| Planetary Photo Techniques (page 1) This page focuses on the collection of exposures and processing exposures from Atik usb 1 cameras |
I use the Takahashi Mewlon 250, modified DK cassegrain, as my planetary OTA. I'd like to
have the 300, but I'd prefer to leave a little money in the bank account (some of these
scope manufacturers seem to think that their mirrors are made of gold, and they want to
extract as much of it as possible, from us, to make more mirrors). A high end OTA helps, but
there are a lot of excellent planetary photos that have been taken through less expensive
scopes. The Meade and Celestron SCTs are very popular. Aperture counts, the resolution
helps, and you need enough light gathering to allow very short exposures, to approach the
resolution capability of your scope. An eight inch, or larger aperture, will do a fine job. A
few nights per year (for my location) a 12" aperture may yield noticeably better results than
my 10" cassegrain, even fewer nights a 14" scope would better do the job. It's really the
seeing that has the biggest impact on your planetary images.
This first page contains some general comments/suggestions for using the older USB1
webcams and associated capture programs. This page was written in 2005, editted in 2006.
The second page, begun in 2007, contains my observations to date, and some suggestions,
when imaging w/ a Lumenera Skynyx monochrome camera, LRGB filter set, and Lucam
Recorder. If you are using the Skynyx, take a look at the comments on this page, or go
directly to page 2 via the link at the top of the page.
General Comments
The focal length needed for pleasingly sized planetary images, and proper spacial sampling
of the scene, using most popular cameras (4-7um pixels) is 5000 to 10000mm. My mewlon
native focal length is 3000mm, I need help. To get the help, I use barlows (or Televue
Powermates). The two inch X2 Powermate provides for X2 increase regardless of the
extension between the powermate and the camera focal plane (chip). The two inch X2
Orion barlow multiplies more than X2 as the extension increases (I use extension to create
an effective X3 multiplier). The ATIK cameras that I use have 5.6um pixels. This requires
~7000mm focal length to sample blue light to the limit of my 10" scope (larger pixels require
longer FL, shorter wavelengths also require longer FL). Keep in mind that as the focal
length goes up on your fixed aperture OTA, your focal ratio also goes up (the lens slows
down). This means that you must expose longer to get the same intensity at the camera
focal plane. This means more apparent subject motion due to the atmosphere, mount
issues, whatever. There is a trade between the FL needed to sample the seeing disc (the
scope airy disc if seeing, and/or mount tracking errors are not the limiting factor) and the
desire to stop subject motion (due to the atmosphere and mount) during the exposure.
The OTA mount is important, but not as much so as with long exposure photography. As
long as it can hold the OTA, track the subject at the very long focal lengths required for
decent photos, it's good enough. Having said that, we require minimal tracking error during
the exposure, and want minimal movement during the exposure capture sequence. A
mount like the CGE or G11 generally fulfils this requirement with a 35-45lb OTA load, the
EQ6 mounts work well for 30-35lb loads, the EQ5 works well at 25-30lb. Be sure to balance
your scope/camera on the mount (slightly east/backend heavy is the norm). Setup your
mount's PEC, and use it during your photo work.
Choose an interesting scene for your photo. Some portions of Mars have a lot more
features than others. Jupiter's GRS/RS Jr. and its moons (especially with shadows) are
even more interesting than just the planet by itself. If you can catch a storm, or catch a
moon on Saturn, so much the better. Uranus and Nepture don't offer much more than a
pretty colored ball. Venus offers some cloud features. I don't have the foggiest notion what
Mercury may, or may not offer. Use the web to plan your photo sessions. Sky and Telescope
offer web pages that show when moons are going to be near Jupiter, or when the GRS is
transiting. They also have sites for Saturn's moons and Mars' surface.
Jupiter moon ref: SkyandTelescope.com/observing/objects/planets/article_830_2.asp
GRS ref: SkyandTelescope.com/observing/objects/planets/article_107_1.asp
Saturn moon: SkyandTelescope.com/observing/objects/planets/article_1136_2.asp
Mars surface features: SkyandTelescope.com/observing/objects/planets/article_997_1.asp
have the 300, but I'd prefer to leave a little money in the bank account (some of these
scope manufacturers seem to think that their mirrors are made of gold, and they want to
extract as much of it as possible, from us, to make more mirrors). A high end OTA helps, but
there are a lot of excellent planetary photos that have been taken through less expensive
scopes. The Meade and Celestron SCTs are very popular. Aperture counts, the resolution
helps, and you need enough light gathering to allow very short exposures, to approach the
resolution capability of your scope. An eight inch, or larger aperture, will do a fine job. A
few nights per year (for my location) a 12" aperture may yield noticeably better results than
my 10" cassegrain, even fewer nights a 14" scope would better do the job. It's really the
seeing that has the biggest impact on your planetary images.
This first page contains some general comments/suggestions for using the older USB1
webcams and associated capture programs. This page was written in 2005, editted in 2006.
The second page, begun in 2007, contains my observations to date, and some suggestions,
when imaging w/ a Lumenera Skynyx monochrome camera, LRGB filter set, and Lucam
Recorder. If you are using the Skynyx, take a look at the comments on this page, or go
directly to page 2 via the link at the top of the page.
General Comments
The focal length needed for pleasingly sized planetary images, and proper spacial sampling
of the scene, using most popular cameras (4-7um pixels) is 5000 to 10000mm. My mewlon
native focal length is 3000mm, I need help. To get the help, I use barlows (or Televue
Powermates). The two inch X2 Powermate provides for X2 increase regardless of the
extension between the powermate and the camera focal plane (chip). The two inch X2
Orion barlow multiplies more than X2 as the extension increases (I use extension to create
an effective X3 multiplier). The ATIK cameras that I use have 5.6um pixels. This requires
~7000mm focal length to sample blue light to the limit of my 10" scope (larger pixels require
longer FL, shorter wavelengths also require longer FL). Keep in mind that as the focal
length goes up on your fixed aperture OTA, your focal ratio also goes up (the lens slows
down). This means that you must expose longer to get the same intensity at the camera
focal plane. This means more apparent subject motion due to the atmosphere, mount
issues, whatever. There is a trade between the FL needed to sample the seeing disc (the
scope airy disc if seeing, and/or mount tracking errors are not the limiting factor) and the
desire to stop subject motion (due to the atmosphere and mount) during the exposure.
The OTA mount is important, but not as much so as with long exposure photography. As
long as it can hold the OTA, track the subject at the very long focal lengths required for
decent photos, it's good enough. Having said that, we require minimal tracking error during
the exposure, and want minimal movement during the exposure capture sequence. A
mount like the CGE or G11 generally fulfils this requirement with a 35-45lb OTA load, the
EQ6 mounts work well for 30-35lb loads, the EQ5 works well at 25-30lb. Be sure to balance
your scope/camera on the mount (slightly east/backend heavy is the norm). Setup your
mount's PEC, and use it during your photo work.
Choose an interesting scene for your photo. Some portions of Mars have a lot more
features than others. Jupiter's GRS/RS Jr. and its moons (especially with shadows) are
even more interesting than just the planet by itself. If you can catch a storm, or catch a
moon on Saturn, so much the better. Uranus and Nepture don't offer much more than a
pretty colored ball. Venus offers some cloud features. I don't have the foggiest notion what
Mercury may, or may not offer. Use the web to plan your photo sessions. Sky and Telescope
offer web pages that show when moons are going to be near Jupiter, or when the GRS is
transiting. They also have sites for Saturn's moons and Mars' surface.
Jupiter moon ref: SkyandTelescope.com/observing/objects/planets/article_830_2.asp
GRS ref: SkyandTelescope.com/observing/objects/planets/article_107_1.asp
Saturn moon: SkyandTelescope.com/observing/objects/planets/article_1136_2.asp
Mars surface features: SkyandTelescope.com/observing/objects/planets/article_997_1.asp
Post Capture Processing
Now that you have the AVI sequence, you'll need to stack individual images in the
sequence/s. I use Registax for this. There are many pages on the web, and in the program
instructions, outlining the operation, so I'll be brief. The first step is aligning, select your
AVI sequence, set for LRGB processing, set the box to fully surround the subject (with
ample margin for subject movement), select a high contrast alignment point (the ice cap on
Mars, or a nearby moon on Jupiter), select local contrast or gradient, set at 90% (if you don't
get enough aligned frames, the seeing wasn't good enough to fool around with the
sequence anyway, if the frames all look good, use 95%), increase the order of the FFT filter
until the FFT shows little change (typically 15 to 20), hit the align button. Go to optimize, use
default settings initially, after the first pass, note error and set to 1 pixel search, repeat
optimize, select best search, reoptimize. Go to stack, set LRGB (if processing a color
image), bring the stackgraph out, use the sliders to limit the error (vertical) and frame
quality (horizontal), this is equivalent to the editing phase in photography, be brutal. If your
sequence isn't any good, try again on another one. Stack, then go to the wavelet page. You
will be selectively sharpening the image here. You can use deconvolution (as in AIP or
MaximDL) in combination with the wavelet processing (I do). Your best bet is to play around
with the wavelet (selective FFT) filtering. Don't sharpen the noise, and don't let the halo
form around the subject, don't get carried away with the weights, keep them relatively low,
generally below 10 or 12 (if the image noise is low, you can increase some of these), keep
the weight of the first and last layers lower than the others, you may need to turn off these
two layers. Guassian processing can be used advantageously, but start with the default.
Look at the image, if it is beginning to look over processed, ... it probably is, back off the
weights, turn on/off filtering layers. Once you're satisfied that the image is as good as it
gets, go to Final. Play with the advanced HSL if you wish, or just save the image.
If you use two sets of sequences, one for lum, the other for RGB, once you finish with
Registax, you will need to stack the component images to form the final image. I use
Photoshop and/or MaximDL for this. Before you worry about aligning the lum and RGB
panes, be sure to correct for differential refraction (if the planet was below 45 or 50
degrees altitude). MaximDL offers a function to move the alignment of the R, G, B
components relative to each other in the RGB image. You can also use Photoshop to do
the same on the components of the RGB image. The proper way to accomplish the diff
refrac correction is at the time of image capture. You should be using a dispersion
corrector, such as a Risley prism. These units should be used whenever you are working at
or below 45 degree elevation (you will see just over 0.5" of blue/red shift re green at 45
deg). Prism devices are available from imaging retailers.
Once you have the RGB image as you want it, bring both lum and RGB into Photoshop. Be
sure the images are equal size. Process both the lum and RGB as desired (sharpen the
lum, adjust curves on both, color adjust the RGB if needed). Select the lum image, copy and
paste on the RGB image. They don't align...too bad (I use MaximDL to get the alignment, but
you can also use Photoshop). Set the lum layer that was copied to the RGB as a difference
layer, use Edit>Free Transform to translate and rotate that layer as required to align with
the RGB. There are many ways to combine the lum and RGB frames. One method is to set
the lum layer to soft light, and adjust the Opacity, typically between 50 to 80%. Another
approach to combine the frames is to set the lum frame as a normal layer (1st layer on the
stack to be formed), duplicate the RGB frame, desaturate (image adjust > desaturate), then
select all, copy, and paste it as a muliply layer on the (normal) lum layer. Select, copy, and
paste the RGB on the layer stack as a color layer. Adjust the opacity of the desaturated RGB
layer (middle layer), typically 40-60%. Flatten the stack, then readjust levels/curves as
desired. Crop as needed, and you are through with the image. Now go back and try another
set of settings on earlier steps with the same sequence/s to build a second, and even third
image, select the best....sometimes you get lucky, sometimes not, if not, try again another
night.
Now that you have the AVI sequence, you'll need to stack individual images in the
sequence/s. I use Registax for this. There are many pages on the web, and in the program
instructions, outlining the operation, so I'll be brief. The first step is aligning, select your
AVI sequence, set for LRGB processing, set the box to fully surround the subject (with
ample margin for subject movement), select a high contrast alignment point (the ice cap on
Mars, or a nearby moon on Jupiter), select local contrast or gradient, set at 90% (if you don't
get enough aligned frames, the seeing wasn't good enough to fool around with the
sequence anyway, if the frames all look good, use 95%), increase the order of the FFT filter
until the FFT shows little change (typically 15 to 20), hit the align button. Go to optimize, use
default settings initially, after the first pass, note error and set to 1 pixel search, repeat
optimize, select best search, reoptimize. Go to stack, set LRGB (if processing a color
image), bring the stackgraph out, use the sliders to limit the error (vertical) and frame
quality (horizontal), this is equivalent to the editing phase in photography, be brutal. If your
sequence isn't any good, try again on another one. Stack, then go to the wavelet page. You
will be selectively sharpening the image here. You can use deconvolution (as in AIP or
MaximDL) in combination with the wavelet processing (I do). Your best bet is to play around
with the wavelet (selective FFT) filtering. Don't sharpen the noise, and don't let the halo
form around the subject, don't get carried away with the weights, keep them relatively low,
generally below 10 or 12 (if the image noise is low, you can increase some of these), keep
the weight of the first and last layers lower than the others, you may need to turn off these
two layers. Guassian processing can be used advantageously, but start with the default.
Look at the image, if it is beginning to look over processed, ... it probably is, back off the
weights, turn on/off filtering layers. Once you're satisfied that the image is as good as it
gets, go to Final. Play with the advanced HSL if you wish, or just save the image.
If you use two sets of sequences, one for lum, the other for RGB, once you finish with
Registax, you will need to stack the component images to form the final image. I use
Photoshop and/or MaximDL for this. Before you worry about aligning the lum and RGB
panes, be sure to correct for differential refraction (if the planet was below 45 or 50
degrees altitude). MaximDL offers a function to move the alignment of the R, G, B
components relative to each other in the RGB image. You can also use Photoshop to do
the same on the components of the RGB image. The proper way to accomplish the diff
refrac correction is at the time of image capture. You should be using a dispersion
corrector, such as a Risley prism. These units should be used whenever you are working at
or below 45 degree elevation (you will see just over 0.5" of blue/red shift re green at 45
deg). Prism devices are available from imaging retailers.
Once you have the RGB image as you want it, bring both lum and RGB into Photoshop. Be
sure the images are equal size. Process both the lum and RGB as desired (sharpen the
lum, adjust curves on both, color adjust the RGB if needed). Select the lum image, copy and
paste on the RGB image. They don't align...too bad (I use MaximDL to get the alignment, but
you can also use Photoshop). Set the lum layer that was copied to the RGB as a difference
layer, use Edit>Free Transform to translate and rotate that layer as required to align with
the RGB. There are many ways to combine the lum and RGB frames. One method is to set
the lum layer to soft light, and adjust the Opacity, typically between 50 to 80%. Another
approach to combine the frames is to set the lum frame as a normal layer (1st layer on the
stack to be formed), duplicate the RGB frame, desaturate (image adjust > desaturate), then
select all, copy, and paste it as a muliply layer on the (normal) lum layer. Select, copy, and
paste the RGB on the layer stack as a color layer. Adjust the opacity of the desaturated RGB
layer (middle layer), typically 40-60%. Flatten the stack, then readjust levels/curves as
desired. Crop as needed, and you are through with the image. Now go back and try another
set of settings on earlier steps with the same sequence/s to build a second, and even third
image, select the best....sometimes you get lucky, sometimes not, if not, try again another
night.
Seeing, and to a lessor extent transparency is what yields the best photos, keep up with
your local conditions. Use the ClearSky forecasts (cleardarksky.com) and your preferred
weather site (like intellicast.com). The Unisys site offers a good indicator of seeing.
Dual Camera Use
I use two of the Atik, modified webcams (1HS and 1C). The monochrome version to collect
lum data, the color version for chrome data. The data collection time is limited (discussed
below), so I use the Tak multiport turret which allows quick changes from one camera to
another. The turret is also great for switching between different eyepieces. There are many
popular webcams out there, Philips, Celestron, Meade, and others, they're all the same for
this type of work. Then there are the USB 2.0 and firewire trinkets. Some of these offer
60fps full frame, to 100fps subframe, without compression, but they are much more
expensive than the standard USB 1.0 webcams (that only run 5fps uncompressed, 10fps
with minimal compression). I recently purchased a Lumenera Skynyx 2-0, usb2 monochrome
cam, the Lucam Recorder capture program, and have attached the cam to my Truetech filter
wheel (using an Astronomik LRGB filter set). You'll find comments concerning its use on
pages 2 and 3.
I use K3CCD to control the usb1 webcams that are used to make the images. This program
(it came with my Atik cameras) offers general control for the functionality of a webcam that
allows for long exposures (that capability isn't used for planetary imaging). Whether you
use K3CCD, or other control programs, the camera settings are important. You want to
minimize image compression artifacts. Keep the frame rate and gain low. I run 5 or 10fps.
Keep the gain below 25%, preferrably below 20% (on the typical slider scale). Set the
exposure as fast as possible (within exposure meter limits, discussed below), 1/50 sec is
good, 1/35 sec is typical, 1/25 sec is required for some planets, look at the exposure meter
(high frame rate cams won't help if the target is dim (Uranus, Neptune), due to the long
exposure required, they would really help on bright planets (Venus, Mars, Jupiter). Set the
gamma and saturation by observing the image on the preview screen, I set mine between
50 and 75%. Adjust the brightness between 50 and 75% using the preview screen and the
exposure meter. You would like to keep the exposure meter around 110 to 130 on the
typical 8 bit webcam 0 to 255 scale (higher end if you're trying to catch one of the planetary
moons, lower end if only the planetary ball). Choose a sequence length consistent with the
apparent motion of planetary features (due to planetary rotation as seen from your vantage
point). This translates into a sequence between 250 and 500 seconds on Mars, 50 to 150
seconds on Jupiter, (arguably) unlimited on Saturn, Venus, Mercury, Uranus, and Neptune
(I won't discuss Pluto, or other KBOs, you need long exposure for these). You are going to
stack all of the individual AVI frames that make up a sequence, your stacking software, or
hard drive may have limitations, so check the number of frames that will come out at the
selected frame rate and sequence duration, you may need to limit the sequence length, but
try to keep the number of frames above 400. It's nice to have 2000 frames (planet rotation
and camera frame rate limitations, especially usb 1.0 cams, sometime preclude collecting
large numbers of frames). You can really run up the frame totals with the usb2 cam, you'll
need more RAM (as in at least 1GB, preferrably 2GB) on that old laptop.
your local conditions. Use the ClearSky forecasts (cleardarksky.com) and your preferred
weather site (like intellicast.com). The Unisys site offers a good indicator of seeing.
Dual Camera Use
I use two of the Atik, modified webcams (1HS and 1C). The monochrome version to collect
lum data, the color version for chrome data. The data collection time is limited (discussed
below), so I use the Tak multiport turret which allows quick changes from one camera to
another. The turret is also great for switching between different eyepieces. There are many
popular webcams out there, Philips, Celestron, Meade, and others, they're all the same for
this type of work. Then there are the USB 2.0 and firewire trinkets. Some of these offer
60fps full frame, to 100fps subframe, without compression, but they are much more
expensive than the standard USB 1.0 webcams (that only run 5fps uncompressed, 10fps
with minimal compression). I recently purchased a Lumenera Skynyx 2-0, usb2 monochrome
cam, the Lucam Recorder capture program, and have attached the cam to my Truetech filter
wheel (using an Astronomik LRGB filter set). You'll find comments concerning its use on
pages 2 and 3.
I use K3CCD to control the usb1 webcams that are used to make the images. This program
(it came with my Atik cameras) offers general control for the functionality of a webcam that
allows for long exposures (that capability isn't used for planetary imaging). Whether you
use K3CCD, or other control programs, the camera settings are important. You want to
minimize image compression artifacts. Keep the frame rate and gain low. I run 5 or 10fps.
Keep the gain below 25%, preferrably below 20% (on the typical slider scale). Set the
exposure as fast as possible (within exposure meter limits, discussed below), 1/50 sec is
good, 1/35 sec is typical, 1/25 sec is required for some planets, look at the exposure meter
(high frame rate cams won't help if the target is dim (Uranus, Neptune), due to the long
exposure required, they would really help on bright planets (Venus, Mars, Jupiter). Set the
gamma and saturation by observing the image on the preview screen, I set mine between
50 and 75%. Adjust the brightness between 50 and 75% using the preview screen and the
exposure meter. You would like to keep the exposure meter around 110 to 130 on the
typical 8 bit webcam 0 to 255 scale (higher end if you're trying to catch one of the planetary
moons, lower end if only the planetary ball). Choose a sequence length consistent with the
apparent motion of planetary features (due to planetary rotation as seen from your vantage
point). This translates into a sequence between 250 and 500 seconds on Mars, 50 to 150
seconds on Jupiter, (arguably) unlimited on Saturn, Venus, Mercury, Uranus, and Neptune
(I won't discuss Pluto, or other KBOs, you need long exposure for these). You are going to
stack all of the individual AVI frames that make up a sequence, your stacking software, or
hard drive may have limitations, so check the number of frames that will come out at the
selected frame rate and sequence duration, you may need to limit the sequence length, but
try to keep the number of frames above 400. It's nice to have 2000 frames (planet rotation
and camera frame rate limitations, especially usb 1.0 cams, sometime preclude collecting
large numbers of frames). You can really run up the frame totals with the usb2 cam, you'll
need more RAM (as in at least 1GB, preferrably 2GB) on that old laptop.
