M8 Lagoon. First ever LRGB image

alpal · #1 14-09-2012, 01:15 AM

This is my first LRGB, Ha picture with my new QHY9.
3 x 3 minutes each of LRGB plus 10 minutes of Ha 7nm.
Total 46 minutes of data.

Telescope 8" Newt f6, modified NEQ6 mount,
guided with TS9 OAG, PHD guiding & Lodestar, used Ezycap for capturing subframes.
No: flats, darks or dithering were used.

This the first time I processed an LRGB image - used PS CS5.
It was easier to process than a DSLR image.
Stacked in DSS.
Used an RCC corrector which needs further adjustment.
Just checked the secondary mirror -
it needs cleaning & is the cause of the dust doughnuts.
The primary mirror & rest of the optical chain is very clean.
Pic is slightly cropped because the Ha subframe wasn't perfectly aligned with the other frames.

All comments are welcome.

A larger version is here:
http://www.flickr.com/photos/2471943...n/photostream/

RickS (Rick) · #2 14-09-2012, 07:31 AM

A good first attempt! The KAF8300 sensor has shallow wells so it takes some effort to get good star colours (adding some short exposures with HDR is worth trying). Surprising how much Ha you got in 10 mins.

The big dust donut isn't going to be from your secondary. The dust will be much closer to the sensor than that. Here's a calculator that will help you determine where it is:
http://www.ccdware.com/resources/dust.cfm

Cheers,
Rick.

gregbradley · #3 14-09-2012, 08:37 AM

A great first image.

Dust donuts are removed by flats and are typically on the filters or the sensor window.

Its a good practice to keep them clean and not rely totally on flats to get rid of them.

An easy way to do flats is a white tshirt over the end of the scope, point towards the NW at dusk and take 3-5 images with each filter
and aim for an adu level of around 20,000. Then take darks for the same length exposure and do median combine in CCDstack and dark subtract them to create a master flat.

Now you have a master flat for each filter.

If you are lazy you can just take one for luminance and use one of the coloured filters only. You can often get away with that unless one of the other coloured filters is really dirty.

Greg.

Meru (Michael) · #4 14-09-2012, 03:54 PM

Very, very nice Greg. It's always a bit of challenge getting the dynamic range right but I think you've got it down pat. Maybe bump up the saturation a little? Just my personal preference anyways

multiweb (Marc) · #5 14-09-2012, 04:58 PM

Pretty cool - nice core details.

alpal · #6 14-09-2012, 07:05 PM

Quote:

Originally Posted by RickS

A good first attempt! The KAF8300 sensor has shallow wells so it takes some effort to get good star colours (adding some short exposures with HDR is worth trying). Surprising how much Ha you got in 10 mins.

The big dust donut isn't going to be from your secondary. The dust will be much closer to the sensor than that. Here's a calculator that will help you determine where it is:
http://www.ccdware.com/resources/dust.cfm

Cheers,
Rick.

Thanks Rick,
When I sampled all the bright stars in Photoshop curves -
none reached the top of the curve so I assumed that I had not saturated
any of the pixels - I therefore expected the star colours to be there.
I also ran the recommended gain of 5% & offset of 100.
I just checked all the bright stars on the curves eyedropper function
& they all read 217 out of 255 gray scale.
I checked the same strength levels on a particular medium brightness star on an 4 original RAW frames & I got:
Red = 72
Green = 76
Blue = 70
Lum. = 150

I'll have to think about those results as that's telling me that the stars are actually white.

Thanks for the dust mote calculator link.
I put in correct values & for the giant dust mote
I got 3.029 mm from the sensor.
For one of the small motes I got 0.68mm from the sensor.
That's telling me that I have dust most likely on the sensor itself or very close to it.
I'll check that out.

cheers
Allan

alpal · #7 14-09-2012, 07:09 PM

Quote:

Originally Posted by gregbradley

A great first image.

Dust donuts are removed by flats and are typically on the filters or the sensor window.

Its a good practice to keep them clean and not rely totally on flats to get rid of them.

An easy way to do flats is a white tshirt over the end of the scope, point towards the NW at dusk and take 3-5 images with each filter
and aim for an adu level of around 20,000. Then take darks for the same length exposure and do median combine in CCDstack and dark subtract them to create a master flat.

Now you have a master flat for each filter.

If you are lazy you can just take one for luminance and use one of the coloured filters only. You can often get away with that unless one of the other coloured filters is really dirty.

Greg.

Thanks Greg,
I'll be sure to take good flats next time.
I probably could still take: flats , darks & bias & then reprocess.
The camera was running at -35 degrees & I couldn't see any noise with
RAW subframes except a few hot pixels. (only 3 minute subframes )
It's a lot better than my noisy old EOS 1000d -

cheers
Allan

alpal · #8 14-09-2012, 07:15 PM

Quote:

Originally Posted by Meru

Very, very nice Greg. It's always a bit of challenge getting the dynamic range right but I think you've got it down pat. Maybe bump up the saturation a little? Just my personal preference anyways

Thanks Meru,
I think you meant Allan?
I'll try to work out a way to get some more star colour.
I think the saturation is about right -
I don't like "chocolate box " pictures but
it's subjective & just a matter of personal taste.
Is it worth re-processing with such little data &
when the new coma corrector needs to be adjusted slightly?

cheers
Allan

alpal · #9 14-09-2012, 07:17 PM

Quote:

Originally Posted by multiweb

Pretty cool - nice core details.

Thanks Marc - it's amazing how little data you need with a mono camera to get detail, especially on an easy target
although there is a lot of light pollution where I am.

cheers
Allan

atalas · #10 14-09-2012, 07:28 PM

Nice start Al!

alpal · #11 14-09-2012, 07:43 PM

Quote:

Originally Posted by atalas

Nice start Al!

Thanks Louie,

What did you think of my reply - am I correct?

quote Allan

Quote:

When I sampled all the bright stars in Photoshop curves -
none reached the top of the curve so I assumed that I had not saturated
any of the pixels - I therefore expected the star colours to be there.
I also ran the recommended gain of 5% & offset of 100.
I just checked all the bright stars on the curves eyedropper function
& they all read 217 out of 255 gray scale.
I checked the same strength levels on a particular medium brightness star on an 4 original RAW frames & I got:
Red = 72
Green = 76
Blue = 70
Lum. = 150

I'll have to think about those results as that's telling me that the stars are actually white.

atalas · #12 14-09-2012, 07:57 PM

Al,at 3 minute exposures I doubt that the stars are saturating....more likely your post processing.....your RGB readings confirm that.

How did you combine your Luminance to your RGB?

alpal · #13 14-09-2012, 08:08 PM

Quote:

Originally Posted by atalas

Al,at 3 minute exposures I doubt that the stars are saturating....more likely your post processing.....your RGB readings confirm that.

How did you combine your Luminance to your RGB?

Hi Louie,
I set the blending mode to luminosity for the luminance.

I had all 5 layers as:
Red to 0 degrees set to screen, opacity 100%
Green to 120 degrees set to screen, opacity 100%
Blue to 240 degrees set to screen, opacity 100%
Ha to 0 degrees set to screen, opacity 18%
Luminance set to luminosity 100%

I made sure when I stretched that I never allowed the stars to reach the maximum 255.
Is that correct?

cheers
Allan

atalas · #14 14-09-2012, 08:54 PM

Al,you combined RGB in PS?and you used screen blend?why?not surprised the star colors washed out!

It's important that you combine your RGB in pre processing and normalize LRGB brightness levels before going into PS,and there's no need to split RGB in PS for LRGB.

alpal · #15 14-09-2012, 09:12 PM

Quote:

Originally Posted by atalas

Al,you combined RGB in PS?and you used screen blend?why?not surprised the star colors washed out!

It's important that you combine your RGB in pre processing and normalize LRGB brightness levels before going into PS,and there's no need to split RGB in PS for LRGB.

Hi Louie,
I just got the idea from a Ken Crawfords video on clipping masks for narrowband.

see here:
http://www.imagingdeepsky.com/Tutorials/Tutorials.html

I didn't follow the video exactly but I could see the idea of having
each colour as another layer in screen mode
& the luminance as luminosity blend mode.
This was the first time I processed LRGB so I didn't know
there was any other way.
It seemed to work quite well & was easy to do except that the stars went all white.
Sorry - I don't know of any other way.

Also here:
http://www.jemastrophoto.com/tutoria...-in-photoshop/

seems to do the same thing?

cheers
Allan

alpal · #16 15-09-2012, 01:42 AM

After some advice from Louie on channels, I reprocessed my photo.
I tried to get some more star colour.
Is this any better?

http://www.flickr.com/photos/2471943...n/photostream/

cheers
Allan

atalas · #17 15-09-2012, 10:59 AM

Al,there's a color bias to the red,the stars have a harsh multicolored halo around them with what looks like damaged cores, the shadows end looks clipped,and also looks over sharpened.

When stretching an image you are altering brightness levels and so, If all channels aren't normalized(equal brightness)(I do this in pre processing)and knowing this, you can now see why color balance can be lost when stretch the image.

Peter.M · #18 15-09-2012, 11:18 AM

Damaged cores are probably the smaller HA stars showing through.