Quote:
Originally Posted by gregbradley
A wonderful post. That last colour image is exquisite. Love it. Such wonderful colours and hues.
One way to handle varying star sizes in RGB subs is deconvolution. You decon the fatter colour to match the thinner colour stars. It may take more than one go at decon to do it. You can overdo decon but one of the advantages of the small high QE Trius is images from it will take a lot of decon without artifacts.
Greg.
|
thanks Greg. Yes, I was surprised at the colour image, but still need to work on processing. The big problem is that the Ha stars are much smaller than the colour ones, so the nebula surrounding the stars has no colour data (it got lost in the fatter stars). I tried decon and morphological processing, but got nowhere. looks like I will need to separate out the stars and the nebula for different processing and "heal" the nebula in the star regions. Rick did something similar, putting broadband stars in a narrowband image, so it can be done. Startools heals the nebula effectively, but I now need to work out how to get the narrowband stars back in there, but with broadband colour.
Quote:
Originally Posted by tilbrook@rbe.ne
Superb work Ray! 
I see what you mean by the variable seeing, but the seagull and NGC 1934 are superb. I love the colour version of the seagull good variation.
Still astonishes me what you can do with your setup!
Cheers,
Justin. |
Hi Justin. Thanks for your comments. the seeing was wind affected and very variable for these images, but last night was wind-free for the first time in yonks - and the blue channel seeing was around 5-6 arc sec at one stage
. Bring back the wind!
The scope is a continuous battle to keep everything aligned and yet not so tightly held that it distorts - you know how it is. The biggest problem is that everything is interconnected - eg recently had some atsigmatism appear - did it come from the primary, the secondary, the coma corrector, the filters or the collimation??? In the end, it turned out that the secondary glass had gone slightly pear shaped. A problem like this that would be of limited consequence at 2 arc sec sampling, but was a major issue at 1 arc sec. Works really well when it is in a good mood though.
Quote:
Originally Posted by marc4darkskies
Cool collection there Ray!
2 to 3 arc sec FWHM isn't bad at all - about average. Within that range, decon should take care of the differences you're seeing IMO. Does your processing s/w do decon?
Cheers, Marcus |
Thanks Marcus. I generally get somewhere around 2-2.5 arc sec on a good night in Ha and always use deconvolution in Pixinsight. However, the amount of data I had here was limited (eg 20 minutes one night, 25 the next etc.) and the objects are fairly faint, so I had to go very light on with deconvolution to keep the noise under some sort of control. The other effect was that wind gusts did not appear to change the central peak FWHM much, but added a low level "jitter" skirt from the occasional big excursions - the skirts blew out the bright stars when stretched (eg on N70), even though the nebulae did not suffer much. Need a stiffer mount to fix this.
regards Ray