Deep Sky Astrophotography with a dSLR

Deep Sky Astrophotography
If you're willing to devote an entire night — and sometimes multiple nights — to one subject, thousands of objects become viable targets for digital SLR cameras. Results such as this exquisite picture of IC 4604 in Ophiuchus, which represents more than eight hours of exposure, can be achieved with these versatile cameras.
Chuck Vaughn

In early 2004 Kodak announced it had stopped producing
Technical Pan 2415 black-and-white film, widely regarded
as the best film for astrophotography. Many of us
were left with a choice — either stockpile whatever rolls
of Tech Pan we could find, or make the transition into the
digital age. At the time I wasn’t ready to invest large sums
of money on a CCD camera and the required accessories.
While I mulled over these choices, Canon announced an
astrophotography version of their popular EOS 20D digital
SLR camera. After much contemplation, I decided that
this EOS 20Da (S&T: November 2005, page 84) was the
best choice for me, and I have never looked back.
The EOS 20Da was affordable and required virtually no
modifi cations to my existing equipment. Only a T-adapter
was necessary to continue my photographic endeavors,
along with two computer programs to calibrate and combine
the RAW format images produced by the camera.

Deep Sky Astrophotography: Getting Started with Digital

After initial tests with the EOS 20Da, I replaced the
camera’s internal infrared-blocking filter with a Hutech
Type 1 filter
to improve its sensitivity to hydrogen-alpha light. Also, just like my experience with film astrophotography, I found that the
longer the total exposure time with the digital camera,
the better the final image. I consider 3 hours to be
my minimum duration per subject when imaging under
dark skies. As with cooled CCD cameras, creating long
cumulative exposures is best done with many shorter
ones combined; due to the greater sensitivity of digital
sensors, they reach the skyfog limit far quicker than film.

Digital single-lens reflex (SLR) cameras advanced so much in recent years that the quality of astrophotos captured with them now exceeds what was possible using film. Using a Hutech-modified Canon EOS 20Da, the author combines dozens of 10-minute expsosures recorded at ISO 800 to create ethereal masterpieces such as this portrait of Van den Bergh 142, port of the larger nebula complex IC 1396 in Cepheus with 420 minutes of exposure.
Chuck Vaughn

From my location, I’ve standardized my exposure
length to 10 minutes per frame at ISO 800. Lower ISO
values cause undesirable posterization (visible steps
between brightness levels) in the faintest areas of my images,
and I don’t see much advantage to ISO 1600, with
its decreased dynamic range. I save all my images in RAW
format, to preserve the entire 12 bits of data produced by
the camera. I also turn off the automatic noise reduction
and instead record dark exposures (images of the same duration
as the light frames, recorded with the camera’s lens
covered) at roughly the same ambient temperature so I can
calibrate my images later. Of course, accurate guiding of
the telescope is a must for any image longer than a few
seconds, so I use an old SBIG ST-4 autoguider to ensure
round stars every time.

Processing the Raw Image

I used Adobe Photoshop to process my first images with the
20Da but quickly discovered that it isn’t a very effective
tool for performing critical steps such as the calibration
of individual frames. Also, the plug-in that was included
with the Canon software package for converting RAW images
to other formats wasn’t as effective as I had hoped,
especially in handling color balance. So I purchased the
program ImagesPlus (reviewed in the
July 2006 issue, page 82), which was specifically created
to convert, calibrate, and process digital SLR camera files.
(Cyanogen’s MaxDSLR performs
many of the same functions.) I now use ImagesPlus for
RAW file conversion, dark- and flat-field calibration, and
combining the individual images.

I try to record at least as many dark frames per night as
light frames and then combine them using ImagesPlus to
create a master dark frame, which averages any temperature
changes and cosmic-ray hits in the individual darks.

I also use a master bias frame, which is a zero-length
exposure that records the inherent electronic signal in
every exposure. Bias frames are used when dark frames
need to be scaled. While I don’t take new flat-field images
every night because my instruments are optically very
stable, users of telescopes with moving mirrors that can
shift over the course of the night should consider recording
new flat-field images each imaging session. Finally,
I’m very careful to clean dust off the filter in the camera
each night. Shadows from any remaining dust motes I can
remove in Photoshop later.

After a night of capturing light exposures and calibration
frames, I convert the RAW files to 16-bit FITS format
in ImagesPlus using no white balance. In the Calibration
Setup menu, I select Auto in the Scale Factors section to
compensate for any remaining differences in my darks
due to temperature changes. (All digital SLR cameras lack
temperature regulation, so differences from one image
to the next are inevitable.) Once all the calibration is performed,
I align the frames and combine them using Adaptive
Addition, which is similar to performing a sum of the
exposures with much of the random noise averaged out.
With the exception of noise reduction, I do all other processing
with Adobe Photoshop.

After the individual frames are calibrated and stacked, the white balance must be adjusted. Using the Levels palette in Adobe Photoshop, Vaughn adjusts the white (right) slider on the green and blue channels to set the white balance of a G2V star similar to our Sun. (Click for larger image.)
Chuck Vaughn

The images from my EOS 20Da are very red because of
the wider red passband of the Hutech filter. To compensate
for this, the image needs to be white balanced. White
balancing simply means that if an image of a gray card is
recorded with a camera in RGB format, the individual red,
green, and blue channels are adjusted until the image appears
gray. The standard technique in astrophotography to
achieve white balance is to photograph a G2V (solar-type) star and adjust the green and blue channels until the star
appears white. My G2V test on an overhead star achieved
proper white balance when the green channel’s white
slider in Photoshop’s Levels palette was set to 133 and the
blue channel’s was set to 126. While other white-balance
calibrations can be used, anything other than G2V calibration
will change the color balance compared to daylight
photos.

The next step is to perform a gamma adjustment. Using
the Levels palette, I move the center slider until it is just
to the right of the primary data within the histogram. This adjustment starts to bring out faint nebulae and galaxies without
saturating the already-bright stars in the image.

Establishing a neutral sky background is an important step in all image processing. Set the Color Selection tool to 5 X 5 average and place it on a region in the background where no faint stars, nebulosity, or galaxy detail is apparent. Image adjustments can be monitored in the Window > Show Info drop-down menu. Using the Levels palette black (left) slider, adjust the black point until the red, green, and blue channels are equalized, but don't make the background completely dark at this stage. (Click for larger image.)
Chuck Vaughn

After the gamma adjustment, I now adjust the sky background.
It’s necessary to find an area of neutral sky in the
image that’s free of any faint detail I wish to preserve in
the final image. Images of nebulae that permeate the entire
field can sometimes complicate this step. To monitor
any changes I perform on the background, I first open the
drop-down menu Window > Show Info (which displays
the pixel values of the point under my cursor) and set it
to display both RGB and Grayscale readings. I also select
the Color Sampler tool and change the sample size from
point sample to 5 × 5 average. I then click on the image in
my chosen neutral area, open the Levels palette, and adjust
the black (left) sliders of the green and blue channels
until they equal the red levels.

Approaching the Curve

Now comes the first really subjective part of my processing.
It’s obvious that much of the information I wish
to display resides in a small area of the histogram near
the faint end. Using the Curves function, I perform the
equivalent of a hyperbolic arc-sine adjustment based on
the image data. I first open the Histogram palette and
determine the values that are slightly darker and lighter
than the bulk of the histogram data. I then apply a steep
linear slope to the curve between these levels. I use a
gentler slope for the curve above and below this mountain of data, which prevents the faintest
data from being clipped to black and
the brightest highlights from clipping to
white.

Using the Curves palette, you can now stretch the image, though carefully monitoring of the white and black points is still required to avoid clipping the ends of the histogram. Often more than one application of the Curves function is necessary to achieve satisfactory results. This particular image required the two curves presented above, plus one additional, though less aggressive, application of the function to establish the proper contrast throughout the image. (Click for larger image.)
Chuck Vaughn

While this adjustment improves the
image greatly, my subject still appears
faint, and the background is still too
bright. I need to apply at least one additional
curve to increase the contrast. I open the Curves palette again and
place additional points to hold the linear
part of my curve straight while simultaneously
keeping the shadows and
highlights from clipping. Exactly how I
adjust this curve depends on what it’s
doing to the image on the screen — I try
to achieve good contrast in the object, a
dark gray sky, and highlights that are not
burned out. I may try several curves until
I find one to my liking.

One rule I apply to all my images during
processing is that objects must fade
into the background sky rather than have
razor-sharp cutoffs. It’s okay for regions
of the subject to be faint. Typically I
leave the sky background in the range of
20 to 30 out of the 255 levels displayed in
the Histogram window. Another, weaker
Curves adjustment might be necessary to
achieve the desired effect.

Finishing Touches

My experience with my modified Canon
EOS 20Da shows that it is necessary to
increase the color saturation in Photoshop
by +20 for daylight images, so I automatically
apply this adjustment to all my astrophotographs.
Next I perform a final color balance
for the sky background. I examine a few
areas of blank sky (if available) and use
the Levels palette to neutralize them.

After sharpening and noise reduction, the final image should appear smooth but not quite noise free. This view of IC 5070, the Pelican Nebula, combines 24 five-minute exposures recorded with an Astro-Physics 155EDF (6.1 inch) refractor operating at f/6.
Kristina Grifantini

At this point I apply any sharpening I
think will be beneficial. While there are
many different ways to sharpen an image,
mostly I prefer to use the Lasso tool with
some feathering to select an area I wish
to sharpen, and then apply the Unsharp
Mask filter with a radius of 1 to 2 pixels.
I rarely, if ever, sharpen the entire image.

My final step is to minimize noise,
which shows up in virtually every astrophotograph.
Here again we are fortunate
to have many choices, though I prefer
Noise Ninja. It’s
easy to use and does a good job. Although
the program works on RGB images,
I find I can achieve better results by
splitting the red, green, and blue channels
in Photoshop, saving them as individual
grayscale images, then using Noise
Ninja
to perform luminance noise reduction
on each channel separately. I feel
this gives me better control.

Some imagers try to achieve the
smoothest possible result. Human eyes
are accustomed to seeing detail in everything
down to the limit of our visual
resolution, but much of the time that “detail”
is noise where our brains fill in the
blanks. The trick is to have just the right
amount of noise. Too little makes the
image look plastic, and too much distracts
from the subject, so I attempt to process
my images to achieve a smooth — but not
completely noise-free — result.

Unfortunately, the 20Da is no longer
manufactured, but similar results can be
obtained with other modified cameras.
Regardless of which digital SLR camera
you choose for astrophotography, these
processing steps will help you get the
most out of your equipment. DSLR astrophotography
has come of age!


Chuck Vaughn has been photographing the universe from the Northern California mountains for nearly 20 years.