Helix - C925

[The_bluester (Paul)]

Another bit of data gathered with the SCT I have gifted to my son. Approx 50 X 300 second subs (I can't recall exactly how many) captured with Voyager, integrated in APP and post in Photoshop using Nik tooks to bring out the contrast better.

It is interesting to note the use of Nik tools given the recent discussion of tools like Topaz Labs AI sharpen and denoise. I actually gave them a try and recognised some things in the results that I have noted in a fair few images on Astrobin and other sites that I did not like the look of. I know people get some good results with the Topaz stuff but at least with this data set I could not produce an effect I liked. The AI sharpen in particular tended to produce artifacts as soon as enough sharpening was applied to actually bring out detail. Little fuzzy galaxies suddenly turned into distinct X shapes.

Nik tools seems to bring out detail that is arguably there already (At least with moderate settings) where Topaz AI sharpening needed very gentle application or it produced "Detail" that plainly never existed to begin with. I even tried it on some traditional images with very mixed results. I am using the trial versions of all the above, I reckon I will have to buy Nik tools but I am going to leave the Topaz stuff to those who are getting better results out of it.

[gregbradley]

A good result Paul.

Sharpening of any kind requires a strong signal to noise ratio otherwise it tends to enhance the visibility of noise.

I found Topaz AI sharpening as well as their noise reduction to be very hit or miss also with very little user control.

Greg.

[Sunfish (Ray)]

Great level of detail and colour in all those background stars and other tiny objects. The clarity and focus must be spot on. There are also clouds floating in the iris which are interesting.

[The_bluester (Paul)]

I found even with a shot taken with my ASI2600 (Which has a smooth noise profile and in this case a pretty decent SNR) that the AI sharpen seems to produce "Detail" that I am pretty sure does not exist, and a very odd looking "Elephant skin" background (To pinch a term from the car industry for a dashboard surface treatment that was popular for a little while in the 90s)

Denoise AI seemed to do better, but I still struggled to make it produce something that looked better than the original input straight out of Astro Pixel Processor, with a million and one ways to make it look worse.

[Andy01 (Andy)]

That’s looking good, funnily enough OlI’ve not noticed that galaxy in the right hand ‘tear duct’ before. Nice one

[The_bluester (Paul)]

Interesting, NB data seems just to swamp it completely.

Originally Posted by Sunfish

Great level of detail and colour in all those background stars and other tiny objects. The clarity and focus must be spot on. There are also clouds floating in the iris which are interesting.

The nights I shot it were very good, and in the case of the SCT in use for it, the focus changes pretty drastically with temperature so I set Voyager up to do a focus run every half hour to make sure it stays sharp. Colour wise I am pretty happy with it, brighter stars saturate to white as I expect in the ASI294, I would love to see how the 2600 does on them as it has effectively four times the pixel well depth as the 294 at the settings I use.

[Sunfish (Ray)]

I would be interested to see the App settings you used.

Originally Posted by The_bluester

Interesting, NB data seems just to swamp it completely.






The nights I shot it were very good, and in the case of the SCT in use for it, the focus changes pretty drastically with temperature so I set Voyager up to do a focus run every half hour to make sure it stays sharp. Colour wise I am pretty happy with it, brighter stars saturate to white as I expect in the ASI294, I would love to see how the 2600 does on them as it has effectively four times the pixel well depth as the 294 at the settings I use.

[The_bluester (Paul)]

Nothing special, 48 lights, a master dark made up from 100 darks, a master flat with a good dark flat (100 frames again) and 50 flats from the flat panel.

Settings wise in APP, normal mode normalisation, and integration is by average with adaptive LN rejection filter, 2nd degree LNC and 5 iterations (I think it stopped at four) and integration type set to the normal "Interpolation" mode. Much of the detail was brought out by Nik tools using COlour effex 4 with the tonal contract filter. It is one of those things as has been discussed elsewhere, you could bring this level of detail out yourself but it would be a lot more work step by step. It still needs a judicious hand setting it up whithin Nik filters or you just end up with a noisy contrasty mess.

[Sunfish (Ray)]

Thanks Paul . Giving APP a whirl and those settings are a start.

[The_bluester (Paul)]

The default settings usually work pretty well. One that you do need to change from the default if you are using an OSC camera though is on tab "0" you need to tick the "Force bayer CFA" box or you will not get debayered data.

The default "auto" integrate setting is usually pretty good but very strong star trails can leak through a little, the "Average" (If you have at least 20 subs, otherwise median is better) integration and local normalisation rejection filter get rid of trails better.

The other tips I would make if you are assembling a colour image in APP or integrating OSC data is use the highest stretch value when selecting areas in tab 9 for the background calibration routine. It tries to make the areas you have selected greay, so if there is nebulosity in the areas you select it will try to make that greay and the colours will go well off.