Dave

Eagle’s Astrophotography Guides.

Welcome to the third edition of my astrophotography guides.

I run a series of successful astrophotography workshops, covering DSLR astrophotography and
Webcam imaging and image processing.

My students asked if I had handouts of the techniques I used, but at the time I did not have anything
prepared.

As time has developed I have started to put together some of the techniques I use into these handy
guides.

These should help the newcomer to using these techniques to become more proficient at using the
software and achieving better results.


The way I outline in these pages is definitely not the only way to do things. Some people may
advise differently, but what follows in this guide is certainly the simplest way I have found to
achieve perfectly acceptable results.

I hope you find this guide useful.

If you have look out for the previous editions:

A Guide to Deep Sky Stacker: Deep Sky and Comet Images.

Solar Webcam Imaging.

More will be published in due course.

Dave Eagle.

January 2017

Essential Photoshop Techniques.

Photoshop has long been the mainstay of processing astrophotographs. It may not be dedicated for
astrophotography image processing, but it does have a lot of flexibility in its use, as well as a lot
of plug-ins that can increase the effectiveness of the software.

There are alternatives, such as Pixinsight, but these also have a steep learning curve and are less
useful for processing normal photographs.

In this booklet, I will guide you through some of the image processing techniques I use within
Photoshop so you can get a decent image from a final stacked image.

You may also have heard that you must use dark frames, flat fields, bias frames and other such
embellishments to produce astrophotographs you are proud of.

Indeed, the use of these techniques will certainly enhance your images, but as you will see, you
don’t have to use them to get a decent image.

When you are first starting out, there really is no need to go to these sorts of lengths.

In fact, virtually all the images I have taken over the year have not had these carried out on them. I
simply don’t have time or the inclination to fuss about that much. In fact, when I have tried to do
them, they seem to create more problems than they seem to solve.

I always try and keep my image processing to use the KISS Method: Keeping it Stupidly Simple.

What follows is my Photoshop image processing work flow.

Photoshop.

This is a fantastic piece of image processing software. The functions buried within it are many
and wonderful. This power is also its biggest downfall as using the software and wandering
amongst its many tools can be quite bewildering and intimidating to the beginner.
Like many processes, there is quite a steep associated learning curve. You can feel like you are
banging your head against a brick wall at times, when you can’t quite get your head around what
is happening or how to tame it.

Having gone through this pain barrier, when the light bulb moment occurs, it is a revelation. This
then enables you to progress much further in your processing skills.

This guide leads you through the essential processes I use to reveal the detail in my images.
Hopefully this will get you over that pain threshold and well on your way to mastering the
software.

In the past, the software was extremely expensive. Recently the licensing model has changed and
can be licensed on a subscription.

Licensing and purchase details for Adobe Photoshop CC here:
http://www.adobe.com/uk/products/photoshop.html

The edition called Adobe Photoshop Creative Cloud (CC) is available for the quite reasonable
cost of £8:57 a month for the Photographer edition (January 2017). Although you never own the
software, as someone who uses the software quite a bit, I think this is an extremely good
investment.

This will also ensure that you have the most up to date version and that it will be able to handle
RAW files from the latest cameras, if required.

Despite the name, Cloud, the license enables the user to download and install a copy of the
software (Windows or Mac) onto two different machines. Once the license is purchased and the
software installed you sign into the software to activate the license. Only one of the two machines
can run the software at a time. If you need to activate the software on a third machine, one of the
existing installations will need to be de-activated before the new machine can be licensed.

This license also enables you to install LightRoom and Bridge on your machines as well. I have
not had much experience with using either LightRoom or Bridge, so cannot really make any
comments on how useful these two pieces of software are, although people I have spoken to rave
about them.

Photoshop Plug-Ins and Actions.
There are quite many Plug-Ins and Actions available for Photoshop.
The following Plug-ins and Actions I would heartily recommend. The first two I consider
essential:

GradientXterminator. Paid ($49:95 US).
Reduces background light gradients. This is almost essential for processing images taken from my
light polluted location.
http://www.rc-astro.com/resources/GradientXTerminator
Comes in 32 and 64 bit editions. You will need to purchase the correct one for your version of
Photoshop.

Deep Sky Colours. Free (Donations Welcome).
Reduces the amount of green in a DSLR Image.
http://www.deepskycolors.com/archivo/2010/04/26/hasta-La-Vista-Green.html

Photoshop Astronomy Tools. Paid ($21:95 US)
http://www.prodigitalsoftware.com/Astronomy_Tools_For_Full_Version.html

Annie’s Astro Actions. Paid ($15:00 US).
http://www.eprisephoto.com/astro-actions

Peter Shah’s Astronomy Actions. Free.

(Requires Stargazers Lounge Account).
https://stargazerslounge.com/applications/core/interface/file/attachment.php?id=11811

Google Nik Collection (Free). Not specifically designed for astronomical images, but these tools
have now been made freely available. Can dramatically improve the appearance of many types of
images.
https://www.google.com/nikcollection
DSLR Imaging Basics.
Digital Single Lens Reflex Cameras (DSLR’s) are fantastic pieces of equipment. They enable you
to take colour images under a wide variety of conditions.

The way these devices are manufactured, does tend to give many inherent problems when used for
astrophotography.

It may come as a surprise to a lot of people that every single camera in production uses a
monochrome chip to record the image. This chip is composed of several light-sensitive pixels.
The greater the number of pixels, the higher the resolution of the image.

1 megapixel = 1 million pixels.
12 megapixels = 12 million pixels.
24 megapixels = 24 million pixels.

The image produced by a 24-megapixel chip will be much bigger file size, and take up much more
disk space than a 1 megapixel image.

While an exposure is being made, each pixel records photons that hit their surface and add these
counts. The more photons that hit the chip, the greater the value assigned to that pixel. A pixel that
receives a small number of hits will record a much lower value.

At the end of the exposure, each pixel will have its own count. These values are read off the chip.

As the camera downloads the resulting image, a level of grey is assigned to each pixel. This can
range from completely black, where no photons have been detected, to completely white, where a
pixel has been reached its maximum possible value.

On many Web sites, you may have seen a chart like this so that you can set your monitor to view
images correctly and can see differences between all levels of grey.

Complete black is on the left-hand side this will be true if any pixel on the chip receives no
photons. On the right-hand side is a pixel that has received a lot of pixels. A pixel recording pure
white is completely saturated. There will be many pixels in between that have received a number
if photons somewhere between the black and white level. The number of levels of grey shown
between black and white in this example has only 14 steps.

Think of Pixels as Light Buckets
The term for a big telescope, such as a large reflector are affectionately known as light buckets.

I like to think of pixels as light buckets as well. When an image is being taken, each pixel will
collect photons and record the number of hits. An empty pixel will be black, a full pixel is pure
white. Every other pixel in-between will be a different level of grey, depending on the number of
photons detected by each pixel.

If we look at the histogram of an image it tells us a lot about the number of pixels at each level of
grey.

Along the lower axis, the black point is on the left-hand side. The white point over on the right,
with varying levels of grey between each one.

The number of levels of grey that can be assigned to each pixel in the camera, will be determined
by the bit depth of the imaging chip.
Different cameras will have different sensors installed, so the number of levels of grey produced
in the resulting image is much more than the 14 shown above.

Below are the levels of grey that can be assigned by different bit level cameras:

12-bit image = 4,096 grey levels (212).
16-bit image = 65,536 grey levels (216).
24-bit image 16,777,216 grey levels (224).

As with the number of pixels, an image with a greater bit depth will produce a bigger file.

The peaks that can be seen in the histogram shows the number of pixels that are at each level of
grey. The more pixels with this value, the higher the peak appears. When taking images, we want
to avoid under-exposing (empty buckets) or over-exposing (full buckets) our images.

Producing Colour images.

As mentioned, a camera chip will only produce a monochrome image composed of different
levels of grey. For a camera to produce a colour image, the pixels need to be fed some colour
information. This is achieved by placing a layer of coloured filters on top of the imaging chip.
This layer, called the Bayer Matrix, is designed so that a different coloured filter sits above each
pixel on the chip. The Bayer Matrix has three different coloured filters, Red, Green and Blue. The
camera knows which pixel can see which colour, and produces a coloured image.

A different coloured filter lies above each pixel. A typical layout of a Bayer Matrix is shown
below.

When the Bayer Matrix is in place if a blue filter lies above a pixel, only light near the blue end
of the spectrum will reach it. The same goes for pixels under the red and green filters.

This is where the inherent problems with using a DSLR start to rear their ugly head.

This means that we are starting to stop quite a bit of light from reaching our imaging chip, as
wavelengths away from the filters preferred wavelength never reach each pixel.

This will effectively make the image much dimmer, so a much longer exposure is required to get a
reasonably bright image.

Most astrophotography (apart from the Moon, Sun and bright planets) is performed using very dim
light sources, so the fact that a lot of light never reaches the chip already inhibits the cameras
performance.

Correcting Colour Shift.

Added to this problem, the way the Bayer Matrix is set up also causes another issue. As you may
be able to see from the diagram above, there are double the number of green-sensitive pixels than
red or blue. This is because the human eye is more sensitive to the middle of the visible
electromagnetic spectrum. So, allowing more pixels to be green sensitive, gives a nice balanced
colour in the resulting image close to what we see with our naked eye.

So, images usually have a greener cast that they should have. In many images, this green colour
shift may need to be removed in post-processing. More of which later.

Another loss of light at the red end of the spectrum is caused by yet another filter fitted above the
imaging chip. This is the IR cut-off filter. Most imaging chips are very sensitive to red light.
Images taken without the IR filter will have a strange red colour cast. The IR cut-off filter
balances this colour, making the chip less sensitive at the red end of the spectrum, so the resulting
image looks how it would appear to the human eye.

Many astronomical objects, especially ionised Ha nebulae, emit light down that end of the
spectrum, so an off-the-peg camera needs to make quite long exposures to record them. Cameras
can be modified, where the IR filter is removed (modded), to make them more sensitive to this
end of the spectrum. However, there will be a strange colour shift if the camera is also used for
daytime photography.

All the images I have taken on my Web site and Flickr site have been taken using cameras without
any modifications. I use my cameras for all sorts of photography, so do not want to modify them,
which would severely restrict their use for normal photography.

Taking and stacking images.
It is a fact of life, astrophotography, in most cases, is trying to record objects that are very faint.
As well as the objects sending us only a small amount of light, the average DSLR camera as
discussed above is conspiring to restrict what we can record.

Thankfully, there are many techniques that we can use to enhance our images and reveal the true
nature of the objects they can successfully manage.

Data is key. The more data (light) you can collect, the better your image will be.

Unfortunately, it is a fact of life that most astronomical images will almost always be very under-
exposed.

To get enough data for processing, we will need to do extremely long exposures, or stack many
images to add data together.

Deep Sky Stacker is perfect for this last purpose, and it’s free:
http://deepskystacker.free.fr/english/index.html

My first astrophotography guide on Deep Sky Stacker will take you through the process for taking
suitable images and stacking images of both Deep Sky Objects and Comets, so I am not going to
go into any more details here.

Where is my object in the stacked image?
A typical stacked image will show an image like that shown below. This image was stacked from
49x 1 minute subs, and shows the stars in The Rosette Nebula.

So where is the nebulosity? It’s definitely in the image. If we look at the histogram we can see that
the data is hidden away down towards the left-hand side black end of the histogram. It’s that very
thin peak almost up against the black point.

As most of our data is, as expected, well down towards the black point. This makes it very
difficult to see the faint nebulosity.

The nebulosity can be revealed several ways. Go out and do even longer exposures, stacking
even more images, or by tweaking the image we have to reveal the hidden data.

Thankfully Photoshop has several tools we can use to tweak the image and move this peak
towards the white end. This will help to reveal the nebula.

What follows is my somewhat rough and ready processing workflow.

It is designed to show you how a reasonable image can be produced using these Photoshop
techniques.

It’s a reasonably easy technique to master, but needs to be performed in an orderly fashion. Once
you become familiar with this, it should enable you to become much more confident in trying out
other tools and techniques and really develop (pun entirely intended) your image processing
skills.

Never ever save files as JPG’s. These compress data and will lose data.
Save all your images as TIF files.

As you go through the processing I would recommend that you save your images as a different file
name as you go.

This will ensure that you never overwrite your original stacked file. If a different file is saved at
each stage of the process, if something goes terribly wrong later, you can pick up the appropriate
saved file and continue from there. If you save a file with layers, the file size increases
enormously.

Yes, I do know that the processing history saved in Photoshop can also get you out of a tight spot,
but this doesn’t help if you’ve closed the file and re-opened it to work on it later.

So, let’s get down to some image processing.
Curve Adjustments.
Open your previously stacked image into Photoshop.

The first step in my imaging workflow is to adjust the curves.

Here is what my original stacked Rosette Nebula image looks like. This was taken with a Nikon
D750 camera and is stacked from 49x 1 minute subs through a 190mm Mak-Newt.

Despite this being almost an hours integration time, only the stars of the cluster within and around
the nebula can be seen, but no nebulosity. The image needs quite a lot of post processing to reveal
the faint nebula.

There will be some artefacts around the edge of the image from the stacking process. This is
unavoidable. Use the crop tool to remove the edges of the image so they do not affect the results
later.

Once the image is cropped, from the menu select, Image, Adjustments, Curves…

If this selection is greyed out you will need to convert to a 16-bit image by selecting, Image,
Mode, 16 Bits/Channel before the previous step.


Once the Curves Tool is open, the histogram will be visible.

The diagonal line running from bottom left (Black Point) to top right (White Point) on the
histogram is the curve adjustment.

We will use this to stretch the image.
Click anywhere on the diagonal line and move the point to the position shown below:

We haven’t finished adjusting the curves yet so don’t click OK, as your image will have been
over adjusted by doing this as shown below.


Leave the curves tool open and look at the new adjustment curve.
The right-hand side of this line now hits the top of the histogram window very early. This is
extremely bad! This means everything above that level of grey will be treated as white, data
above that level of grey will have been lost. Despite not being able to see any peaks over on the
right-hand side, there are some there, like stars. This data will never be able to be recovered. We
need to make sure we don’t affect those pixels.

Click again on the curve to add a couple more points onto the line. Adjust the position of these
points so it looks like the curve shown below. When you are happy with the curve shape click
OK.

On the left-hand side where the curve is steep, we are highly multiplying the values of those
pixels at those darker levels of grey. What this means is that the faint stuff down the black end
will increase in brightness and move towards the right.

Over on the right-hand side where the slope is more level, in the brighter area, we are hardly
multiplying these pixels at all. The values of the pixels where we have stars and other bright
objects have hardly changed at all.
Your image should now look something like this.

Don’t worry about the green colouration, this is caused by the colour channels not being correctly
registered and some light pollution. We will get rid of this in the next step.

Open the Levels Tool.
This is available from the menu: Image, Adjust, Levels…

Looking at the histogram in the Levels Tool, we can see that more than one peak is visible in the
histogram. These should be aligned as a single peak. This means that one of the colour channels
has lower values than the other two. This gives the colour shift that can be seen. In this case, it is
the red channel. Remember the IR-cut filter mentioned earlier?

Adjusting Colour Channels.
Method 1. Levels Tool.
The way to adjust the curves so that each colour channel is aligned is to select each channel
individually. Below the Preset: Default setting can be seen the Channel. The default is set to
RGB.

Below the Preset: Default, setting can be seen another drop-down menu, Channel. The default is
set to RGB. Use this drop down menu to select each channel in turn to identify which peak
corresponds to which channel.

This is the red channel:

This is the Green Channel:

This is the Blue Channel:

The peaks of the Green and Blue channels are a bit further to the right in the histogram than the
Red. Not surprisingly, the red channel is fainter than the other two channels. To align the peaks in
the three channels, the black point needs to be moved in all three colour channels. To do this,
move the black point in all three, so it is in a similar position to their histogram peaks. To move
the black point, grab the black triangle under the left-hand side of the histogram and drag it to the
right.

Move each black point and place it a little way just before the peak starts rising. Repeat for each
colour channel.

Red Adjusted:


The colour of your image will look a bit strange as you make these adjustments in turn, so don’t
worry about the colour until you have adjusted all three channels.

Green Adjusted:

Blue Adjusted:

Click OK and the colour should now be reasonably balanced and your image should now look
like that shown below.


It might be worthwhile looking in the Levels tool again.

In this case, although the peaks are very close, they can still be seen diverting slightly towards the
top of the peaks. Do a similar Levels Adjustment as before to get the peaks aligned as perfectly as
possible.

Once the colour channels have been correctly adjusted the colour shift should now be properly
removed. Your image should now look something like this.

Method 2. Gradient Xterminator Plug-In.
An alternative way of doing this step is to use the plug-in Gradient Xterminator. You can do it the
hard way, but this utility is well worth the money as it saves you a lot of hard work and produces
much, much better results, enabling you to get more out of your image.

It is also more flexible that it can be told to ignore bright objects in the field of view, but this is
more difficult with an image like this where the object like the Rosette Nebula is wider that the
field of view. But I’m not going to go into more detail about that here.

Once the plug-in is installed in Photoshop, from the menu select Filter, RC-Astro, Gradient
Xterminator…

I generally use these settings:

Click OK to run the plug-in.

The image will lose its green colouration and now have a grey background.

Next step is to remove this grey background.
From the menu select Image, Adjust, Levels

The histogram will once again be displayed and should look something like this. Gradient
Xterminator has aligned the three peaks perfectly.

The histogram peak is about two thirds of the way up the grey level scale. There are markers for
the black point, mid-point and white point underneath.

Grab hold of the black point over on the left-hand side with the mouse and drag it to a position
just before the histogram starts to rise.


The image should now look something like this.

Getting the nebulosity out.
In the last few steps we have removed the background, and adjusted all channels so they are
evenly balanced. The image I am using for the rest of this process was balanced using Gradient
Xterminator.

Now we need to work on the image to reveal the faint nebulosity hidden within it.

Go back into the Curves tool. It should now look something like this.

As before, add a few extra points onto the diagonal line to adjust the curve. This time the
adjustment will be much subtler than the previous adjustment that was made. Pay close attention
to the lower part of the curve.

If the image is stretched too much, we start to lose data and our image will rapidly degrade. The
more data you had in your original image, the further you will be able to stretch it.

The image should now look something like this. Some of the faint nebulosity is now starting to be
revealed.


Go back into the Curves Adjustment and give an even more subtle adjustment.

The image should now look something like this and the surrounding nebulosity is even brighter.

A look at the histogram in the Curve Tool now looks like this.

The histogram peak is now nicely central but if we try to stretch this image any further it will start
degrading rapidly. We have stretched this image almost as far as we can before we start to
destroy it.

The image still looks a little wishy washy, so next step is to fix that.

Make sure that your Layer Window is visible by selecting it from Windows, Layers.

From the menu, go to Layer, Duplicate Layer.
Call this new layer whatever you like (I called mine Luminosity Layer) and click OK.

Make sure that you click the mouse on the new upper Luminosity layer to select it.

Above the layer list there is a drop down menu which is at the default setting of Normal. From the
bottom of the drop down menu, select Luminosity.


Now click on the lower original Background layer to select it.

From the Menu, select, Image, Adjustments, Hue/Saturation…

Move the middle Saturation adjustment to no more that about 40%.


The colour of the nebulosity should now start to become more vivid.

If you have enough data in your image, it may withstand another step or two of increasing the
colour saturation. Watch the image as you adjust to make sure you do not overdo this step.

You will probably notice that the quality of the picture has degraded a small amount since the
colour saturation was increased. The image will take on a bit more noise and look grainier. Any
adjustment made can degrade the image. We never get anything for nothing.

But we can cope with a bit of degradation if you don’t take things too far.

To restore some of the original quality we will use the upper luminosity layer.

Keep the lower layer selected and from the menu go to, Filter, Blur, Gaussian Blur…

Move the adjustment to the right. This will blur the image in the lower layer.

This should improve the appearance of the overall image and remove some of the graininess that
was just created.

As this layer tells the image what level of grey is assigned to each pixel, and the lower pixels are
blurred, the luminosity layer is now dominating.

Next another curve adjusted needs to be made.

Select the upper Luminosity layer again and from the menu click,
Image, Adjustments, Curves…

I normally put in a small amount of adjustment, bringing the darker end of the curve down a little
and increasing the upper end of the peak a little.


This adjustment is very subtle, but it adds quite a punch to the image.

Depending on the quality of your image, you may want to make a few more small saturation and
curve adjustments to get the best out of your image.

Removing the Green Bias.
Remember the Bayer Matrix gave the camera a bias towards the Green part of the light spectrum?
This is where the plug-in Deep Sky Colours now comes in.

If the plug-in has been copied to the correct folder, it is selectable from the menu.

Click, Filter, DeepSkyColours, HLVG… And click OK.


These are the settings I normally use. Click OK.

The green bias should now be removed.
This is the final image.

Considering our original stacked image looked like this, what a transformation.

I hope that this guide has enabled you to become more familiar with the tools used in Photoshop
and has helped you get the best out of your images.


If I have made any mistakes in laying out this guide, please let me know so I can improve future
editions.


I am also open to correspondence and love to hear from fellow enthusiasts, so please send me a
message (A kind one that is!).


dave@eagleseye.co.uk



Please follow me on:

My blog: www.eagleseye.me.uk

Facebook: https://www.facebook.com/Eagleseye45

Twitter: https://twitter.com/DaveEagle45

Flickr: https://www.flickr.com/photos/eagleseyeonthesky


I have some other astrophotography guides planned, based on my successful astrophotography and
webcam imaging workshops, so keep a look out for those as they are published.

Keep Looking Up.

Dave Eagle.

January 2017.