Thoughts on which imaging rigs to concentrate on

[Gina]

Been sorting out all my data - I think I now have it organised.  I'm putting source files on my USB3 2TB external drive except those found to be poor or duff in Blink - they get deleted.

Now processing the Heart and Soul in Ha - I find I have 419 lights of 60s with gain of 440 and mostly -30C.  The data looks good so I should have a good Ha result.  Nearly 7 hours of data

[Gina]

Been processing the Ha data.  BPP crashed at the integration stage so I'm running ImageIntegration separately on the calibrated and registered frames - all 419 of them - it's taking hours!!  On the last phase of pixel integration now.

I looked that the OIIIs  too after collecting them all together - don't know how I managed to capture so many - the total is around 1200!!!  Quite a percentage of these are dud, unusable due to bad framing or just poor.  With so many there's no point in using any but the best.  I've been through several hundred in Blink and dumped a lot.  There's still more to do.

[Gina]

Well...  the result of integrating 419 Ha subs of the Heart and Soul is absolute RUBBISH!!!    The result from the other night with just 35 subs was very much better!

[Filroden]

That looks like your star alignment hasn't worked. *edit* Actually, it does look aligned. Just noisy. Did some cloudy subs sneak in?

Edited December 6, 2016 by Filroden

[Gina]

I think I'll have to Blink all those Ha subs again but I thought I'd dumped all the poorer subs.  Now I have calibrated and registered frames I guess it's best to use Blink on those rather than the originals.  I'll do that...

[Gina]

Now been all through the Ha subs and towards the end there was definitely thin cloud going through so I've dumped all those.  279 subs left which all seem very good.  I make that 140 dumped.  Now I'll run ImageIntegration again...

[Gina]

Curses - PI has dropped out!

[Gina]

Still going...

[Gina]

I think I'm better off limiting capture to just a hundred or two - or even fewer.  Otherwise processing takes an inordinate amount of time.  I need a super-computer

[Filroden]

11 minutes ago, Gina said:

I think I'm better off limiting capture to just a hundred or two - or even fewer.  Otherwise processing takes an inordinate amount of time.  I need a super-computer

Indeed. I upgraded mine (not just for this but it helped). It was even worse processing DSLR RAW files as they exploded in size once debayered. I find anything more than 300 subs something to walk away and do something else territory! The good thing is, it should only be a one time process.

[Gina]

Yes, I've been reckoning on around 300 subs and that's mainly what I've been doing.

Finished and result is much better.  Here is an unstretched screenshot.  278 subs integrated.  I'll post process in Photoshop but I think it only needs a slight "lift" in levels.

 

[Filroden]

It looks like it calibrated very well.

[Gina]

It needs the DBE treatment but here is just levels applies in Photoshop.  (I really should see how I do this in PI).  I shall need to remove the uneven borders resulting from registration but might as well leave this until after I register the different wavelengths.

[Filroden]

Use HistogramTransformation, selection your image in the view then clicking the live preview. Make sure to remove the visual stretch. Then it's just like Levels. Just drag your midpoint and apply when happy. It retains settings so your live preview will go crazy until you click reset (lower right icon in bottom bar). Repeat until happy. You can adjust black point too and it will show you number and percentage of pixels clipped. I aim never to exceed about 0.01% clip.

[Gina]

Thanks Ken

Here is a "preview" from Photoshop after applying levels and "Enhance DSO and Reduce Stars" plus cropping the borders.

[Filroden]

6 minutes ago, Gina said:

It needs the DBE treatment but here is just levels applies in Photoshop.  (I really should see how I do this in PI).  I shall need to remove the uneven borders resulting from registration but might as well leave this until after I register the different wavelengths.

You can register the images again afterwards. You don't want to apply DBE without cropping first. My workflow for L channel is crop, gradient removal, L and star masked deconvolute, L or range masked linear noise reduction, stretch, L or range masked non-linear noise reduction. I also throw in some L or range masked curves adjustments too if needed once non-linear.

[Gina]

Ah yes I remember now

[Gina]

Been through the OIIIs again and pruned drastically - now have 373 which I'm feeding into BPP.

[Gina]

Still going...  I'll let it run overnight.

[Gina]

I ordered the ZWO adapter to allow me to use the 45mm lens but since the delivery time is 10 to 20 working days, I've been looking into using my 35-105mm zoom lens with M42 mount.  This is as heavy or heavier than the 45mm medium format lens and more tricky to support - not much space to fit a bracket.  It would be nice to have remote zoom control as well as focus but this would involve more hardware and more software writing - probably won't bother - I can just set it to whatever FL I want by hand.

[Gina]

I actually used 301 OIII subs in BPP - here's the result, screenshot of the auto-stretched image in PI.

[Gina]

And now after post processing in Photoshop.

[Gina]

SII subs Blink'd and 95 60s g440 -30C fed through BPP.  Auto-stretched in PI.  Screenshot.

[Gina]

Photoshopped.

[Gina]

Think I might try combining the three integrated images into a colour version.  Starting here.

Edited December 7, 2016 by Gina