2007 – New Home, New Scope, New
Challenges
In May 2007 we
decided to sell our house in the ‘burbs and move to a
condo in downtown

The
above view shows the adjacent Bankruptcy Courthouse (could come in handy!),
central park, and adjacent hotel. In the distance is
another condominium in town. The sunsets can be quite spectacular:

As you can
tell, our condo faces West. I donated my old telescope to the Tallahassee
Astronomical Society, since I had no place to put it in our new home. Furthermore, I figured the light pollution of
downtown would pretty much prevent any astrophotography. Also, the FSU football stadium is directly
West and they keep high-powered sodium lights on 24/7 (regardless of whether it
is being used or not).
However, I
read an article where NASA was using video imaging from land-based telescopes
as a means to compensate for atmospheric distortion. Due to the nature of such distortion,
individual frames collected at 30 fps will generally be in focus, but longer
exposures will be fuzzy. So, the process
involves taking video and then aligning each frame and summing the images. This requires an extremely sensitive video
camera, fast lens (<f2.0), software for stacking, etc. Amateur astrophotography has been utilizing
this approach for planetary imaging. In
this case, the planets are quite bright and good images can be taken with
insensitive video cameras and slow lenses.
However, Watec has recently developed a
sensitive video camera for amateur astrophotography. Furthermore, one company (Starizona
Inc.) has developed a modification to consumer Schmidt-Cassegrain
telescopes that converts them from f10 to f1.8 lenses. The combination of the two could permit videoastrophotography at an affordable price. This technique also does not necessarily
require an accurate tracking mount. I
decided to try to see if this method would work, and started by purchasing the Watec video camera.
It needs a lens so I purchased a reasonably fast (f2.8) lens from ebay. Here is my new
“telescope”:

Pretty tiny compared
to my prior scope, but it will allow me to test this new method of
astrophotography. One of the first
problems was that the lens would not come to focus. The Watec camera
appears designed for telescope use, where there is much more range of focus. The lens is from 16mm video cameras and
assumes a precise focal plane distance. Anyway… I had to file down the adapter on the telescope to
get the lens in focus.
Next I took
several test videos, and this identified various other issues associated with
file size (huge – on the order of Gigabytes after only a minute of video),
video compression artifacts (lossless compression results in larger files but
is essential to astrophotography), file format compatibility with image
stacking software, and so on. I tried
various compression codecs, different capture
software, and different image stacking software, and found the following to be
best:
Capture
software: VirtualDub (www.virtualdub.org)
Frame
decimation software: VirtualDub
Image stacking
software: Registax (www.astronomie.be/registax/)
The Watec 120N+ (StellaCam3) video camera allows on-chip
integration of various multiples of frames (i.e. 1, 2, 4, 8, 16, etc.) all captured
at 30 fps. The raw video remains 30 fps,
thus, if the camera setting is to integrate 8 frames, for example, the video
does not output at 4 fps, but simply duplicates the last integrated image at 30
fps until the new one is ready. Thus,
the video (in the case of 8 frames being integrated) is 8 times larger than it
needs to be. Registax
has a size limitation of 10,000 frames in the loaded AVI file and this limits
the file size (exposure time). The VirtualDub software has a “decimation” feature where you
can specify how to sample a captured video file (in this case, decimate by a
factor of 8 to eliminate duplicate frames).
This worked pretty good. The VirtualDub
software also save video in a lossless AVI format that
can be read into the stacking software.
The following
is a single frame of video at 0.25 sec exposure (8 frames integration at 30
fps):

The following
is a stacked image composed of ~2400 frames:

This looks
really promising. One thing I found was
that at 100mm focal length (i.e. the focal length of the lens) the image moves
off-screen due to Earth’s rotation after about 5 minutes. Thus, while a mount is not necessary for a
sharp image, you still need to track the object to keep it in the imaging
frame. Thus, an equatorial mount is
still necessary from a practical standpoint.
I have not yet decided which f1.8 optical system to purchase (they are
expensive), I think a mount will be necessary.