter [Y2] set to [1477 pixels (-537)]
--- Wipe
Default values, except for Parameter [Dark Anomaly Filter] set to [7 pixels] (this fixes the problem you had, as indicated earlier)
I also set Parameter [Temporary AutoDev] to [Yes], so I can better see the result that Wipe comes up with.
--- Auto Develop
Final global stretch, with Region of Interest over the galaxy.
Parameter [Ignore Fine Detail <] set to [2.8 pixels], in order to not optimize for noise grain.
Parameter [Outside ROI Influence] set to [8 %], in order to optimize even less for anything outside the ROI - we *really* don't care about anything else :P
--- Deconvolution
Let Decon generate a mask automatically.
Parameter [Radius] set to [2.1 pixels]
--- HDR
Reveal preset.
--- Wavelet Sharpen
Use same mask that Decon created.
Parameter [Small Detail Bias] set to [96 %]
--- Life
Clear mask, and invert (e.g. 'reset' the mask to everything 'on').
Isolate preset. This pushes back the busy star field a little (as well as noise), while retaining the galaxy.
--- Color
Final color calibration. We can see that alignment by DSS was not great (it often is quite bad unfortunately) with some color misalignment in some of the stars.
Parameter [Cap Green] set to [To Yellow] (to suppress any green, which is very rarely dominant in an image)
Parameter [Dark Saturation] set to [3.20] (bring color into the darker portions of the image).
--- Wavelet De-Noise
Final noise reduction (switching Tracking off)
Parameter [Read Noise Compensation] set to [26.03 %]

And that's it!

You could use the Magic module to reduce the star sizes a little and use the Heal module to fix the bad pixel column, but I'll leave that as an excercise to the reader. :)

(I had to crop the image a little further to fit < 200kb for posting)

Hope this helps!