Saturated star cores and LRGB imaging Ritchey Chretien telescopes GSO Family · hughsie · ... · 4 · 204 · 0

hughsie 0.90
...
· 
·  Share link
Hi everyone.

I am using a RC8 for galaxy season and my first two images have shown saturated star cores. My camera is an ZWO ASI1600mm pro cool using unity gain (139) and an offset of 30. Filters of choice are Chroma 1.25” LRGB.

For M94 (see gallery), I have been using exposure times of 120s on L and the same on RGB with binning of 2x2. I then switched to M100 and used 1x1 binning. L was reduced to 60s but I kept RGB at 120s.

The full well depth of the 1600mm is low so I was considering reducing the gain to increase this depth and avoid saturation on my broadband images and was wondering what other RC images do? I only ask this as I see a mix of gains when reviewing images, some 0, some 75 some at unity gain...all pretty confusing.

I would really welcome your thoughts.

Thank you in advance.

John
Like
astrod 2.15
...
· 
·  1 like
·  Share link
Hi John

I see you have 1.814 arcsec/pixel which is too high for galaxies.  You should aim for about a third of that, which means 1/9th of the number of photons per pixel of what you're doing at the moment.  This number is a mixture of optical, telecompressor, pixel size and binning.  Are you running at F/8?  Seems that in the post processing you're resampling.

Centres of bright stars will always saturate a bit.  Generally I'm using 60sec for L at 0.65 arcsec/pixel, and 120sec for RGB, at F/6. So your exposure times seem fine. With these settings the subframes look quite thin. I just use gain 139, so varying the gain probably isn't addressing your problem.

Perhaps if you can share your subframes and L master frame (calibrated and stacked but un-stretched) then some diagnosis can be done.

CS Rod
Edited ...
Like
hughsie 0.90
...
· 
·  Share link
Hi Rod, thank you for your reply.

I use PixInsight for image processing and with M94 which was binned 2x2 at the camera I used the ‘resample’ tool to reduce the image width to 1200 pixels. With M100 where I used 1x1 binning, I used the tool IntegerResample to downsize by a factor of 2 then resampled the image size to 1200 pixel width. 

From what you are saying my exposure times are ok but I should use IntegerResample if needed where I have captured the image at Bin 1x1 and avoid the second resampling altogether?

All the best

John
Like
hughsie 0.90
...
· 
·  Share link
Just revisiting my most recent comment. 

When I run the ImageAnalysis script on M100 bin 1x1 it reports a resolution of 0.487 arcsec/px  which is what I expect. Software binning the image to 2x2 increases the resolution to 0.973 arcsec/px. If I then ran the Resample process and reduce the image width to 1200 px the resolution increases further to 1.814 arcsec/px which is where I do not want to be. So, from now on I will not use the Resample process. I am happy to share the L master, however, the max. file size is 10mb and I the file is way in excess of that.....unless I am missing something.
Like
astrod 2.15
...
· 
·  Share link
OK I can see the numbers now make sense. Technically you would say the resolution is decreasing when the arcsec/pixel goes up; but I know what you meant. When you double the arcsec/pixel you're essentially putting 4x the number of photons in each pixel, so that increases the tendency to saturate.

Personally I would do processing at the bin 1x1 level and if I wanted to resample do that at the last step.  Of course 0.487 arcsec/pixel is oversampling but after a lot of testing found 0.64 arcsec/pixel is the sweet spot (when the seeing is excellent) at least for galaxies.  I'm using a tele-compressor to get this number and run at about F/6 rather than native F/8, but I think you'll be ok without a tele-compressor.

With sharing the Master image etc I think we can exchange using email or a dropbox link rather than spamming as an Astrobin image.  I'll try and contact you directly through the Astrobin messaging interface to facilitate this (btw we are out of sync as I'm in Australia).  I use PixInsight too.
Like
 
Register or login to create to post a reply.