sh2-119 in Ha

[Rodd]

FSQ 106 with .6x reducer, ASI 1600 with 3um Ha filter.  209 60 sec subs.  I was a bit underwhelmed by the results of this one, having been very satisfied with the Ha stack of the Veil.  I suppose the Veil Nebula is a bit brighter than Sh2-119, but I thought 209 subs would make up for it somewhat.  I am tempted to double my exposure time to 120 sec for this one.  I would get less subs, but I would add those to these and benefit from a high count with perhaps  third being 2min in duration.  Then again, the longer I look at this image, the less I feel it needs a boost of depth and signal--except perhaps around the edges of the image where there should be smooth, faint nebulosity.  On the other hand, when I find myself trying to convince myself that "its fine", most often it is not.  Conditions were questionable.  It was clear, but even with good focus the FWHM values were between 3.5 and 4.5--sometimes 5.  The stars don't look bloated.  I suppose with a pixel scale of 2.46 arcsecs per pixel, a FWHM value of 5 is still only 2 pixels.  Nyquist says 3 is optimum.   

 

[Rodd]

Here's 194 best subs.  Not sure if the differences are due to the subs election, or more careful processing.  Still unsure about collecting 60 or so 2min subs.  I was doing 5min with the TOA 130 at F7.7.  I figure at F3, 1-2min should be enough.  Full resolution mode is where this image deteriorates.

 

 

 

[Petergoodhew]

Stunning detail and clarity Rodd.  Seriously good image.

[Rodd]

48 minutes ago, Petergoodhew said:

Stunning detail and clarity Rodd.  Seriously good image.

Thanks peter. I think I improved it allot with the last one

Rodd

[wimvb]

You can use the subframe selector script in PixInsight to add a weight to each sub, based on fwhm value. This way, the sharper subs add more to the overall result. As an alternative, you can set a limit in the script to accept only subs with the best fwhm.

[Rodd]

2 hours ago, wimvb said:

You can use the subframe selector script in PixInsight to add a weight to each sub, based on fwhm value. This way, the sharper subs add more to the overall result. As an alternative, you can set a limit in the script to accept only subs with the best fwhm.

I don't think that having more weight in subframe selector means that the subs add more to the image.  I thought that was just a ranking system to identify the best subs.  I chose the best 194 by using FWHM, eccentricity and SNR.  Amounts to the same.

Rodd

[wimvb]

3 hours ago, Rodd said:

I don't think that having more weight in subframe selector means that the subs add more to the image.  I thought that was just a ranking system to identify the best subs.  I chose the best 194 by using FWHM, eccentricity and SNR.  Amounts to the same.

Rodd

Have a look here:

https://pixinsight.com/doc/tools/ImageIntegration/ImageIntegration.html#description_002

In image integration, when calculating the average of pixel values, these values are multiplied by the weight assigned to each sub. The weight parameter that is used by default is the snr which is calculated during noise evaluation in the calibration process. Selecting subframes is in principle the same as assigning a weight factor of either 1 (in) or 0 (out).

[Rodd]

8 hours ago, wimvb said:

Have a look here:

https://pixinsight.com/doc/tools/ImageIntegration/ImageIntegration.html#description_002

In image integration, when calculating the average of pixel values, these values are multiplied by the weight assigned to each sub. The weight parameter that is used by default is the snr which is calculated during noise evaluation in the calibration process. Selecting subframes is in principle the same as assigning a weight factor of either 1 (in) or 0 (out).

I stand corrected--but I wonder how much of a difference there would be.  I have used SFS to choose my best subs and output those subs to a new directory so they carry the SFS weight header--but I don't see a big difference.  Sometimes just by virtue of a higher sub count, leaving inferior subs in can bring down noise and increase SNR (as long as the subs are not unfit).  I took out 20 subs with high FWHM and the final image had a FWHM less than .1 arcsec lower--but the SNR was lower and the noise higher.  I know its all a balancing act.  I guess the experiment would be--leave all subs in but use SFS to weigh them and integrate using SFS and then without the SFS weight analysis.  I think this is a case where theory may outshine a bit (assuming the same subs were used, one with the weight data and one without the weight data).

Anyway--I added 58 2min subs (couldn't get 60 due to clouds).  So the image now has 252 subs, 194 1min and 58 2min.  I also used less noise control and was more conservative with the brightness.  I think its the sheer number of stars that has me annoyed.  Maybe if my FWHM values were much lower, it would be better.  But that would mean starting over with new data!

[tooth_dr]

That last one is a signicant improvement in my untrained eyes!  Very well done!  Dont be too hard on yourself!

[Rodd]

1 hour ago, tooth_dr said:

That last one is a signicant improvement in my untrained eyes!  Very well done!  Dont be too hard on yourself!

Thanks Doc.....I think I have enough Ha now.  Looking forward to getting the other channels.  

Rodd

[ollypenrice]

The object and framing are sensational. There is something slightly 'scratchy' about the look which, I guess, just needs more data to rectify. The Veil is bright - and easy, visually, even in small apertures. I doubt that this object is at all comparable but what a treat to see it.

Olly

[Rodd]

44 minutes ago, ollypenrice said:

The object and framing are sensational. There is something slightly 'scratchy' about the look which, I guess, just needs more data to rectify. The Veil is bright - and easy, visually, even in small apertures. I doubt that this object is at all comparable but what a treat to see it.

Olly

Thanks Olly--"Scratchy" is a good way to put it.  So you think more data huh?  I guess it makes a degree of sense--I only have a bit over 5 hours--not a huge amount.  I did see quite an improvement with 120 sec vs 60 sec subs.  Maybe this target is just not bright enough for 60 sec subs.  A couple more hours of 2min subs should help.  

Rodd

[Allinthehead]

That's an interesting target nicely done. I find a lack of data in combination with short subs with the 1600 gives that scratchy look. I'm imaging at a similar  scale to you at f3.3 and find i need at least 7 hours ha to get a smooth background. If you're using 2 minute subs is this at unity? I would increase to 5 minutes if this is the case.

[Rodd]

26 minutes ago, Allinthehead said:

That's an interesting target nicely done. I find a lack of data in combination with short subs with the 1600 gives that scratchy look. I'm imaging at a similar  scale to you at f3.3 and find i need at least 7 hours ha to get a smooth background. If you're using 2 minute subs is this at unity? I would increase to 5 minutes if this is the case.

I did 5min with the TOA 130 at F7.7 for M16 and M17 and only needed 3-4 hours.  Those targets are brighter, I relaize.   If I have to go much above 2min it defeats the purpose of shooting at F3.  The veil worked well at 60sec and it only had 4 hours total (2 of Ha and 2 of OIII)-but again, its a bit brighter.  I think another 2 hours of 2min subs should solve it.  I could change teh gain--but I do not really know about gain.  I guess if I make the higher I can leave the exposures shorter?  Don't really know.

Rodd

[wimvb]

Your 2nd version you managed to reveal the contrast between bright areas and dark patches within the cloud. This is a great capture of a very interesting target.

If fwhm bothers you, you can try deconvolution on just the stars. Either use a star mask to target the stars, or increase the wavelet layers and noise threshold in the deconvolution process. This will tighten the stars without affecting the nebula. At the end of postprocessing you can dim/reduce many of the smaller stars with MT (morphological transformation), and make the nebula "pop" a little more.

 

[Rodd]

1 hour ago, wimvb said:

Your 2nd version you managed to reveal the contrast between bright areas and dark patches within the cloud. This is a great capture of a very interesting target.

If fwhm bothers you, you can try deconvolution on just the stars. Either use a star mask to target the stars, or increase the wavelet layers and noise threshold in the deconvolution process. This will tighten the stars without affecting the nebula. At the end of postprocessing you can dim/reduce many of the smaller stars with MT (morphological transformation), and make the nebula "pop" a little more.

 

I once compared a high FWHM sub with a low FWHM sub and there were discernable differences in the non stellar details.  So, its not necessarily the stars that bother me when the FWHM values are high--but the reduction in sharpness if fine structural details.  It makes sense--if FWHM is used to assess focus--higher than normal FWHM can mean a softness in resolution. (I refer to significant differences in FWHM). But this one is not bad with respect to that.   I have never had much luck with Decon in the linear state--can't seem to get the hang of it.  I use it for fine sharpening though (sometimes).  MT is another tool that I sometimes use but sometimes I don't like its appearance.  In this image, for instance, a touch of MT is ok--but too much and it looks unnatural.   

More data I suppose is the best solution to many poblems.  That and color--adding the other channel wwill certainly help.

Rodd

[Rodd]

4 hours ago, ollypenrice said:

There is something slightly 'scratchy' about the look which, I guess, just needs more data to rectify

 

8 hours ago, tooth_dr said:

That last one is a signicant improvement in my untrained eyes!

 

3 hours ago, Allinthehead said:

I find a lack of data in combination with short subs with the 1600 gives that scratchy look

Het guys--just thought I'd try  drizzle integration.  Drizzle is best if used with allot of subs--well 250 seems like enough.  The image is 4x the size of a normally integrated image--but I think that only shows up under full resolution viewing.  Not sure its the way to go--thought I'd try it and see.  I like the normal size--the full is too much as usual.  Edit:  I down sampled the drizzled image so its the same size as normal (because I like the version best I think.  Just experimenting)

 Down sampled