AstroPhotography Tips:

The following "tips" are recorded here mostly for my own later use.  It is unwise for anyone else to try to bypass the "floundering learning process" that leads to an understanding of why some things don't work and others do.  The learning process is the fun part!

Wide-Angle Shots

A 35-mm cameras is ideal for constellations, the Milky Way, aurora and other big area shots.  Larger format cameras (i.e., 4.5x6 cm medium format) are a poor investment because film scanners for medium format film are very expensive (and quality commercial film scanning can cost $25/frame).

Use color negative film, not slide (positive) film.  This assumes you're going to work with a digital version of the image obtained using a film scanner (I can't imagine someone NOT working with a digital image obtained with a film scanner).  Affordable film scanners use light detectors that saturate whenever the sampling light beam abruptly encounters a star image's transparent spot on the film, so the resulting image has annoying dark ghosts.  The same scanners have no problem when the detectors encounter a star's dark spot on the film, which means negative films are preferred.

Wide angle shots require dark skies, that are only possible from mountain sites.

Since exposures of a minute or longer are needed, it is essentially necessary to mount the 35-mm camera to a telescope that has sidereal tracking.  Tracking requirements are relaxed for wide angle shots, so a small equatorially-mounted tracking telescope can be used - such as a Meade ETX125 (costing $500).  Short exposures (<2 minutes) won't require manual adjustments (provided you do a good polar axis alignment).

Airplanes are becoming a bigger problem for wide angle photography, especially if you live below a highly travelled air corridor like I do.  I'm ever-watchful and ready to cover the camera lens temporarily whenever I see an airplane approaching.

Medium-Angle Shots

This is where a small refractor with good optics is useful.  Attach it to t big scope for guiding, and count on hand guiding adjustments every 30 seconds or minute. If you don't have one, you'll have to use a telephoto lens made for your camera.  I have a 400 mm FL f/4.5 which works OK (if you're willing settle for degreaded image quality near the corners).  The camera and telephoto lens must be mounted to a big telescope with good tracking properties (7-inch meade or bigger, the ETX125 won't give good enough tracking).

Dark skies are not as important for medium-angle photography as it is for wide-angle photography.

Narrow-Angle Shots of Dark Sky Objects (i.e. CCD Photography):

CCD's are ideal for dark sky objects because they're very sensitive.  However, they're so small that they are limited to narrow fields of view.  With my Meade LX-200 f/6.3 and "3x focal reducer/field flattener" I get an effective focal length of 695 mm!  Even with this relatively short focal length (for a large aperture), the CCD imager I use (Meade 4216XTE) is so small (6.9 x 4.6 mm) that the sky coverage is just enough to include the first quater moon (image size is 34.1 x 22.5 'arc).  It's too small for the Pleides, the Andromeda Galaxy or the North America Nebula, for example.  Vignetting is a problem for extended dark sky objects, and will require a field flattening frame correction (for some objects you can cheat by using software to figure out the vignetting fucntion, and removing it - MaxIM DL's "Flatten Background" does a good job on most occasions).

Good focus is always essential, even for dark sky objects.  So spend time focusing on nearby stars.  Having some faint stars in the focusing field is a good idea, because they appear only when focus is good, whereas a bright star mostly changes size as the focus setting is adjusted.

Cold CCDs are less noisy, so "go for the cold."  Take several dark frame images after the CCD has stabilized at the temperature setting, and use an average of several dark frames for subtracting CCD noise from image frames.

If a "focal reducer/field flattener" is used (such as Meade's 3x, which is really a 2.4x), take exposures of the daytime sky (away from the sun) to serve as "frame flattener" images.  Expose the flattener images as long as needed to bring the brightest pixels close to saturation (>50,000 counts is a good goal).  Average a few such frames to reduce introducing noise into the flattened image.  If you're going to mosaic images together (to cover a large area of the sky), it's imperative to employ flattening to remove as much "vignetting" as possible before joining the mosaic image elements.

Exposure times will depend on whether you're using some kind of CCD guiding.  I don't, so my exposures are limited to 30 seconds, usually.  The drift rate of the star field will depend on the quality of your polar axis alignment, which can be about 2 pixels per minute when observing at a home station, where a good polar alignment can be achieved.  If the drift rate is greater than 3 or 4 pixels per minute, which is typical for me, I simply set the Pictor View "Number of continuous frames" to a large number, like 10, and do the combining the next day with software.  It may be labor intensive, but at least I get to reject periods with poor seeing or poor guiding.

Planet Photography

I've had the best luck imaging planets using a regular digital camera (not a CCD imager) placed against an eyepiece (focused for infinity).  This technique is called "eyepiece projection."  I've constructed a simple mounting mechanism which at one end screws to the telescope tube and at the other end provides for the fastening of the camera using a standard camera bolt (1/4-10 size).  However, it is now possible to buy a simple device that mounts any common digital camera to a standard eyepiece:  ScopeTronix sells a "Digi-T" for $60.  The sequence for use is to center the planet image in the eyepiece visually, focused in the normal way, then the digital camera is mounted to the same eyepiece (without changing the focus setting).  Hopefully, the planet image appears in the center of the camera's LCD display.  The optical and digital zoom features can be used to produce a large image, while making small pointing adjustments to keep the image centered.  Focusing for the camera can be done two ways:  1) force the camera to focus on infinity, and adjsut the telescope focuser until sharp images appear, or 2) allow the camera to auto-focus for each exposure.  The focusing should be done with a full aperature, and then for the desired pictures you can use an aperture cover with whatever opening size gives the best "atmospheric seeing" improvement.  I have a shutter relase cable for my digital camera to minimize vibration during the exposure.  My camera has a feature of exposing multiple pictures 1/2-second apart while the shutter release is pressed, and this is useful for obtaining a couple dozen shots with one operation.  Typically, only 1% of the images will be used in the processing phase, so you can never have too many raw images to work with.  Taking 2000 images during a 2 or 3 hour session is reasonable.  Keep in mind, that processing will usually take more time than the acquisition of raw images.  Sometimes processing can consume MUCH more time!  But that's OK, since processing can be lots of fun.

When a planet is low in the sky it will have the appearance of color distortion, with blue at the higher elevation part of the disk and red at the bottom.  This is caused by atmospheric refraction which bends blue light more than red.  This color problem can be removed during processing by splitting the image into its three component colors (red, green blue) and then re-merging them with appropriate offsets.  For Mars, I've found that green and blue can be kept together, making just a two component splitting and re-merging.  I use Ulead System's PhotoImpact for this.

Combining several good images is useful, as it reduces the appearance of "grain" - or pixel noise.  It also forces features to be located in their proper position, although resolution is sacrificed for this.  Remember, the atmosphere distorts the image in ways that "move" features by several seconds of arc, causing a circular planet to take on odd, non-circular shapes. (Seeing is discussed in the next section.)  It is easy to quickly visually review 1000 or 2000 images.  I use the versatile viewing program ACDSee for this.  While stepping through the many pictures at a fast rate I stop whenever I notice a good image, and write down it's picture number.  Later, I transfer the good pictures to another directory.  It makes sense, usually, to try to combine images that were taken close together in time.  Jupiter and Mars rotate, so feature movement dictates how long an interval can be used for selecting images for averaging.  For Jupiter it can be as short as 15 minutes if the seeing is good.  For Mars the "window" can be 1/2 hour long.  I use Ulead's PhotoImpact for all my planet processing.  I'll load 6 to 8 images, and determine if there are clear winners.  Sometimes it's necessary to enhance all the images (the same way) before deciding if there are winners.  I'll average the two best images, then the two next best.  Then I'll decide whether to average the averages, etc.  When a good average image is created, it then is subjected to the color splitting and recombining to remove the blue-top/red-bottom effect.  After several repeats of this process there may or may not be an image that is worthy of keeping, but usually there isn't.  Processing is a lot of work, and if it's not enjoyable then you won't be good at it.

Outer Planet Satellite Imaging

Jupiter, Saturn, Uranus and Neptune have satellite systems that are within reach of many amateurs for CCD imaging.  The paramount challenge to overcome is the immense brightness of the parent planet in relation to the satellties.  A 16-bit CCD imager is a great help in this regard.

Clearly, the farther out the satellite is form the planet, the easier it is to image.  Jupiter's four Galilean satellites are easy to image, as is Saturn's Titan, and Neptune's Titan.  The rest require good technique.

The main items to consider are: 1) image scale (long focal length), 2) exposure time (short), and 3) processing technique (the next day).

Use the longest focal length possible to increase image scale (pixels per "arc).  Don't use a "focal reducer" lens, even though it improves f-ratio and theoretically shortens exposure time for faint objects.

Frequent focus checks may take time, but they are crucial.  My Meade LX-200 10-inch SCT can require focus knob changes of 10 degrees of rotation during a half hour of cooling (or possibly it's related to orientation).  It's possible to establish a best focus setting to about 3 degrees of knob rotation.

Exposure time should be such that the planet is just barely saturating each image.  This will enable processing to use the "maximum entropy" method for resolution enhancement (since a star-based "point spread function" can be used to process the planet's image, and reduce its areal coverage in the processed image).  For Uranus and Neptune a 3-second exposure (using the Meade 416XTE) works well.  Many short exposures can be used to noise, and thereby encrease the signal-to-noise for the satellite images.  Colder CCDs generate less pixwel noise, so "go for the cold."  Processing requires subtraction of dark images, of course.  Remember to take dark frames long after viewing a bright object, since the CCD circuitry has a "memory" that can last a minute of longer.  Pixel cleaning with PictorView (using the Median Filter in Ver. 7.14) has resolution penalties, so try not to use it unless it's necessary.  Taking many images ("number of continuous frames" setting) with one command is a good idea.  It's important to take all the images that will later be averaged together in as short a time as possible, since the planet is moving with respect to the star field (and it's esthetically good to have point stars as part of the image).  For Neptune, for example, all frames that will be averaged together should be taken within a half hour.  For Uranus, an evenshorter time is needed, since Uranus moves across the star field faster, and it has satellites with shorter orbit periods.

Processing of images can take many times the time required to obtain the images, and this is where most of your time should be spent.  The day after observing I use Pictor View to go though all (dark frame subtracted) images of a given object, and record my subjective impression of image quality, and the time each acceptable image was taken (I note any differences, such as exposure time).  Then I decide which images are to be averaged together, and create a list of intended image groups.  I use the software imageing program "MaxIm DL" to "Combine/Overlay" the images in a group.  Then I use MaxIm DL to enhance the averaged image in many ways.  The two I've the best luck with (so far) are: 1) Maximum Entropy (specify a star for the "point spread function"), 2) FFT High Pass, Custom, 40/35.  Using these two processing techniques produces a noticeable increase in the dark space between the bright planet and its faint satellites.  If the image will require notation (labelling objects, showing the satellite orbit, drawing circles around special objects), I adjust the FITS image for best appearance, and save it as a JPG (or BMP) file.  I then use either Photo Impact (JPG) or Paint (BMP) to add the notations, before creating a final JPG image file for a web page (or e-mail attachment).  When saving a JPG file try different "compression percentage" settings, and view the outcomes (with ACDSee) to determine when losses begin (this will keep file sizes to a minimum).  Some images may need a scale change (image size in pixels), and I use Graphic Workshop Professional for that.

Atmospheric Seeing

"Seeing" is typically not a problem for "dark sky" objects.  It mostly affects planetary and lunar work, since resolution is far more important than brightness levels.

The following link is devoted to "atmospheric seeing."

Moon Photography

For the moon (close-ups), I've had good luck using an astronomical CCD imager at the prime focus of my telescope.  Also, you  can use a regular digital camera (such as the Nikon Coolpix 990) mounted to an eyepiece (using a ScopeTronix Digi-T) in "eyepiece projection" configuration - which is what I recommend for planetary photography.  This will give highly magnified views.

For images of the entire moon, there are two main choices: 1) a good medium focal length refractor (20 to 46 inches focal length) and a 35 mm film camera, or 2) a 3x "focal reducer/field flattener" lens attached to a large telescope with a focal length of 1600 mm or less (using  CCD imager).

Image Processing

This is where the action is.  I count on spending more time "processing" images than taking them.  Because of this, astrophotography is really a mostly daytime activity!

My favorite programs are used in the following sequence:
    1) MaxIm DL:
        a) to flatten background (for wide angle, dark sky shots),
        b) to re-map brightness levels using a Gamma value of about 0.7 for dark sky objects (to enhance low brightness level pixels),
        c) to improve visibility of a faint satellite that's near a bright planet, using FFT or Maximum Entropy,
        d) to average images to improve signal-to-noise (Combine, Overlay),
        e) to combine RGB or RGBL images into a color iamge, then balance color background and bright areas.
    2) Ulead System's PhotoImpact:  to adjust background level of each color separately, further use of Gamma if appropriate, averaging of several images (after rotation, if necessary),
    3) Graphic Workshop Professional:  to re-scale (change pixel size of image), and convert to JPG format.

Drawing constellation lines is easily done using Microsoft Paint, but it requires conversion to BMP-format & re-conversion to JPG.  There must be an easier way, but I haven't discovered it yet.

[This is a work in progress]


This site opened:  May 24, 2001.  Last Update:  August 28, 2001