Thoughts on which imaging rigs to concentrate on

[Gina]

Well there's not a lot of OIII in Orion!  Just the stars and M42, and some in the Rosette but can't see much else.  This seems to be a general thing - so much more Ha than anything else.  I don't think there's much point in increasing exposure.

Edited January 18, 2017 by Gina

[Gina]

Now processing the two sets of Ha subs.  Blink rejection of duff frames and calibration done and now registering.

[Gina]

Here's a screenshot of the two integrated sets in PI. Plus same histogram stretched.

Edited January 19, 2017 by Gina

[Gina]

Full resolution versions.  Both are the integration of 98 60s subs, gain 600, -30C.

[Gina]

I'll leave the imaging system taking OIII 60s subs overnight (what's left of it) and go to bed.  I'll process the OIII and play with the Ha star removal and subtraction tomorrow (later today actually).

Night all

[Filroden]

That's a beautiful Orion with the Ha 3nm. It's amazing how small the stars are in comparison to using the L filter. I can't wait for mine to arrive so i can have a go at the Rosette Nebula. However, it looks like your Ha 5nm filter is not parfocal as the image looks out of focus.

I was also thinking overnight about how to balance the image intensities so subtraction would remove the Ha from the image. I think the subtraction needs to be done in the linear state but some stretching might be needed to equalise the Ha intensity (or match background intensities). Otherwise you'll also see the difference in transmissions between the two filters and not just the different bandwidths. I wonder it you could measure the intensity around the Horsehead in one image and stretch the other to get a matching value? Or would you just add a constant value to the entire image to equalise (given it's linear and you want it to remain linear)? So if the background had an intensity of 0.1 in the first image and 0.15 in the second, to add a fixed 0.05 to the entire image in the second. I guess comparing the values in a few areas where there should be no Ha or N, or where there is only Ha to ensure they remain in balance?

[Gina]

Yes indeed. I was surprised to see rings instead of dots for the Ha 5nm indicating it's out of focus - I did think they would be parfocal so didn't check.  Might be clear again tonight so I'll have another go and refocus.

I too have been thinking about how to correct for variation in intensity and contrast when subtracting images.  This is undoubtedly going to be tricky.  Yes, the images must remain linear but I agree will most likely need matching in both intensity level (offset) and contrast (multiplier).  There is supposed to be a PI function to do just this for matching up RGB images but I don't think I'll rely on that as previous usage has not proved very good on RGB.  Forget what it's called ATM - PI naming doesn't quite match my logic

Once I have good focus on both sets of images I plan to take the integrated images, match using StarAlignment then play with PixelMath to try to match them up.  I think this is going to be interesting

[Filroden]

I've heard of a beta module called dnaLinearFit that might do the trick. I think it's linked on the PixInsight forums. 

Edit: found it http://trappedphotons.com/blog/?p=994 and go to the Matching Intensity section

Edited January 19, 2017 by Filroden

[Gina]

Very interesting and shows the problems I've had with LinearFit.  This new version looks like it should do the job

[Gina]

OIII integration auto-stretched in PI.

[Gina]

Full resolution version straight from integration in PI.  130 OIII subs of 60s gain 600 and -30C.

[Filroden]

Looks like plenty of OIII in M42 area, the Flame and Rosette, plus a few small bits here and there.

[Gina]

 Just refocused the Ha 5nm and it's all of 25 counts away from the 3nm filter (-23 counts v +3 counts).  Amazing!  It's as if the filter was a different thickness.

[Gina]

Now capturing images with the new focus position and have also adjusted the framing slightly to include a previously unseen object near to the Rosette Nebula - there is nothing on the other side of the image being lost in this readjustmant of FOV.

Here is an auto-stretch of a single 60s frame in PI.

[Gina]

Here is a full resolution PNG saved from PI.

[Gina]

Focus is alright now but the field is not flat or may be slightly tilted - stars are elongated in the RHS of the frame.  I plan to retake at least 100 of both 5nm and 3nm Ha subs with the new framing.

Edited January 19, 2017 by Gina

[Gina]

109 Ha 5mn subs captured - 103 passed Blink and now calibrating.  Currently capturing Ha 3nm subs having refocused - it was 1 count different from yesterday (now +4 counts).

[Gina]

Calibration done now registering - StarAlignment.

[RichLD]

Watching with much interest!

[Gina]

Registered now integrating (stacking) images.

[Gina]

Finished integration   Auto-stretched integrated image in PI.  Some amp glow not fully subtracted but not too bad.  Angelfish looks like an alien

Edited January 19, 2017 by Gina

[Gina]

Full resolution integration untouched from PI - just saved as PNG.

[Filroden]

54 minutes ago, Gina said:

Angelfish looks like an alien

I'll never be able to not see that now

[Gina]

I have plenty of data from last night to process   I gave up relatively early as I couldn't find anything else to image with that FOV - everything of interest had sunk out of view.  I think I've now exhausted the available views with the current 28mm lens so planning to change lenses for tonight hoping that the sky will remain clear.  Have to have a good think about what to image  

I have realised why I was having a problem with M42 - I'm now not using high clipping as M42 is expected to saturate in order to capture other DSOs in Orion.  So I might have another go at Orion - Flame through Horse's Head to M42 area.

I would normally take flats but the images captured since the clean-up and lens change are not showing dust bunnies or any vignetting so there seems no need to bother (unless data processing from last night shows a problem).

[Gina]

Update...  109 Ha 5nm subs processed last night.  That leaves Ha 3nm subs of Orion, Barnard's Loop etc. and other odds and ends to be checked with Blink and further processed.