Jupiter - First Attempt

[gnomus]

I tried Jupiter a couple of nights ago (6 Jan 2015).  This is from my ZWO ASI120MM, through an 8 inch Celestron Edge.  I took 4000 frames through R, G and B filters and stacked the best 2000 in AS2.  They were subject to some Lucy-Richardson Deconvolution in Astra Image 3.0.  The images were then aligned using a Demo version of Maxim DL.  I then did a Linear Fit in a trial version of Pixinsight before combining the three channels in Pixinsight.  

I would say that the seeing on the night in question was not great.  Furthermore, these were taken around 10pm when Jupiter was not all that high in the sky.  Furthermore, there were occasional wisps of cloud in the vicinity.

Overall, I am pleased to have got something that looks vaguely like Jupiter.  I was a little disappointed, however, by the lack of any fine detail.  I did try a bit more work using wavelets in Registax.  This seemed to result in some increase in contrast, a lot of increase in noise, but no real increase in detail.  I am trying not to over-process my images.

Can I capture finer detail using a C8?  (I have a 2.5 barlow, but I am awaiting n extension tube before I will be able to try that.)

Secondly, I am not going to be able to afford both PixInsight and Maxim DL.  I am a little confused by the various options available for purchasing Maxim DL and don't know what I would need.  There are also certain things that I really like about PI.  If I can only buy one, which one should I get?  If PI, what should i use in place of Maxim to do the alignment (I am correct in thinking that PI does not do planetary alignment, right)?

This is my final image at 100% size - be gentle (oh ... whatever ... be as rough as you like...)

[cuivenion]

Looks pretty good to me, I imagine you will get more detail when you start using your barlow. Much better than my first attempt, not that I'm an expert or anything.

[Freddie]

That's a good first pic. For planetary, the majority use AS2, and Regi6 then Image Analyser and or Photoshop/GIMP so Pix and Max are not really needed/a must for planetary imaging.

[Astrosurf]

That's better than my first attempt too! I've always had problems with planets and I've been told the seeing is crucial. One night you could get a stonking image and the next it would be blurry and awful. Just keep at it!

I'm afraid I'm not familiar with PixInsight and Maxim DL. I'm sure someone will be along soon.

Alexxx

[michael.h.f.wilkinson]

A Barlow will get you more detail, but that is a very good first effort.

[leelee970]

Excellent first attempt at Jupiter. Well done.

Lee

[MattJenko]

I am itching to get a crack at Jupiter with my new ASI (the non-S version), and this has done nothing calm me down - quite the opposite. Excellent effort.

[MattJenko]

I bought PI as a complete image processing app, whereas Maxim does a lot more to do with the capture side of things. You can do all the aligning/registering work for planetary with AutoStackert and Registax from my understanding, but seeing as I haven't done a single planetary image, I am theorist only! JamesF did a great primer on it here : http://stargazerslounge.com/topic/184821-beginners-guide-to-stacking-planetary-images-with-autostakkert2/page-1?hl=+planet +image +jamesf

[glowingturnip]

better than any planet image i've got to date, well done.

if it's just for processing, and you want one or the other, Pixinsight beats Maxim pants down.  I've got both and never use Maxim nowadays, mistaken purchase if I'm honest.  Some use it for capturing though.

You can align RGB channels with Registax - load the stacked RGB file into registax, and on the wavelets page there's a tool called RGB align, use that, draw a box round the planet and click the button.  Remember to 'apply to all' before saving.  Works for lunar detail too

try van cittert deconvolution in pixinsight, and/or multiscale transform for sharpening

[gnomus]

better than any planet image i've got to date, well done.

if it's just for processing, and you want one or the other, Pixinsight beats Maxim pants down.  I've got both and never use Maxim nowadays, mistaken purchase if I'm honest.  Some use it for capturing though.

You can align RGB channels with Registax - load the stacked RGB file into registax, and on the wavelets page there's a tool called RGB align, use that, draw a box round the planet and click the button.  Remember to 'apply to all' before saving.  Works for lunar detail too

try van cittert deconvolution in pixinsight, and/or multiscale transform for sharpening

Thanks everyone for the positive feedback.  Thanks too for these suggestions, Stuart.  I am leaning towards PI.  I don't seem able to get very far with Van Cittert - are there any tutorials to give me some idea what to do with respect to these deconvolution processes?  There are so many parameters that can be adjusted it all seems a bit hit or miss.

I did a little more work using Lucy Richardson in PI.  I got this, which I think is a little "sharper".  But the differences are marginal.  I am not sure if this version is starting to get that 'over-processed' look.

[Vox45]

Very nice ! Good job on your 1st attempt

[glowingturnip]

there was a tutorial on using it for lunar, but it got 404'd - not sure if this link works for you, it doesn't get past the naughty-filter here - http://web.archive.org/web/20120621031059/http://pixinsight.com/examples/deconvolution/moon/en.html

basically, go to the gaussian psf tab, set the algorithm to van cittert, the iterations to something neutral, say 15, wavelet regularisation turned on and left at defaults, deringing off for now, and then it's trial and error on the shape and the stddev - I set up lots of previews and try them with different shape numbers first to see which one I like, and then using that shape number try again with different stdevs to see which one I like then.  Once you have found your favourites, turn on derringing and set the global dark and global light numbers to very low levels, like 0.01 and then try increasing the iterations - you will hopefully find you can get away with more iterations if you increase the derringing numbers slightly, but be careful not to oversharpen or introduce artifacts - common artifacts are light stringy-looking things, you'll know what I mean when you see them.

Artifacts of over-sharpening on planets are halo effects either just inside or just outside the limb of the planet, which I'm afraid you have on the right hand side of your second picture.

for DSO's, use the Richardson Lucy algorithm, and you don't have to do the shape+stddev guesswork, you use the Dynamic PSF tool, click on a bunch of stars in the image (about 40), click on the 'camera' icon on the Dynamic PSF and it gives you a little psf file, then load that in the External PSF tab, and do the rest of it as above.

[gnomus]

there was a tutorial on using it for lunar, but it got 404'd - not sure if this link works for you, it doesn't get past the naughty-filter here - http://web.archive.org/web/20120621031059/http://pixinsight.com/examples/deconvolution/moon/en.html

basically, go to the gaussian psf tab, set the algorithm to van cittert, the iterations to something neutral, say 15, wavelet regularisation turned on and left at defaults, deringing off for now, and then it's trial and error on the shape and the stddev - I set up lots of previews and try them with different shape numbers first to see which one I like, and then using that shape number try again with different stdevs to see which one I like then.  Once you have found your favourites, turn on derringing and set the global dark and global light numbers to very low levels, like 0.01 and then try increasing the iterations - you will hopefully find you can get away with more iterations if you increase the derringing numbers slightly, but be careful not to oversharpen or introduce artifacts - common artifacts are light stringy-looking things, you'll know what I mean when you see them.

Artifacts of over-sharpening on planets are halo effects either just inside or just outside the limb of the planet, which I'm afraid you have on the right hand side of your second picture.

for DSO's, use the Richardson Lucy algorithm, and you don't have to do the shape+stddev guesswork, you use the Dynamic PSF tool, click on a bunch of stars in the image (about 40), click on the 'camera' icon on the Dynamic PSF and it gives you a little psf file, then load that in the External PSF tab, and do the rest of it as above.

Thanks again Stuart.  I will try some of these things on other images.  I think I've about reached the limit of where I can go with my Jupiter (first post).  I see the halo with the second attempt.  One of the issues with this astrophotography lark is knowing when to stop.

Steve