StarTools

A rare and unusual clear stable night during our summer ( Southern Hemisphere) presented itself, so took the opportunity to image. The moon was a decending slither in the west and gone by 10pm. Humidity was up around 89% and no wind but I’m baffled why there was little or no dew around. Another unusual phenomenon in summer , no mosquitoes or bugs ?? Anyway enjoyed the night immensely

NexDome Obs
Bortle 3 Sky
10” f5 Klaus Helmerich Carbon fibre Newt
Skywatcher EQ8-R pro mount
ZWO 2600MC cooled to -10C , Gain 101 ( I always add 1 digit just in case the ZWO HCG setting calibration is out on the driver , some folk add 5 )
TS Optics GPU coma corrector
Antlia ALPT 5nm Dualband OSC filter
EQMOD and Stellarium
APT
PHD2 Multistar guiding ( O.45 to 0.55 total rms ) smooth tight guiding ( image scale is 0.62 arc sec per pixel )
60 x 3 min dithered subs
40 x Flats
60 x Bias
Stacked in DSS
Processed in Startools v 1.8
I only processed in Synthetic Luminance so far , chasing more data to finalise the image with Color.
Used HDR tame to tone down the keyhole region which is always the brightest part.
Used Spatially Variant PSF Deconvolution again with excellent results ( this time as the data set was virtually without noise I did not use Binning but left it at native size ( huge image file ) Definitely SV Decon works better with low noise unbinned data. I did a re process with 50% Binning and the smaller stars ended up a bit blocky and some with artefacts in the cores when zoomed in.
The stars with no Binning are pin point , bright and round points of light , brilliant !!!
Iterations setting 15
Ringing 80% and 0.90
Sampling at least 25 stars with green cores ( plenty of great stars to choose from due to quality low noise unbinned data )

Original framed version
Crop version showing Keyhole Region

Comments most welcome

Thanks
Martin

Wow nice stuff, Martin.

You're becoming an SVD master. Such crisp detail. Will be great to see this with color applied, though, I wonder how it will be in bicolor from the duo.

My only concern here was the one bright star, which seems to be awful wobbly. But, afte finding some references here and there using ESO images, CN, even Stellarium with DSS on and star augmentation off, there does seem to be something going on here. Sometimes more or less showing, but to the extent it does I think your wobble directionally matches the references. So, multiple stars that can't be split?

I do wonder if perhaps it is being overenhanced here though, perhaps artifacted, which makes it appear stronger than it ought to? I think there's additional room for deringing, which if I pixel peep I can see on some other stars against nebula background as well. Not many, but a few.

There were definitely cases in 1.8 where I would need to bump detect to about 85, and dering to say 0.92. Though that's on my far less crisp and far more binned data.

Due to getting tripped up some with ringing in the latest 1.9a, I have been thinking about going more full scale processing also. Takes longer, too long really, and I doubt my data is ever clean enough to not want that extra SNR from binning. Maybe I'll try some 50 and 71 bins instead of my typical 35.

Thanks Mike,
The central large Star is Eta Carinae which suffered a massive explosion 150 years ago. This created two gigantic lobes either side of the core ( refer Hubble image ) This might be the reason for the unusual shape of the stars core ?

I don’t know about an SV Decon Master , but I am getting my head around it after many months

Yes in hindsight I should have used 100% for de ringing detect and 0.95 for de ringing as my iterations were bumped up to 15 ( highest I’ve ever used )

SV Decon performance and results are definitely tied to Binning and image quality ( SNR )

This image was not Binned due to the high quality data ( image size was 6100 x 4100 )
Noise reduction was only 2 pixels ( I could have skipped noise reduction as there was little difference between sing it and not )

For an experiment I binned at 50% and SV Decon did not give good results , blocky stars , cores with a dark pixel and so on ….

So SV Decon has its limitations like everything

I find at my image scales , Scalable Binning from 65% and up to no Binning works best and of course this is dependent upon the quality of data ( SNR )

I used to Bin at 50% or default in 1.7 and previous, now with SV Decon in 1.8 , I gauge my Binning on the quality of data through experience ( anywhere from 65% to No Binning )

I still have an issue with HDR 1.8 ( only use Tame which works well ) Reveal and Optimise are too Aggressive or powerful and my nebulosity regions in images end up looking like shattered safety glass. Reducing the value still doesn’t do a satisfactory job. I much preferred HDR in 1.7 which worked beautifully.

If I can’t collect additional quality data on Carina soon , I will process the Color ( Bicolor HOO or OHO ) and move on to other targets

Thanks again
Martin