NGC 7635 - an 82-hour bubble

[Petergoodhew]

A more conventional target for a change - and one of my favourites. This time revisited with more data.

82.5 hours total integration HaOIIILRGB
 

Image captured on my dual rig at EyE, Extramadura, Spain

APM TMB 152 F8 LZOS, 10 Micron GM2000HPS, QSI6120ws8

 

 

Edited July 26, 2019 by Petergoodhew

[Brian Maurer]

Wow, you got so much detail.

[Ouroboros]

Wow is the word!  Amazing image.  

[Petergoodhew]

7 hours ago, Brian Maurer said:

Wow, you got so much detail.

Thanks Brian - the secret is lots of data, and then using just the best for pulling out the detail. Here I had 128 30-minute Ha subs but used only 15 for the fine detail. Deconvolution, high-pass filtering, and Annie's Actions helped too.

[alan4908]

That is a very impressive Bubble Peter.  Excellent stars and good detail.

One thought....on the Ha 30min subs used for the detail you mentioned above that you decided to use the best 15 out of 128. I was curious to know if you'd tried using an alternative approach: use a much higher percentage of subs,  which would give you a slightly less detailed result but much improved signal to noise ratio.  This would enable you to apply a much stronger deconvolution which would assist in recovering the lost detail. 

Alan

Edited July 18, 2019 by alan4908

[Petergoodhew]

On 18/07/2019 at 15:50, alan4908 said:

That is a very impressive Bubble Peter.  Excellent stars and good detail.

One thought....on the Ha 30min subs used for the detail you mentioned above that you decided to use the best 15 out of 128. I was curious to know if you'd tried using an alternative approach: use a much higher percentage of subs,  which would give you a slightly less detailed result but much improved signal to noise ratio.  This would enable you to apply a much stronger deconvolution which would assist in recovering the lost detail. 

Alan

Hi Alan, yes I've experiemented with different selections of subs.  The sugnal is very strong in these and so signal to noise isn't a problem. 15 subs represents 7 1/2 hours of data. Here I was able to do 200 iterations of deconvolution. I find that larger numbers of subs does reduce the clarity, and so settled on 15.
Peter

[upahill]

Amazing picture, the time invested alone shows the dedication and the result is superb.

 

[Datalord]

6 hours ago, Petergoodhew said:

I find that larger numbers of subs does reduce the clarity, and so settled on 15.

This is so interesting. I'm currently processing NGC6888 and have a thread on the selection on subs here: 

I came to the exact opposite conclusion. Difference is that I don't have the final image to show for... yet. 🙂

And btw, fantastic image. Well done!

Edited July 19, 2019 by Datalord
Forgot praise

[ollypenrice]

Very fine result, Peter. The refractors strike again!

Olly

[sloz1664]

Excellent image, but....... 82 hours, that's nearly a years worth 😉

Steve

[Petergoodhew]

On 19/07/2019 at 23:23, Datalord said:

This is so interesting. I'm currently processing NGC6888 and have a thread on the selection on subs here: 

I came to the exact opposite conclusion. Difference is that I don't have the final image to show for... yet. 🙂

And btw, fantastic image. Well done!

I don't this there's a right or wrong to this.  I've some images where selected the best subs made no difference at all - and others where it does. I think it will be down to the data. If there's very little difference between the quality of the subs, and especially if the target is faint, adding more will of course improve signal strength and thus help with deconvolution. However if a few subs are of exceptional quality, only using them could turn out to be the best strategy.