Astra Image help

[Rigel003]

Hi
I'm just trying out Astra Image for the 1st time (thanks for posting the link, David) and having problems. I've split the image of Jupiter into the 3 colour planes and run a deconvolution on each. When I do the RGB combine I end up with really odd colours. It looks almost like the tones are reversed -the brown bands are blue and the blue zones are pink. Can someone help here please?

I'm having a "bad software" week - problems occuring in Registax that I've not had before but I might leave that to a separate thread. One thing at a time.

Graeme

[iceman]

You've gotta make sure that when you recombine, you're picking the right images for the red, green and blue channels.

Until you get the hang of it, when you do a split, it's named red/green/blue, but after you do a deconvolution the names are gone. So, after each deconvolution, go Edit->Image Title, and call it something like "red-deconvolution" etc, so when you RGB combine you can find the right image for the right channel.

[Robert_T]

what Mike says:thumbsup: , do one channel at a time and make sure you rename with the colour before moving on - there are also colour weights defaulting to 1.0 for each colour channel - you may want to check they are set to 1.

cheers,

[[1ponders]]

It also sounds like your images after the RGB combine aren't lining up. If that's the case then using the xy rgb coordinate box at the bottom right of the RGB combine dialogue box will allow you to shift each channel in xy directions to line them up. I'd suggest leaving the green channel alone and shifting the Red and Blue channels

[Rigel003]

Thanks, Mike, Rob & Paul.
I had the R and B channels reversed. As you said, you can't tell which is which in the dialogue box. I tried Lucy Richardson with 8 iterations and a radius of 1.1. The result is sharper but also much grainier - any way around this? The blurring function seems pretty severe, even on the "little" setting. Do people use this or transfer the file into Photoshop or something else for noise reduction? Can you tell me the difference between the 3 deconvolution methods?
So many questions!
Graeme

[asimov]

Lol. I have my mate Lester here right now with me for the sole purpose of teaching him how to use Astraimage.

I never use the blurring or the sharpening options in Astraimage, miles too severe.

Try 5-7 iterations at say, 0.4-0.6 radius & see how you get on.

[Rigel003]

Thanks, Asi. Do you do house calls to Adelaide?
Graeme

[Robert_T]

Hey Asi - looks like you might have a calling as an AstraImage Processing Consultant... maybe if you gave advice on drop curtains and cork flooring at the same time:P

[davidpretorius]

Janoskiss knows the theory and the rest of us have fiddled enough to know generally what will work for our scopes. So by no means is this gospel:

I have found that ME 3 or 4x1.1 exponential deconvolution works best for me. I am imaging with a 10" reflector at f38.5 or 1684x.

When i was imaging at less magnification, then LR seemed to do the trick.

Bird and Asimov have been experimenting with VC and Asimov has adding FFT into the mix as well.

There is no idiot proof recipe here, the best thing to do is experiment.

Many processing / image capture settings have been turned on their head in the last 12 months and there is so much more to try.

You will be happy to know that what we are discussing / nutting out here on IIS is more indepth and leading edge than other forums in the US and UK. I get the sense we here in australia are much more keen to experiment and discuss amongst each other and it shows! There are no secret processing techniques, we share and thus we are moving forward quicker than anywhere else in the world in my opinion.

This results in Mike's jupiters popping up in all the BIG websites.

[Rigel003]

Thanks for the detailed info, David. Lots of stuff to try out now.
Graeme

[davidpretorius]

my pleasure.

http://www.precons.com/iceinspace/gallery/jupiter_060426_1147/

grab some of the .tiff files and then split in astraimage and then have a play with what you think is best ie with deconvolution and unsharp mask and fft etc

there are various stack sizes and wavelet settings