M51 stretch comparison - General Hyperbolic Stretch

Astrophotography: share your photos & discuss techniques
User avatar
menardre
Vice President
Posts: 920
Joined: Tue Jan 17, 2012 6:09 pm

M51 stretch comparison - General Hyperbolic Stretch

Unread post by menardre »

This past April I imaged M51 and processed using PixInsight.

One of the most important components of image processing is 'stretching' the image. When you look at the histogram of an 'un-stretched' deep sky image you will see that the only 'data' is at the extreme far left of the histogram. The histogram essentially shows how many pixels there are in the image at different levels of brightness with total black on left and total white on the right. If you were to zoom in the the left side of the histogram you will see that there is a 'hump' of data at the far left. The act of 'stretching' essentially pulls that hump to the right where it becomes visible.

There are several methods for stretching. When I am processing galaxies, nebulae, or globular clusters I always utilize ArcSinH stretch. This type of stretch enhances the dimmer objects more than the brighter objects, which is important for those types of objects. For open clusters I use Histogram Transformation. In both cases I use an iterative process where I employ several small stretches rather than one major stretch.

PixInsight has a new stretch script called General Hyperbolic Stretch. This script actually allows you to use ArcSinH, Histogram, and other types of stretches, but also includes a new stretch GHS. The beauty of the GHS is that it allows you to select the symmetry point at which the stretch starts. You can also do several stretch iterations focusing on particular aspects of the image. This type of stretch is a bit more complex, but gives you a lot more control. I just starting playing with this script and processed M51 since the tutorial also used M51 so I could just follow along.

Attached is my original processed image processed using ArcSinH and a second image using GHS. It should be noted that the original image also used star reduction, Range selection mask, High Dynamic Range histogram, and curves processes while the GHS version only used the General Hyperbolic Stretch. I could have further enhanced the GHS version but I thought it would be more informative to leave it 'as-is'. Also, this is my 1st attempt at using GHS and I know I have a lot to learn to fully use its capabilities. I think that both images are really good, but I can see benefits from using GHS once I learn the fine point.

Just thought I would pass along this new stretching technique to anyone interested.

Roger

Original
M51 Master_DBEdiv_PCC_EzDN_ArcSinH_RSmask_Curves_HDR_EzSR_Curves.jpg
M51 Master_DBEdiv_PCC_EzDN_ArcSinH_RSmask_Curves_HDR_EzSR_Curves.jpg (14.93 MiB) Viewed 1288 times
GHS version


M51_GHS1.jpg
M51_GHS1.jpg (14.26 MiB) Viewed 1288 times
Roger M.
Celestron CPC1100 EDGE, Stellarvue 130T refractor dual mounted on iOptron CEM120 on permanent pier mounted in Observatory. Imaging camera ZWO ASI2600 OSC, guide camera Lodestar or ZWO ASI290MM.
User avatar
Apollo XX
MSSF Coordinator
Posts: 2795
Joined: Sun Jul 11, 2010 8:11 pm

Re: M51 stretch comparison - General Hyperbolic Stretch

Unread post by Apollo XX »

It always amazes me the hidden data that an image file contains that may not necessarily be revealed in the processing workflow. It's almost as though one should never discard data, for the possibility that it contains previously unrealized features that may be revealed by future processing attempts or new processing developments. Good for you for keeping up with the developments, Roger!

Mike M.
"The purpose of life is the investigation of the Sun, the Moon, and the heavens." - Anaxagoras
User avatar
menardre
Vice President
Posts: 920
Joined: Tue Jan 17, 2012 6:09 pm

Re: M51 stretch comparison - General Hyperbolic Stretch

Unread post by menardre »

Mike

Thanks ...

The problem is that sometimes I find a new technique and I am tempted to re-process older images.

Roger
Roger M.
Celestron CPC1100 EDGE, Stellarvue 130T refractor dual mounted on iOptron CEM120 on permanent pier mounted in Observatory. Imaging camera ZWO ASI2600 OSC, guide camera Lodestar or ZWO ASI290MM.
User avatar
mark.m
President
Posts: 310
Joined: Fri Feb 03, 2017 1:13 pm

Re: M51 stretch comparison - General Hyperbolic Stretch

Unread post by mark.m »

I much prefer the image with the GHS processing (I saved it as wallpaper on one of my computers).

I think the little tiny background galaxies are really cool!

- Mark M
Mark M, AJ1B
Portsmouth, RI
Celestron 14" and Meade 10" SCTs
QHY268M + SBIG ST-9
GM2000 (10Micron)
American Association of Variable Star Observers (AAVSO) observer code: MMU
User avatar
menardre
Vice President
Posts: 920
Joined: Tue Jan 17, 2012 6:09 pm

Re: M51 stretch comparison - General Hyperbolic Stretch

Unread post by menardre »

Mark

I agree ... The advantage of the GHS is that you graphically view the image and 'probe' the image at various points to identify what is background and what is foreground. Since you do not want to stretch the background, you then select the point where the foreground image 'starts' (i.e. is dimmest but above the level of background). Then you use that value as the symmetry point for the stretching. This gives more control over the stretching process.

Once you do the initial stretch you then view the histogram on the log scale. This helps to identify if there are any areas that need further selective stretching. This way you can focus on particular aspects of the image without stretching everything.

I think GHS is a powerful tool, but in order to utilize all of its capabilities you have to spend some time with it.

Roger
Roger M.
Celestron CPC1100 EDGE, Stellarvue 130T refractor dual mounted on iOptron CEM120 on permanent pier mounted in Observatory. Imaging camera ZWO ASI2600 OSC, guide camera Lodestar or ZWO ASI290MM.
Post Reply