Imaging Workflow - The New Astro™ Zone System For Astro Imaging

Transcript

DVD Bonus Chapter Imaging Workflow  Imaging Workflow Overview There are many possible steps involved in imaging. Some things you do all the time; other things you do when they are needed, or only with certain types of cameras. And some things you only do when your skill and experience reach the level required! This chapter provides a detailed list of the steps involved in imaging. In addition to various planning, preparation, and equipment issues, there are 6 steps in processing an image: Acquisition You might not think of taking images as part of image processing, but there are things you can do when taking images that have a major impact on processing. Most of these have to do with keeping the noise in the image as low as possible. Reduction (Also called calibration) Reducing noise and removing instrument (system) artifacts from the data. Balancing (Color only) Sometimes referred to as normalization, this is the process of balancing the contribution of the red, green, and blue channels to each other. Includes color bias, color balancing, noise-related issues, and saturation. Histogram Includes both linear and non-linear histogram changes. Linear changes (Levels) set the background. Nonlinear changes (Curves) expand the Dim zone and compress the Bright zone. Zonal Processing Applying specific types of processing to the different zones, such as blurring the Dim zone and sharpening the Bright zone. Cleanup Some types of cleanup can be done at any time during image processing, but some types of cleanup are easier to do after you are nearly done with the image. Cleanup includes removing hot and cold pixels, tweaking contrast to enhance details, removing gradients, removing halos around stars, etc. This chapter is only one-half of the story outlined above. This chapter is provided on the DVD, and focuses on the acquisition phase of imaging. To learn about processing-related workflow, please see the printed book.  Chapter Structure There is more to workflow than just the steps involved. First and foremost, the steps don’t line up cleanly in a certain order—you will find yourself going back and tweaking things that you thought were complete. Some activities, such as hot and cold pixel removal, can occur at different times (after reduction, or after zonal processing) and in different ways (pixel filters, the Clone tool, the Healing brush). The order you follow will also depend on your experience level, personal preferences, available tools, desired outcomes, and so on. Since there is no one way to process, there is also no one way to teach you about workflow. I have therefore included multiple ways to look at workflow: • Advice and Workflow for Beginners • Underlying Concepts (Nuggets of CCD Wisdom) • Planning and Preparation/Equipment Issues • Image Processing Procedures Preparation is at least as important as execution, so you’ll find plenty of information on preparation and setup here, too. Advice for Beginners Unless you have extensive film experience, you should consider starting with the easier objects: moon, planets, and/or globular clusters. Then you can move on to bright nebulae and galaxies. Unless you have a one-shot color camera, starting with monochrome images keeps things simple. Focusing is a very important skill. Motorized focusing and computer-assisted focusing will greatly ease the job. Use small steps to adjust focusing so you don’t go past focus unintentionally. Start the process close to focus whenever possible (mark the focus position when you find it the first time). Bright stars and the edge of the moon make good first focus targets; use very short exposures (0.001 to 0.1 second). As you gain more skill, you will move on to focusing with dimmer stars and longer focus exposures. Polar alignment is much more important for imaging than it is for visual use. Learn to drift align, or use a tool like PoleAlignMax, TPoint, or MaxPoint to get more accurate polar alignment. It may seem like wasted time to get excellent polar alignment, but you will in fact save even more time in increased efficiency. Bad polar alignment ruins many an image! Start out with automatic dark subtraction (also called AutoDark). Your camera control software will have a setting for this. You can move on to more sophisticated calibration techniques later. Always use autodark for focusing and for guiding. Always! When processing, watch out for the basics: don’t wash out bright areas, don’t remove dim details when raising the black level. Take your time; there may be more data in the image than you realize. If you have light pollution, you will likely be dealing with backgrounds that vary in brightness from gradients. Don’t panic; as you gain more experience, you’ll be able to deal with these gradients more effectively.  CCD imaging isn’t like regular photography. Take the time to learn about image reduction/calibration early in the process. Take and use dark frames, learn when to use bias frames, stick with taking flats until you get it right. Start with sky flats. If you are having trouble with your mount, you can use a faster focal ratio because it will put fewer demands on the mount. You can do this with a reducer, or by using a different telescope. In the long run, learn to evaluate the performance of your mount, and how to tune it for best performance. Guiding and color imaging are not the first things you need to learn. Learning to master your mount is job #1. Many subsystems must work together in order to get good images, but every one of them sits (literally!) on the mount. Social learning is often useful—locate and team up with imagers in your area, at star parties, and at regional and national conferences. Expect problems, and learn to be a detective. Your particular equipment will have its own set of typical problems (e.g., mirror flop on most SCTs). You can learn what problems to expect with your setup by participating in user forums and web sites related to your equipment. Spend the time to familiarize yourself with the common types of problems, and what causes them. Learn how to recognize them and how to fix them. Some issues will be subtle. For example, you’ll learn that soft stars could be the result of bad focus (use a motorized focuser; use a dial indicator for manual focus; get a bigger pixel camera or bin your existing camera), collimation errors (use a refractor instead; learn how to collimate; get a telescope that holds collimation better), and poor seeing conditions (use a shorter focal length scope, get a bigger pixel camera or bin; find a better location or wait for a better night). As you can see, there are often many ways to deal with a problem. Don’t assume that new concepts mean what they seem to mean. Things like binning, collimation, and dark frames (just to name a few) often have subtle side-issues that take time to learn. Be patient with yourself, and be careful when you use familiar ideas to understand CCD imaging. Sometimes, CCD is so different from other types of imaging that trying to use old concepts can take you down the wrong road. Pay attention to your mistakes. How did they happen? Is it really something under your control? What could you do to prevent the same problem the next time? When you see good images taken by others, start asking questions. For example, if you like the way that background galaxies are shown in an image, ask the imager what technique they used to capture those dim, distant galaxies. Workflow for Beginners Polar align carefully. The longer your focal length, the more accurately you must do this. Fifteen extra minutes polar alignment can save you an hour later on by avoiding errors and lost images. Learn basic control of your mount. This includes knowing the available movement rates and how to select which one is active (e.g., slewing, centering, and guiding). Later on, you’ll  want to learn how to change the guide rate if available, but use the default rate to start with. Focus carefully! My first book on CCD imaging, The New CCD Astronomy, contains detailed information on manual focusing. But if you just want great results, the fast-track solution is to get a computer-controlled motorized focuser. Especially in the beginning, take the time to verify that it has reached best focus. The dimmest stars in the image are always your best indication of accurate focus; they are the first to show when you are out of focus. When you are out of focus, the dimmest stars simply disappear! When you are focused, they show up as very dim points of light. Whatever method you use, take the time to get good focus. Nothing makes an image useless faster than poor focus. When a computer is involved, take steps to get your software in order. Make sure you have current drivers for hardware—mount, focuser, camera, etc. Learn what settings to use for your hardware. If there is an order you must follow, learn what it is and follow it. The typical computerized setup will include a planetarium program that can talk to your mount (I use TheSky from Software Bisque), camera control software, specialized software for certain hardware (e.g., a focuser), image processing software, and perhaps a pointing program (e.g., TPoint) or a focus assist program (e.g., FocusMax). Determine what you want to image. You can use planetarium software, books, advice from friends or online resources (e.g., Yahoo groups), magazines, etc. Think about the object’s brightness, angular size, position in the sky, and so on before you make a choice. Objects with high surface brightness make the best targets - M42 is going to be much easier to image than the Rosette Nebula, and both are best with a widefield telescope, such as a small refractor. Point your mount to the object you want to image. If you are pointing manually (no GOTO), you can use an aligned finder to do this, or remove the camera and use an eyepiece to position the mount. Digital setting circles, connected to either a control box or a computer, provide increased accuracy. GOTO with a hand controller or connected to a computer is by far the easiest and most productive method to use. Learn the lingo of CCD imaging (see the next section, Nuggets of CCD Wisdom). Here’s a suggested outline for your first imaging sessions: 1. Start your camera control program, and either set the camera to Simulator or connect your camera and explore the camera control program’s features. 2. Connect the camera and experiment with imaging indoors, perhaps using a camera lens (if you have the right adapter) or use some aluminum foil on the camera nosepiece to make a pinhole lens. 3. Attach the camera to the telescope in daylight and confirm proper fit and operation. Make sure that the camera can be attached firmly—wobble must be avoided. 4. At night, practice focusing on a bright star. Take some images of the moon. It can be challenging to focus on the moon, but you might try tweaking focus as you take a series of images.  5. At night, use the camera to get a really good polar alignment using one of the methods mentioned earlier, such as TPoint. Once you have done these basic things, you can try your hand at longer exposures. Using your camera control program’s autodark feature, you might try M42 for 10 seconds, or a globular cluster for 10-15 seconds. Some galaxies are bright enough to be recognizable in exposures as short as 10-30 seconds. The better your polar alignment is, the longer the exposures you can take. Your first exposures can use the AutoDark feature of your camera control program. At some point, you will need to start using real darks, flats, and perhaps bias frames. But for starting out, AutoDark gives you decent enough results. If you are feeling adventurous, take several images and learn how to combine them using your camera control program’s image processing features. If you are getting round stars, take longer individual exposures to see how long of an image you can take. The better your polar alignment, and the shorter your focal length, the longer you can expose. If you can’t take exposures of more than a few seconds, you may be using a very long focal length telescope, and will need to learn guiding sooner rather than later. Compare your various exposures and exposure times, and see what kind of difference you get with the longer exposure times. You should see less grain, and more object detail, with longer exposures. Nuggets of CCD Wisdom The usual method of supplying a few definitions won’t work for this chapter. The things you need to know change as you gain more experience. Rather than overwhelming you with definitions covering the full range of CCD imaging, I have broken the definitions down into groups based on various criteria. The definitions you find here are intended to introduce you to concepts. You will often find more detailed definitions and descriptions in the printed book. For now, I suggest that you become familiar with the tools, concepts, and functions involved in CCD imaging. You can get more information on these terms elsewhere in this book, and in my earlier book, The New CCD Astronomy. Starting out Darks - Also called dark frame. Records the dark current (also called thermal current) that occurs during an exposure of a given duration and temperature. The dark frame is subtracted from the light frame (same duration and temperature) to remove the dark current. This makes the image much cleaner looking. Taken with the shutter closed. If the dark and light frame do not match in duration, a bias frame can be taken to scale the dark frame, and allow it to be applied to the light frame. Light Frame - Also called simply an image. Taken with the shutter open. The quality of the image depends heavily on the duration of the exposure, (and how good the focus is!).  Bias - Also called bias frame. Records the starting levels of all pixels. A dark records the state of the pixels at the end of the exposure, and the bias records the state of the pixels at the beginning of an exposure. Subtracting the bias from the dark yields the state of the pixels during the exposure itself. Cooling - CCD chips are cooled using thermo-electric coolers. Cooling reduces the dark current, and permits deeper exposures with excellent details. Buffer - Also called window. An image is downloaded into a buffer/window where you can see it. Very important: In many programs, you need to turn on the Autosave feature or the image will be lost when the next one is taken. Autosave - A feature available in most camera control programs. Turn it on and specify a folder where you want images saved. This is a nice safety feature; you can delete bad images later. If Autosave is not turned on, you risk losing images! AutoDark - When this feature is turned on, the camera control program will take a light frame and a dark frame, one right after the other, and automatically subtract the dark frame to show you a clean image. Dark frames are most effective when multiple dark frames are combined (to lower noise), so AutoDark is useful when starting out, and as you gain experience it becomes a quick way to check focus and framing. Always use AutoDark for focusing and guide exposures. Subframe - A portion of the CCD chip used to take the image. Useful for focusing because it takes less time to download a portion of the chip, allowing you to take many focus images while you (or a motorized focuser) adjust focus position. Calibration/Reduction - The process of applying dark, bias, and/or flat-field (see page 9) frames to light frames. Different software packages use different names for the same procedure. Tracking (of mount) - A polar-aligned mount tracks at the sidereal rate—the rate at which the stars appear to move across the sky. The accuracy with which the mount follows the apparent motion of the stars is critical to successful imaging. Tracking accuracy is limited by periodic error and random mechanical errors. Periodic Error (PE) - Tracking errors caused by irregularity in the worm gear. All mounts have at least some periodic error. The lower the periodic error the better. PE is measured in arcseconds of mount deflection. PE in mounts ranges from very low (a few arcseconds) to barely adequate for imaging (20 arcseconds or worse). RA and Declination - The coordinate system used to find objects on the Sky. Similar to the latitude/longitude system used for specifying locations on earth. RA (Right Ascension) is the night sky equivalent of longitude, and Declination is the equivalent of latitude. Slew - The act of moving the telescope to a new location. Pixel - The fundamental imaging unit on a CCD chip. When photons strike a pixel, they are converted into electrons (abbreviated as e-). Circuitry in the camera counts the number of electrons in each pixel, and assigns a brightness level. Blooming - Pixels are like little buckets. When they fill up with electrons, they can spill over into adjoining pixels. This spillage is called blooming. Blooming typically runs along  a column because most CCD chips link pixels in columns rather than rows. Some cameras use anti-blooming (ABG) chips, which have special circuits to bleed off electrons before they spill over into adjoining pixels. The circuitry blocks some of the incoming photons, however, so most anti-blooming CCD sensors are less sensitive than their non-anti-blooming (NABG) cousins. Exposure time - How long you expose the CCD chip. For deep-sky objects, lots of long exposures are ideal. But when you are just starting out, exposures of even 15-60 seconds can yield interesting images of deep-sky objects. Focusing - A black art involving magic incantations. Seriously, focusing is the key to getting the best possible images but it can be challenging to do. My earlier book The New CCD Astronomy has a full chapter on nothing but focusing. My favorite method for visual focusing: when the dimmest stars in the subframe are small and tight, focus is right. Polar Alignment - The act of aligning the telescope RA axis with the earth’s axis of rotation. Common methods are drift alignment, firmware built into the telescope controller, and specialized computer software (e.g., PoleAlignMax, TPoint). Basics of Image Processing Histogram - The histogram is the basis for most of what you do in image processing. It’s a graph that shows how many pixels exist at each brightness level. By manipulating the histogram, you can change what brightness levels are visible in the image. Stretch - A word used to describe changing the histogram settings. A linear stretch refers to setting the black point and white point. A non-linear stretch involves a Gamma change or the use of Curves (e.g., in Photoshop®). Autostretch - A feature in most camera control programs that does automatic linear histogram changes. Designed with settings that will show the objects in the image clearly. In most cases, there is a way to change the behavior of the autostretch feature. Goes by other names as well, such as Auto Contrast. Screen Stretch - A MaxIm DL feature name, but exists in some form in all camera control programs. Refers to how the image contrast is adjusted for display on the screen. Background/Range (Min/Max) - These settings determine the screen stretch in programs such as CCDOPS. Black Point - Applied to a position on the histogram. All pixels darker than the black point will be rendered as black on the screen. A non-destructive setting in camera control programs, but in Photoshop® changing the black point actually sets the darker pixels to black permanently. White Point - The opposite of the black point. All pixels brighter than the white point will be rendered as white on the screen. The same Photoshop® caution described for the black point applies to the white point. Levels - A tool in Photoshop®. Used to manipulate the black point, white point, and gamma setting of the histogram. For most users, most of your interaction with the Levels tool is to adjust black and white points.  Curves - Another Photoshop® tool. Used for advanced histogram manipulation. Levels adjustments are linear; Curves adjustments are non-linear because some portions of the histogram are expanded, and some are compressed. Combine/Stack - Individual images have noise. Combining images lowers the total noise level. You can’t combine an image with itself to lower noise; you have to use separate, multiple images for combining. Alignment - (offset, rotation, scale, projection) It is often necessary to align the images to a reference image before combining them. Alignment may involve horizontal and vertical shifts (offset), rotation about an arbitrary point, resizing (scaling), and geometric corrections (projection). The last is only available in Registar at the time this is being written. Brightness and Contrast - A simple tool for adjusting an image. Can be useful for quick adjustments to an image for beginners. Brightness changes the black and white points simultaneously. Contrast applies a simple stretch to the histogram. You will get better results by adjusting black point, white point, and histogram stretch individually. Guiding, Color Imaging and Intermediate Topics Binning - Combining groups of physical pixels to create a single virtual pixel. Binning lowers read noise, giving you the most out of short exposures. Most cameras support unbinned (1x1), as well as 2x2 and 3x3 binning. There are two good reasons to bin. If the seeing isn’t good enough to support unbinned imaging, you can bin to lower the resolution to match the seeing. You can also use binning for color (RGB) images that will be combined with unbinned luminance data. Binning improves the signal to noise ratio on dim objects. When you bin 2x2, the output virtual pixel contains only one dose of readout noise. This is half the readout noise you would get from four physical pixels. The result is an image where dim details show up more clearly. (Derived from work by Stan Moore, by permission.) Flats - An image which records the optical issues of your telescope and camera. Typically, the flat records vignetting as well as shadows from dust motes. Camera control software applies the flat to the image to clean up the darkening/brightening from these sources. Saturation - A pixel is saturated when it is full of electrons. Don’t assume that saturation is the same as the maximum value available—this is typically not the case. For example, many cameras have a maximum value of 216 (65,535), but may saturate somewhere in the range of 50,000 to 60,000. Technically, saturation is the full well capacity in electrons divided by the gain (electrons per analog to digital unit (ADU)). Full Well - A technical specification for a camera that defines the maximum number of electrons that a pixel can hold. Typical full well capacities range from 50,000 to 600,000 electrons. A camera with a larger full well will take longer to bloom. Gain - The rate at which electrons are converted into ADU (analog to digital units, or brightness levels). For example, suppose that a camera has a full well of 100,000 electrons, and the analog to digital conversion is being done at 16 bits. 10 That gives only 65,535 available brightness levels. The gain of the camera must be set so that approximately 1.4 electrons equal one brightness level. FWHM - Short for Full Width at Half Maximum. Expressed in arcseconds. Describes the size of a star image, and lower numbers are better. Often used to describe seeing conditions, as in “Last night, the seeing was about 3.5” FWHM.” If you graph the brightness of a star image, it will look like a slightly rounded mountain peak (approximately a Gaussian curve). FWHM is the width of the curve halfway between the base and the top. Most camera control programs provide a way to measure the star FWHM in your images. All unsaturated stars will have similar FWHM values—the bigger/brighter stars will measure FWHM higher up on the curve. Bias - Taken with the shutter closed using the shortest available exposure time. (The ideal bias frame would have a zero exposure time.) The bias frame records thermal current at the start of an exposure. A dark frame records the thermal current at the end of an exposure, thus one can obtain the thermal current during the exposure by subtracting the bias from the dark. This allows the dark to be scaled for a non-matching exposure time. (After scaling, the bias is added back.) Luminance - Describes an image taken through a clear filter, (sometimes without any filter). There are two types of clear filters: those that pass all wavelengths of light, including IR and UV, and those that pass only what the human eye sees. For some telescopes with refractive elements, colors are not focused to the same point, and the type of clear filter that blocks IR (and sometimes UV) is better. For reflecting telescopes, you can use a clear filter that passes all wavelengths since reflection does not introduce chromatic effects. Note: if you are making LRGB images, you may want to use an IRblocking clear filter even if you are using a reflecting telescope. The IR data in the luminance channel may wash out portions of the color data, especially red. RGB - A type of image that is made up of separate red, green, and blue images. The filtered images are combined to produce a full-color image. LRGB - Similar to an RGB image, but the color data is added to high-resolution luminance (L) data, hence LRGB. The color data is typically binned 2x2, and the luminance data is unbinned. The combination allows you to capture more overall data (more accurately, achieve better signal to noise ratio) in less time. The lower resolution color works well because the human eye sees color less precisely than it sees luminance. Grayscale - Same as a luminance image. Grayscale also refers to an image that lacks any color information. L*a*b*/CMYK - Alternate color models. L*a*b* (LAB in Photoshop®) is a model based on human perception. CMYK (cyan/magenta/yellow/black) is often used in the printing industry, but is not appropriate for Web viewing. CMYK filters are available, but offer no noise improvement over RGB. Resizing - Changing the size of an image; scaling. The most common reason to resize is to match the size of binned color images with an unbinned luminance image. However, you can also resize an image if the stars are soft or bloated, or there is little detail in the image. 11 Color Balance - The act of balancing the amount of red, green, and blue in an image. Because color images are taken through filters, and the filters pass different amounts of light, you typically need to adjust exposure time to get balanced color. Even if you take the correct exposure times, other factors may alter the incoming color balance and you may need to use software to correctly balance the color in the final image. Such factors include dust, moisture, and atmospheric blue extinction at lower horizon angles. Color Bias - Often confused with color balance. Color bias occurs when there is too much light coming through a filter. For example, light pollution is often strongest in green, and this extra green light creates a green bias in the image. The solution for color bias is simple: use Photoshop® Levels to raise the black point in the offending color channel or channels. This removes the bias, and allows you to work on color balance. Important: Always correct color bias before you attempt to balance color! Telescope Rates - (Slew, Center, Guide) Most mounts can move the telescope at more than one rate. It may use one rate for slewing (very fast), and another for centering objects (slower). For guiding, extremely slow rates are used, typically around one-half of the sidereal rate (the rate at which stars appear to move is called sidereal). Guiding - Adjusting the tracking of the mount to match the apparent movement of the stars. The adjustments are made on the basis of star images. Software measures the shift of the star from a base position, and adjusts the position of the mount to compensate and keep the star centered. Guide star - The star used for guiding. The ideal guide star must be bright enough so that software can measure the centroid (center position) of the star accurately, but never so bright as to bloom or saturate. Guide error - The amount by which the position of the guide star is off from the desired location. Expressed in terms of X and Y error in most guide software. Guide error should generally be less than the image scale (arcseconds per pixel). Aggressiveness - Determines how much of the guide error is adjusted at one time. Aggressiveness is typically expressed as a number from 1 to 10. If aggressiveness is set to 5, then 50% of the guide error will be corrected for with each correction. Since seeing variations often conspire to introduce error into the guide star position (centroid), an aggressiveness less than 10 is typically used. Correction - A movement of the mount caused by the guiding software. The amount of the correction is based on the guide error and the aggressiveness setting. Collimation - The act of aligning the optical elements of the telescope correctly. A scope that is not properly collimated will produce softer images than it otherwise would, so time spent on collimation is critical to obtaining optimal results. Seeing - The air that we image through is not steady; it shifts and changes over both short and long periods of time. The combination of these shifts results in a softening of the view through the atmosphere. The amount of this softening can be roughly approximated by the FWHM of a star as imaged through the atmosphere. The softening occurs because air of 12 different temperatures refracts light by different amounts. As the air moves, you are looking through different combinations of air temperatures, resulting in shifts that show up as twinkling for visual observers and a larger FWHM for imagers. Seeing can occur in the atmosphere, at the boundary of your observatory and the free air, and at the telescope optics. For example, a telescope mirror that is warmer than the ambient air temperature creates convection currents. That is the very definition of mirror seeing. Backlash - The looseness in the gears of a mount. The mesh of worm and gear should be adjusted to minimize backlash. However, not all mounts provide an adjustment for backlash. Arcsecond/Arcminute - Terms used to describe small portions of a circle. One degree is 1/360th of a circle. An arcminute is 1/60th of a degree, and an arcsecond is 1/60th of an arcminute. Filter/Filter Wheel - Most CCD cameras take color images using filters. Although you can buy a manual housing for filters, most imagers use a motorized device that moves filters into position. This is called a filter wheel because nearly all of them keep the filters in a wheel that rotates filters into position. A few sliders do exist, but they are the exception. One-Shot Color - A type of camera that uses a filter mask attached directly to the CCD sensor to take color images. It does not use external filters or a filter wheel. The mask has tiny filters, one for each pixel, arranged in a specific type of grid called a Bayer mask. A Bayer mask has four pixels in a 2x2 grid (shown at top right). The entire chip is covered with this grid (bottom right). Software extracts luminance (bright- ness and detail) information from all pixels so you get a nearly full-resolution image. Color data is half the resolution of luminance, but this is OK because the eye sees color with less resolution than luminance. One key thing to be aware of with one-shot color cameras: because they have a built-in filter mask, they aren’t as useful with speciality filters such as h-alpha. Overall, oneshot color cameras provide greater convenience, slightly lower resolution, and significantly less quantum efficiency (the mask blocks some of the incoming photons). Shutter - Many astronomical cameras use physical shutters the same way that regular cameras do. Some astro cameras use an electronic shutter. The data is rapidly shifted out of the imaging area of the chip into a separate storage area, also on the chip, allowing for extremely fast shutter speeds. Such chips are called frame transfer CCD chips. Conventional chips that require a physical shutter are called full-frame CCD chips. Frame transfer chips are better for planetary imaging because of the very short exposure times they offer, but they often have lower efficiency. 13 Advanced Image Processing DDP - Short for Digital Development Processing. This is an image processing filter that was developed several years ago to address the differences between film imaging and CCD imaging. Film records light non-linearly, so that the darkest and brightest portions of the image are somewhat compressed. This compression allows the eye to see more detail in the dim and bright portions of the image. DDP is reasonably successful—it’s sort of a very simplified Zone System, but the Zone System allows you to tune all aspects of the process. DDP combines fewer adjustments into a single step. At the same time as it applies histogram changes, DDP also sharpens the image (although in some implementations, such as MaxIm DL, this can be turned off). Although DDP is somewhat effective, with a little practice you can get better results with histogram adjustments using Levels and Curves in Photoshop®. DDP allows a small measure of control, while manual processing in Photoshop® gives you complete control over every aspect of the process. In particular, you can separate sharpening and histogram processing with Photoshop®, choosing the best method for each. Sum/Average/Median/Statistical combines - Combining multiple images (light, dark, bias, or flat-field frames) reduces noise. Each of these combine methods has specific qualities that make it a good choice for specific situations: • Sum should be used only when you are certain that your software can handle the huge numbers involved. Averaging is simply a scaled sum, and is often a better choice. Sum is useful when you have a really large number of images, as it avoids quantization errors from rounding. However, you need a lot of images to realize the benefit - on the order of a hundred or so. However, if your software supports large brightness value (greater than 216), there is absolutely no harm in using sum. Some programs, like CCDStack, are best used with sum, in fact. • Average (mean) provides the lowest possible noise in the combined image. However, it does nothing to clean up satellite tracks and cosmic ray hits. (Use the Photoshop® Clone tool to clean them up.) • Median does a great job cleaning up satellite tracks and cosmic ray hits, but at a cost: the noise in the final image will be higher than for averaging. At least three images are required for a median combine. • Statistical combines—Sigma reject, min/max clip and other fancy combine algorithms—use statistical techniques to decide how to combine images. Typically, you need at least 5-8 images or more, with the exact number depending on the technique you are using. These combine methods lower the noise level and they eliminate cosmic ray hits/satellite tracks effectively. This comes at a slight cost: typically, these techniques may smooth (blur) the image by a small amount. In most cases, the price is right. Logarithmic Stretch - Also called a non-linear stretch. Linear/Non-Linear Stretch - The two types of histogram adjustments available. A linear stretch makes a simple adjustment to the brightness or contrast of the image and does not 14 change the relative brightness of the objects in the image. However, a linear stretch limits your ability to show both bright and dim details effectively. A non-linear stretch compresses some portions of the histogram and expands others, changing the relative brightness relationships. A non-linear stretch (e.g., Curves) is much more effective at simultaneously showing both dim and bright details. However, it takes some skill and experience to achieve optimal results. Gamma - A simple non-linear brightness and contrast adjustment. In Photoshop®, you adjust gamma using the middle slider in the Levels dialog. Gamma starts at 1.0. Increasing it brightens the image, reducing it darkens the image. A Gamma adjustment changes the way that brightness levels are mapped in the image. Increasing gamma gives more of the 256 available brightness slots to the dim portion of the image. For example, a gamma of 2.0 gives 50% of the brightness slots to the dimmest 25% of the image, doubling the portion of the image devoted to this dim data. (In other words, a pixel that was at the 25% brightness level is now at the 50% brightness level.) The reverse happens if gamma is less than 1.0. Clipping - A somewhat technical but important term. If you create a very aggressive curve in Photoshop® Curves, you can create a curve that intersects the top of the graph. This clips (eliminates) all brightness levels to the right of the intersection point. Identical to lowering the white point to the same level. Generally, you want to avoid clipping as it removes data. Color Calibration - The act of determining the color characteristics of a device (monitor, printer, paper, etc.). This is a complex area, but it is very important if you are printing or publishing your images to the web. For a simple web-based monitor calibration tool, see: http://www.easyrgb.com/calibrate.php Gradient - A serious problem that often occurs in images taken in light polluted skies. Technically, a gradient is a variable bias (the brighter portion of the gradient has a greater bias.) Gradients come in two flavors: linear and radial. A linear gradient occurs when light pollution affects the image. A radial gradient typically results from vignetting in the optical path. You can remove gradients manually in Photoshop®, or using tools in camera control programs. Gradients are especially troublesome for color images, and they can be a real challenge to clean up. Flat-field frames cannot correct gradients. The best gradient removal tool I have found is a Photoshop® plugin from Russ Croman, the GradientXTerminator. See: http://www.rc-astro.com/resources/GradientXTerminator/index.html Vignetting - Occurs when the width of the telescope, the focuser, the camera window, or some other part of the optical path is narrower than is required to fully illuminate the field of view. In most cases, dealt with effectively using flat-field frames. Filters - (Image processing filters, not glass filters!) There are a zillion image processing filters out there, such as kernel filters in MaxIm DL, sharpening filters (unsharp mask, FFT, etc.), smoothing (Gaussian blur, third-party utilities like NEAT, etc.) and so on. Filters alter the image data permanently, and should be used with discretion. The main advantage of Photoshop® with respect to filters is that you can use selections to 15 apply the appropriate filter to the appropriate region of the image. For example, you can sharpen areas with high signal to noise ratio (S/N), and smooth areas with poor S/N. This idea is the foundation of the Zone System. Layers - A Photoshop® feature that allows you to stack images in interesting and powerful ways. For example, you can select how to blend two versions of an image. You can choose different processing options for each layer in order to bring out different aspects of the image. You can also use layers to apply completely different processing to different objects in an image. You might apply different Curves to M81 and M82, for example. Or you might apply a completely different color balance to a planetary nebula. Blend Modes - Define how layers combine with each other. For example, you can use the Darken blend mode to correct out-of-round stars (see chapter 8 in the printed book). Photoshop® has many blend modes, but only a few are useful for astronomical images. These are Darken, Difference, Luminance, and Color. You may occasionally find use for other blend modes, but these four are the workhorses. Opacity - Determines the transparency of a layer. You can use this to combine two versions of an image and get a result somewhere between the two originals. For example, you might have one version of a galaxy that is very sharp but lacks body, and another version that has excellent density but poor sharpness. Setting the opacity of the top layer to 50% gives you an equal blend of both images. Selections and Masks - Selections are very powerful Photoshop® tools. Masks are similar, and in fact there are times when one would be hard pressed to tell these apart (for example, when you save a selection, it is saved as a mask). Using a selection, you can select a portion of the image to operate on. For example, you can select a galaxy and sharpen it, or select the background and smooth it. There are a large number of selection methods, and you can save and load selections in interesting ways (add them together; subtract one from another; get the union of two selections). Very powerful stuff, and I highly recommend taking the time to learn all you can about selections. Pay special attention to the many ways you can create selections because there will come a time when you will have to get very creative with selections! Since selections apply to both normal and astronomical images, any good Photoshop® book will give you the details you need about selections. Show/Hide Selection - The selection boundary (affectionately known as “crawling ants”) can be very distracting when you are processing an image. The Control + H hey combination will show and hide the selection boundary. (Do it twice if any guidelines are visible; once to hide the guides, once more to hide the selection.) Feathering - Applied to selections. Softens (fades) the edge of the selection. If you feather by 10 pixels, half of the feathering is applied inside the selection boundary, and half of it outside the boundary. In other words, be careful to make your selection large enough to allow for this behavior. Desaturate - To remove the color from an image or layer. Invert - To make a negative of the image. Useful for revealing extremely faint details. 16 Fade - A tool on the Edit menu. After applying a filter (e.g., sharpening), you can use the Edit | Fade menu item to only apply a percentage of the filtered result. Resizing methods - When resizing an image, you generally want to avoid creating artifacts. When enlarging by a big amount, you can get increased pixelation if you don’t choose the correct resize method. In addition to the methods included in Photoshop®, various third-party tools provide excellent resizing capabilities when you have to make huge changes (I’m talking on the order of 500% larger, for example). For most uses, Photoshop®’s built-in resizing methods are just fine. Stay away from Nearest Neighbor; it retains pixelation. Transformations - Includes such operations as rotate and scale. For example, if you take luminance on one night and color on another, you might need to scale, shift, and then rotate the color image to match the luminance image. Merge Layers - Combine two or more layers. For example, you might have created two layers containing the stars from an image in order to remove elongation due to guiding problems. When you are done, you can combine the two stars layers with the original image by merging the layers. Color Range - A very powerful and useful method for creating selections. The selection is based on the color and brightness of the pixels in the image. You can use Color Range to select stars, or the background in an image, or to select just a certain range of brightness in a galaxy. Once you select the brightness range with the Color Range tool, you can then operate on just that portion of the image. A quick tip: when selecting based on brightness, always use the luminance layer to create the selection (create one with Duplicate Layer and Desaturate if you don’t have one handy). Load/Save Selections - The important concept here is that you can save selections and recall them later. I recommend that you save the base form of a selection (i.e., before feathering) so that you can start over later. Modify Selection - Photoshop® allows you to make some interesting modifications to selections. Feathering is explained elsewhere in this list. Expanding and contracting selections are very useful. For example, you can use Color Range to select stars, and then Contract by a few pixels to remove the smallest stars from the selection. You can then Expand the selection back to its original size. Only the larger stars will now be selected. Offset Filter - Used for shifting a layer up/down and left/ right. Often used with the Darken blend mode to correct badly shaped stars, such as elongation from windy conditions. Minimum Filter - A Photoshop® filter that can be used to shrink stars. The filter looks at adjoining pixels, and selects the darker (minimum) value. This effectively cuts a full pixel off the radius of a star, so it can be overkill for small stars. See Modify Selection above for a tip on how to select just the larger stars. (Puzzle for the reader: how would you use Modify Selection to get just the small stars? Answer: Save the selection of all stars, then use the Contract/Expand technique and save the large stars selection. Load all stars, then load the large stars selection by subtraction. This leaves you with just the small stars in a selection.) 17 Photoshop® Palettes - This includes such things as History, Layers, Channels, Histogram, and Info (used for pixel-bypixel brightness info). You can turn palettes on and off using the Window menu. You can also drag and drop palettes from one palette group to another - just click on the tab that has the palette name on it. This allows you to arrange palettes to fit your working style. I like to put Layers, Channels, and History in a single palette group because I use these three the most. I also like to keep the expanded Histogram Palette handy for color image processing. Rulers - I suggest that you turn on rulers (View | Rulers menu item) and set them to pixels. This allows you to judge the size of selections for things like feathering and Contract/ Expand. Note that when you move the mouse cursor around an image, the position is shown by a thin line on the rulers. Actions - You can record a series of Photoshop® steps into an Action. There is an Action palette that displays the available actions. This can be handy for repetitive tasks. To record an action, display the Action palette, then click the Create New Action icon at the bottom. Name the action, and you are automatically in Record mode. When you are done recording, click the Stop button. Underlying and Advanced Concepts Quantum Efficiency - Often referred to simply as QE, this is the efficiency with which a CCD chip converts photons to electrons. All other things being equal, a camera with higher QE is usually a better choice. However, there are often other considerations - antiblooming chips and low-noise chips being two good examples. QE is important, but it’s not the only consideration for all imagers. For example, under really dark skies, you can use the ability of an ABG camera to go really deep without blooming. Try extremely long individual exposures (30–60 minutes). Dynamic Range - The number of steps between the dimmest and brightest values a camera can record. The step size is determined by the read noise, so a camera with lower read noise will have more steps in the dynamic range. The ideal camera has a large dynamic range with small steps. That requires a high full well level and a low read noise compared to other cameras. Noise Reduction - The process of limiting the noise in an image. Since noise is actually uncertainty in brightness values, you can’t reduce noise in a single image (though you can apply smoothing to hide it). The most common form of noise reduction is to combine images. Half Flux Diameter - The radius of a star image at which half the signal is inside the radius and half is outside the radius. Flux in this case refers to the total recorded light from the star. So in simpler terms the HFD is the radius within which half of the light from the star is contained. FocusMax, an automated focusing program, has popularized this measurement. X and Y Axes - Pertains to guiding. If the guider is lined up with the RA and Dec axes, then the X axis is RA and Y is Dec. But nearly all guiding software allows you to rotate the camera to any angle. In that case, X simply refers to the horizontal axis of the guide image, and Y the vertical. 18 FITS header - I recommend storing images in the FITS format because it’s reasonably universal. The FITS format stores information about the exposure in a header at the start of the file. The information stored depends to some extent on the software you use to take or modify the image. Information ranges from exposure time and temperature to the RA and Dec of the center of the image. camera relay. Note: even though the term relay is used, many cameras do not use relays, they use TTL (short for transistor-transistor logic) circuits instead of relays, but the concept is the same: the camera makes the mount move for a short period of time to affect the correction. The Paramount ME from Software Bisque does not need to use relays; the mount can be slewed in very small amounts using DirectGuide. Extinction - Atmospheric extinction refers to the fact that blue light is scattered more readily by the atmosphere. (This is why sunsets appear red - the red light makes it through, but the blue light is scattered). When you image away from the zenith, more and more blue is scattered as you approach the horizon. Blue isn’t the only color affected; as you get lower, green is also scattered more than red. This causes color balance to change as you move away from the zenith. Above about 45 degrees, the difference is small, but below 45 degrees, the difference is large enough that you should increase your green and blue exposures to deal with the problem. Image Scale - Defines how much of the sky each pixel on a CCD chip “sees.” There are many way to express image scale arcseconds per micron, or, most useful for assessing the match between a camera and telescope, in arcseconds per pixel. To calculate the image scale, divide the pixel size in microns by the focal length of the telescope, and multiple the result by 206. For example, consider my half-meter f/8.3 RitcheyChrétien with the STL-11000 camera (9-micron pixels). It has an image scale of 9/(500*8.3) * 206 = 0.446”/pixel. • At 45 degrees, add 5-10% to blue exposure duration. • At 40 degrees, add 10% to blue. • At 30 degrees, add 15-20% to blue and 5-10% to green. • At 20 degrees, add 40% to blue and 20% to green. For a complete breakdown of extinction recommendations, please see the table in chapter 6 of the printed book. Camera Relays - The guide camera needs a way to make guide corrections. It makes the correction by sending a signal to the mount to move. The signal comes from what is called a Reverse X - When guiding with a a German Equatorial mount, crossing the meridian involves a 180-degree flip. This reverses the orientation of the camera. This causes the direction of guiding to reverse as well. Camera control programs usually have some kind of “reverse X” checkbox that will allow you to guide without re-calibrating when you flip the meridian. Just check the box when you cross the meridian, and uncheck it when you slew back. Note: Some software does this automatically. Signal to Noise Ratio (S/N) - This is the measure of how effective a CCD image is. Signal (certainty) is good; noise (uncertainty) is bad. You want lots of signal and as little noise as possible. A large S/N yields an image that is clear and crisp; 19 a small S/N yields a grainy image with indistinct details. How do you get a large S/N? By taking multiple images, and by using the longest practical exposure time for individual images. Hydrogen Alpha Filter - A type of filter that passes the light of a specific emission line associated with excited hydrogen gas. This is one of the most common (and brightest) emissions in many nebula. A Hydrogen-alpha filter is also a good choice for light polluted locations or for moonlit nights. There’s very little H-alpha in light pollution, and only a small amount of light in this wavelength is reflected by the moon. Integer/Floating Point - Types of data storage for CCD images. Most camera control programs support 16-bit integer values, which gives 65,535 brightness levels. Floating point supports a larger range of values, but that isn’t a significant advantage for a single image. Floating point becomes most useful when combining a large number of images using sophisticated statistical combination techniques (typically used for rejection of outlying values such as cosmic ray hits). 8/16/32 bits - The number of bits used to store the brightness of each pixel determines how many brightness levels can be stored for each pixel. For 8 bits, that’s only 28 or 256 brightness levels. You should avoid 8-bit data storage until you are completely finished processing the image—such as when you create a JPG image for the web. With 16 bits, you get 216 or 65,535 brightness levels. That’s more than enough for one image, or for combining all but the largest numbers of images. For 32 bits, the number of brightness ranges is extremely large (232). This is useful for sophisticated statistical techniques and for combining large numbers of images. Planning and Equipment When I was just getting started in imaging, I was very impatient to get started every night. I would do a quick polar alignment—I just set up and imaged. I wound up fighting the mount all night long, losing images to streaking, jumps, drifting and what I thought was just plain bad luck. More out of curiosity than anything else, I decided to see what would happen if I took some time to carefully polar align the mount before imaging. The moon was still out and I thought I’d “waste” some time perfecting the polar alignment. When I started imaging with excellent polar alignment, suddenly my luck took a turn for the better. Image after image was coming through OK. There were no strange bumps to ruin the images, and guiding was dreamlike in its accuracy. I cannot overemphasize the importance of having your equipment in solid working order before you image. You may be impatient to get started, as I was, but in the end you’ll get more productive imaging time if you “waste” some time early on to get your equipment well adjusted and properly set up. Mount Without question the mount is the one key component in the entire imaging chain. No matter how sharp your optics are, no matter how wonderful your camera or telescope is, your mount sets the limit on what the other components can do. This applies to both buying a mount and setting it up—don’t skimp on the most important device in the imaging chain. 20 Stiffness Periodic Error (PE) The more weight you put on any given mount, the more likely it is to flex from the load. The stiffer a mount is, the greater the load it can carry. Some small mounts are surprisingly stiff, such as the Takahashi EM-10 or the Astro-Physics 400. But every mount has a maximum load it can carry before it will flex so much that tracking and guiding will suffer. Periodic error is a fundamental property of your gears. It is periodic because it occurs once for every rotation of the worm. The worm drives the gear that moves the RA axis. If the worm is out of round, the rate of tracking will speed up and slow down as the radius of the worm varies during rotation. (The speed of the gear depends on the radius of the worm—a larger radius will speed things up, and a smaller radius will slow things down.) Even the very best mounts have at least a small amount of periodic error. It is worth noting that both the weight and the length of the telescope should be considered in this context. A long refractor puts a lot of weight (the objective) far from the rotational center of the mount. This puts more stress on the mount than the same weight in a more compact package. Tracking Accuracy Every mount has some measurable error in tracking rates. There are two kinds of tracking errors: random and periodic. The screen capture on the facing page shows a graph of random plus periodic error (red). The blue line shows drift in Declination from a not-quite-accurate polar alignment. Random Error Random errors are due to things like rough edges on the gears, insect nests in the gears, poor finishing of gear, etc. Cleaning and/or burnishing your gears will take care of such problems to some degree, but for some mounts replacement or upgrade of the gears may be required. If the problem is severe, you may need to upgrade to a higher quality mount. In any case, random errors set a limit on how high a resolution you can image at. Total PE Periodic error is measured as the number of arcseconds of error. For example, if the mount speeds up enough to get 10 arcseconds ahead of where it should be, that is a periodic error of 10 arcseconds. The mount also slows down during part of the worm cycle, so the total periodic error is the distance between the slowest and fastest points. Depending on your focal length and the amount of total periodic error, autoguiding may be necessary to get round stars. A high-quality mount will have a worm that is only very slightly out of round, leading to periodic error on the order of 3-5 arcseconds or even less. A good mount, and one reasonably suited to CCD imaging at high resolutions, will have a periodic error of no more than 10 arcseconds. Mounts with PE of 20 or more seconds are a challenge to use for imaging, and mounts with PE over 30 arcseconds may be tough to use for imaging at any resolution. 21 Rate of change of PE Total periodic error is only half the issue. The rate at which the error changes is also a critical point to ponder. If PE changes slowly and regularly, then guiding will be able to keep up with the changes and you can get good results even with fairly large PE. But if the changes in speed are sudden, guiding will not be able to correct for the error fast enough and you will see elongation of the stars in the image. In most cases, a large PE is associated with rapid errors. For example, if the periodic error is 10 arcseconds, and the worm period is three minutes, then a smooth rate of change would be 10 arcseconds in 1.5 minutes (half of the time spent reaching the error 10 arcseconds, and the other half coming back). This is about 0.1 arcseconds per second, or one arcsecond in 10 seconds. 22 If each pixel of your camera sees one square arcsecond of sky, then it would take 10 seconds for periodic error to move a star from one pixel to the next. Any guide exposure up to 10 seconds would be capable of providing adequate accuracy. Instructions for CCDSoft, version 5: • Open Camera Control • Select the Autoguide tab On the other hand, if the surface of the worm is not smooth, and the rate of error changes, there could be times during the turn of the worm when the rate of change is faster or slower than the average rate of 0.1”/second. • Click Settings button Here’s a real world example. An otherwise highly regarded brand of mount showed a total periodic error of 10 arcseconds. However, the rate of change was a serious problem - the mount jumped 8 out of those 10 arcseconds in a 10-second period, for a rate of change of nearly 1 arcsecond per second. The autoguider was not able to keep up, and no matter how carefully we set up the mount, it always showed elongated stars. • Check the box labeled “Log autoguider data...” The solution required a trip back to the manufacturer, who reported that the worm had a significant kink that was causing the sudden jump. Measuring PE Most camera control programs, such as CCDSoft or MaxIm DL, inlucde the ability to record guiding corrections. A simple trick enables you to record the actual PE of the mount rather than the guide corrections—turn off corrections, and turn on the recording capability. The recorded “corrections” will measure the actual movement of the mount during one or more turns of the worm. You can then load the file into a spreadsheet and graph the periodic error (see page 21). • Make sure that the “Enabled” checkboxes for the X and Y axes are unchecked. Instructions for MaxIm DL, version 4: • Open Camera Control • Select the Guide tab • Click the Options button and choose Guider Settings • Make sure that the “X axis” and “Y axis” checkboxes in the “Guider Enables” area are unchecked. Similar options are usually available in other camera control programs. Tip: Square the camera to the RA and Dec axes before you take these measurements. If the camera is square to the axes, the RA and Dec movements will be confined to a single axis when you make a graph of the data. If the camera is not square, the RA periodic error will be partially included on each axis, and it will be harder to sort out the information. During the recording, the movement of the guide star will be recorded to a file. The movement of the guide star will be a combination of many factors, including: periodic error, 23 random mount error, random movements from seeing, drift because of polar alignment error, and so on. Even so, you can usually look at a graph of the errors and pick out the periodic error. The simplest way to graph the data is often to load it into a spreadsheet. I have created an online tutorial that shows you how to do this with a CCDSoft log file; MaxIm DL is similar: http://www.newastro.com/newastro/tutorials/pe/frames.htm The graph on page 21 shows what a plot of a log file looks like. In this example, the dark blue line is a plot of the declination (Y) data, and shows a pronounced drift due to a poor polar alignment. There is also a slight drift in the RA data (pink) but the periodic error is clearly visible. There is considerable noise in this example, and this is typical unless you collect the data on a night of good seeing. If the seeing is good, and you still have a pretty ragged looking PE curve in RA, you are probably looking at mechanical noise in the mount. Tip: You can estimate the seeing noise by looking at the Dec data. The vertical variations in Dec are caused by the seeing, not by the mount’s movement (since the Dec axis is not moving). Tip: If you do not know the worm period, you can deduce it from the graph. Find a repeating sinusoidal shape in the graph (that’s an up and down wave), and look for a pair of matching peaks or valleys. Refer back to the times in the spreadsheet data for those matching events. Subtract one time from another to determine the worm period. Periodic Error Correction (PEC) Once you can measure the periodic error, you are in a position to correct it. Not all mounts support periodic error correction. Those that do offer different methods. Most allow you to record the periodic error, and then save it to the mount’s memory. A few mounts do not save this error information when you power the mount down, but most do. If your mount does support periodic error correction, it probably only allows you to sample the periodic error through one pass of the worm. PEMPro from CCDWare takes multiple samples of the periodic error, calculates a very accurate result, and then uploads correction to the mount. Highly recommended. Please see: http://www.ccdware.com/products/pempro/ A few mounts allow you to not only record periodic error, but to smooth the curve to remove sources of noise. The typical noise sources are seeing and mechanical. PEMPro’s averaging capability provides this feature for all mounts it supports. Seeing noise Seeing noise results from random fluctuations in the atmosphere. You should avoid recording periodic error when the seeing noise is large relative to the size of the periodic error. For example, if your mount’s typical periodic error is 5 arcseconds, and the seeing noise is 4 arcseconds, you should wait for a better night to measure periodic error. Unless you are using PEMPro, measure PE on a night when the seeing noise is no more than one-fourth the size of PE. 24 You can estimate the size of the seeing noise by measuring the variations on the non-driven axis (Declination). However, this will be mixed with mechanical noise so it is only an estimate. If you have access to a DIMM monitor (either using an STV or the web site of a nearby observatory that is making DIMM measurements), you can get a more accurate assessment of the seeing noise. DIMM stands for Dual Image Motion Monitor. DIMM uses two images of the same star in a large aperture instrument (greater than 10”) to measure seeing. The two images move differently, and that difference provides the measure of seeing. Mechanical noise Your mount has some level of mechanical noise that results from small (or, unfortunately, sometimes large) mechanical errors. Dirt on the gears, bugs in the gears, irregularities from machining of the gears, and other mechanical imperfections all contribute to the mechanical noise. Wind shake is another important component - do not try to record PE on a windy night! You can improve the mechanical noise level of the mount by cleaning or modifying it. For LX200 owners, there are many web sites with tips on how to improve the mechanical performance of the drive mechanism and bearings. For other mount types, you may need to search a little harder to find tips. The bottom line, however, is that you can get better PEC with a mount that has low mechanical noise. You’ll also get better imaging performance. Mechanical noise, in fact, is often the key culprit when guiding is challenging. Benefits of averaging One way to reduce the effects of seeing and mechanical noise on your periodic error recording is to average the results. This is how PEMPro provides very accurate results. Some mounts allow you to upload a custom periodic error correction table. If your mount supports this feature, you can record several rotations of the worm to the log file, and then use a spreadsheet to average the results from multiple rotations. Then upload the averaged data to the mount. Averaging of periodic error data cleans up the data by reducing noise. This is really the same thing as reducing noise in images by averaging them together. Backlash When gears meet, there is always some looseness between the gears. The amount of looseness varies considerably from one mount type to another. The very best mounts have vanishingly small amounts of backlash. One example is the Paramount ME from Software Bisque. It uses a spring-loaded worm block to maintain tight contact between the worm and gear. Even if the gear is slightly out of round, the spring loading will allow the worm to adjust as needed. Many mounts have fairly sizable backlash, even some mounts that otherwise have excellent (low) periodic error and minimal mechanical noise. It is very important to know how much backlash your mount has, and what to do to minimize the amount of backlash. I always recommend that you take the time to measure backlash before you try to use the mount for serious imaging. 25 Measuring Backlash There are two ways to measure backlash: by hand, and visually (using a CCD camera or an eyepiece). To measure by hand, simply rock a telescope axis (RA or Dec) with your hands. Can you feel some motion? If yes, then that is backlash, and lots of it. If you can feel the backlash in a system, you have a lot of backlash to deal with. To measure backlash visually, slew the mount in one direction, and then slew it in the opposite direction. Any delay you experience in reversing direction is the result of backlash. A video camera is ideal for measuring backlash; you can comfortably use a timer to determine how long it takes the mount to reverse direction. For accurate results, make these measurements using guide speed. You won’t be able to to accurately estimate backlash if you are moving the mount rapidly using slew or centering speeds. Adjusting Backlash There are two ways to adjust backlash: by using backlash compensation, and by actually changing the way that the worm and gear mesh. Backlash compensation, commonly available in both hardware and software, is the least desirable way to deal with the problem. Backlash compensation works by “running out” the gears to take up backlash. This typically involves running the gears at high speed briefly so that the backlash is quickly taken up. Then the mount moves at the regular rate. This is OK for simple things like centering a star visually, but it is problematic for high-precision things like guiding. The amount of backlash changes with the rotation of both gears, so you can’t just set one value for backlash compensation and expect precise results. It is far better to take the time to learn how to adjust the mesh between the worm and gear. If this step is not detailed in the manual that came with your mount, you will need to contact the manufacturer or consult with user communities such as Yahoo groups to find out how to adjust out the backlash. You won’t be able to adjust out all of the backlash, however. You need to leave some backlash so that the gears don’t bind. It may take some experimentation to find the sweet spot for your mount. Tip: Be careful about how tightly you mesh the gears. A mesh that is too tight can damage the gears! Smallest Practical Movement - How Big Is It Really? Every mount has a smallest move it can make reliably. For some mounts, stiction will prevent the smallest movement from being the same each time during guiding. Stiction is like friction. It describes a situation where the mount doesn’t move easily for short distances. When stiction is at work, the mount sits still while forces build up from guide corrections. When stiction is overcome, the mount breaks free and moves a larger distance than you would like it to. Tuning and lubricating your mount may help minimize stiction, but to some extent stiction is a result of mount design and you can’t reduce it past a certain point for any given mount. 26 Be aware of the existence of stiction, and if you think your mount is not making small moves reliably, try increasing the minimum move size in software. (Different programs implement this in different ways). carbon dioxide is directed at the optic using a special type of gun. Various models are available in the range of $500–2000. The gun makes CO2 snow, which dislodges dust from the optic and then evaporates leaving no trace. Does not work on optics which have become wet from dew. Telescope Distilled water bath - Often used on mirrors, which are removed from their support and placed in a distilled water bath. A stream of carefully dried and filtered air can be used to remove the distilled water after flushing the surface. While it is obvious that the optical quality of the telescope has a bearing on the quality of images, there are a number of other things about telescopes that come into play. Even something as obvious as cleaning gets overlooked, probably because cleaning a telescope optic is somewhat intimidating. Cleaning For most of us, the thought of cleaning telescope optics is a frightening proposition. Whether it is a 10” mirror or a 4” refractor, the fear of damaging the optical elements is always there. This is a well-founded fear, but from time to time it is essential to clean the optics of any telescope. There is some good news, however. It is not necessary to clean telescope optics frequently. Only when the optic has accumulated considerable dust does one need to clean it. For a typical instrument in a typical location, once or twice a year should take care of the problem. There are many methods for cleaning optics, and various methods are best for various types of telescopes. The most common cleaning methods are CO2 guns, and good old H2O. CO2 guns - This cleaning technique was originally developed in the semiconductor industry. A very high-pressure stream of Soapy water bath - Used prior to a distilled water bath for especially grimy optics. A very small amount of liquid soap is mixed with water and flushed over the optic. Cotton swabs can be used lightly for stubborn stains. In all cases, heavy scrubbing is never used because it can scratch the optic. For more information about cleaning, try a Google search using “mirror optical cleaning.” If you have any doubts about your ability to clean an optic, contact the optic or telescope manufacturer for advice or to learn about their cleaning services. Flexure Just as a mount can flex under load, the telescope itself has a number of ways in which it can flex. These include: Tube flexure - The optical tube of any telescope will have some flexure. If the tube is well-designed, the flexure will be too small to notice. Cheaper optics tend to have less stiffness in the optical tube, and are more likely to show a problem. 27 Focuser flexure - Just as the optical tube can flex, a focuser can flex. Focusers typically have a smaller diameter than the optical tube, and are therefore more likely to flex. A focuser specifically designed for imaging will have a larger diameter and use thicker materials to limit flexure. A focuser made for visual use may be too light for photographic use. You can sometimes replace or reinforce a visual focuser to improve performance for imaging. Focuser shift - In addition to flexing, a focuser can also be loose. This is more common on visual focusers. You can sometimes add set screws to lock the focuser in place. The downside is that you’ll have to loosen the set screws to adjust focus. Optical element flexure - The optical elements in a telescope can sometimes move, shift, sag, tilt, or even change shape slightly as you slew the telescope around the sky. This is a particularly troublesome type of telescope flexure because the changes sometimes happen randomly in large chunks. A selfguiding camera is the only solution and even then you will still lose some images to large changes. Optical path shift - For best results, the optical elements in a telescope should be precisely aligned (collimated). A poorly designed telescope may allow the optical elements to move out of alignment during slewing. This takes the telescope out of collimation and can cause such things as coma, bloated stars, loss of focus, etc. If you are using a separate guidescope mounted alongside your imaging telescope, there can be differential flexure between the two scopes. For example, the main telescope might suffer from movement of the primary mirror as the mount tracks the sky, while the guidescope suffers from tube flexure. If this occurs, the guide scope will cheerfully guide on a star but the two telescopes will be slowly moving to different areas of the sky. The longer your imaging focal length, the greater the chance that differential flexure between the imaging and guide scopes will limit the length of guided exposures. Two types of guiding are best for dealing with all types of telescope flexure: off-axis guiders and self-guiding cameras. Loose Parts and Collimation This is another area that seems obvious, but when you are imaging in the field it is all too easy to miss a loose part in the telescope. A loose focuser has already been discussed; this can come from the factory and require considerable time and effort to remedy. But a loose secondary mirror is also fairly common, and it’s easy to look everywhere else for the problem. When optical performance is poor and you can’t seem to find the cause, look for loose optical elements. Other parts of the telescope can also get loose - baffles can shift and cast shadows or cause unwanted diffraction effects, for example. The more you use your telescope, and the more familiar you get with it, the easier it will be to identify when something is loose. Just looking at the images from the telescope can tell you a lot about the state of the telescope. Never assume that bad seeing must be the cause of some trouble— the telescope itself is a delicate instrument, and anything that isn’t optimal will have an impact on image quality. It’s worthwhile to regularly check your instrument’s collimation. Some telescope types, such as refractors, tend to hold 28 collimation extremely well but that doesn’t mean you should ignore collimation. Other types, such as converted Dobsonians, may require frequent collimation, even several times in one night. The more solidly your optical elements are mounted, the more likely your telescope will retain collimation. For most telescopes, there is a wide range of pixel sizes that are suitable. To learn more about pixel sizes and matching camera to telescope, please download my CCD Calculator program: Different types of telescopes are collimated in different ways. For refractors, if you do have a collimation issue, a return to the manufacturer is usually your best bet due to the special equipment required. Cassegrain family telescopes can have two collimation adjustments, one for the primary and one for the secondary. In some cases, as with the popular SCTs from Meade and Celestron, only the secondary adjustment is available. For Mak-Cass scopes, only primary collimation is available. Cleaning Camera The most common surfaces to clean are the optical window of the camera, and the filters. Sometimes it will also be necessary to clean surfaces inside the camera. Surfaces inside the camera can be tricky, especially if the camera exposes the CCD chip (some chips have a cover slip, some do not). You may want to simply send the camera back to the manufacturer for interior cleaning. Camera choice plays some role in how effectively you can use a camera with a given telescope. Pixel size is a key consideration. You want a camera that has pixels that “see” an appropriate chunk of sky. If the pixels are too small, the chunk of sky will be smaller than what your seeing supports, and soft fuzzy images will be the result. Worse, if the pixels are too small, the light from a given source is spread out across many pixels and you’ll need much longer exposures to get a decent result. On the other hand, there’s no harm in having pixels that see a very large chunk of sky. This is common when doing widefield imaging. You won’t get optimal resolution, but you will get beautiful widefield images. http://www.newastro.com/ibp/ccdcalc Cameras (and the filter wheels typically used with tem) have multiple optical surfaces that also need to be kept reasonably clean. Unlike telescope optics, which are typically far enough away that casual dust is not a problem, the optical surfaces in the camera are often close enough to the CCD chip to cause identifiable shadows. A flat-field can be taken to remove the effects of dust on these surfaces, but if you have large dust motes or too many dust motes, cleaning is in order. Consider obtaining the cleaning kit from Custom Scientific. It includes cleaning aids and solutions that are specifically designed for cleaning optical surfaces. For more information: http://www.customscientific.com For casual cleaning of loose dust, the canned air available for cleaning electronics is a useful alternative, but not without 29 some risk. The air expands as it leaves the can, and cools. And if you do not orient the can correctly, propellant can come out with the air. In either case, thin optics (like filters) can crack from the sudden temperature change. Use with caution! Clean your camera’s optical surface whenever flats show excessive amounts of dust or large-sized dust motes. Large motes cast strong shadows that are difficult to clean up with flats. take at least ten 5-minute darks. Combining images lowers noise, so if you combine light frames, you should do at least as much combining for your darks so that they are no noisier than your light frames. This does not mean that there are no benefits to taking more darks than lights; more darks means less noise in the final image. The total noise adds quadratically. The equation for total noise from light and darks frames looks like this: Noise Control CCD cameras are an electronic device, and that means each camera has some level of noise. Noise is a complex, often non-intuitive subject that is addressed in many places in the printed book. But there are some simple procedures you can follow to keep the main sources of noise out of your images. Other noise sources (e.g., bias, flats) also add quadratically. For best results, take your darks with the same exposure time and temperature as your light frames. You can scale darks using bias frames. But your results will be more precise, and lower in noise, if you use same-time, same-temperature darks. Reduction Frames Darks are prone to cosmic ray hits, which cause bright spots and streaks in the dark frames. When you apply a dark with cosmic ray hits to a light frame, the bright spots become dark spots because those pixels are over-corrected. Combining multiple darks reduces artifacts from cosmic ray hits. You should also choose a combine method that is effective at removing cosmic ray hits, such as median, min/max clip, or sigma reject. See Combine Methods below. Also called calibration frames, these include dark frames, flatfield frames, and bias frames. You can find specific descriptions of what these frame types are and how to take and use them in my earlier book, The New CCD Astronomy. This section describes the best way to use these frames to control noise in your images. Dark frames Dark frames record the thermal noise of your camera for a given exposure time and chip temperature. The first rule of thumb for darks is to take at least as many darks as light frames. If you take ten 5-minute images, then If you want good results, plan on taking at least eight darks. But don’t hesitate to take 16 or 32 darks if you seek the best possible results! You don’t need to take all those darks at once. Take them in groups, perhaps eight each night for several nights. Taking darks over several nights helps average out hot pixels more effectively. 30 Bias Frames Bias frames record the fundamental thermal noise of your camera before an exposure starts. If you follow my advice for dark frames, you won’t need bias frames. Bias frames are used only when you scale a dark frame. For example, if you have a 10-minute dark frame, and want to scale it for a 5-minute image, your software will subtract the bias from the dark, scale by 50%, and then add the bias back in. The result is a dark that is approximately correct for a five-minute exposure. Scaled darks are in fact only approximations, so whenever you can, take darks of the same exposure time and temperature as your light frames. Darks scaled for exposure time are reasonably accurate in a pinch. Changing temperature, on the other hand, is risky and you should not expect clean results. The reason for the difference: thermal noise changes nearly linearly for exposure time, but non-linearly for chip temperature. As with darks, take plenty of bias frames. The exposure time on a bias is nearly zero, so take plenty of them. Bias frames add noise, so it’s reasonable to take 30–100 for best results. Flat-field Frames Flat frames record the optical obstructions of your camera and telescope, such as dust on optical surfaces and vignetting. Dark and bias frames are easy to take - just set the CCD temperature and the exposure time, and hit the Take Image button. Flats are a different story. Flats require significant operator experience and skill. The brighter your skies, the more important it is to take flats. This applies whether the brightness comes from light pollution, a bright moon, or high moisture content in the air. The average brightness of the brightest part of the flat should be roughly half the saturation level of your camera. You can get saturation levels from the camera manufacturer. In some cases, you can calculate the saturation level from the camera specifications: divide the full well capacity by the gain. For example, for the ST-10XME, the full well is 77,000, the gain is 1.5 electrons/ADU, and the saturation level is 77,000/1.5 = 51,300. A good average value for a flat would be 25,000 to 30,000. Three ways to take flats, in order of increasing complexity, are: Sky flats - There is a brief window of time at dusk and again at dawn when the sky is dim enough for flats but still bright enough to allow you to take flats with short (e.g., 2–20-second) exposures. Just point your telescope near the zenith and take flats. Some tips for success: • Turn off tracking on your mount so the light from stars is spread out in lines. • Darks are not essential for good flats. The exposure time is so short that the signal to noise ratio is extremely good. Adjust exposure time to make sure that the average brightness of the flats is as similar as possible. • Make sure your exposure time is at least a few seconds. Otherwise, you may wind up with subtle irregularities in the flats from uneven illumination as the shutter moves. 31 • You can also use a diffuser (see next section) to remove the stars from sky flats. If you do not use a diffuser, take at least 8–12 exposures and use a statistical combine method with normalization turned on so that bright stars will not show up in the final master flat. Programs like CCDAutopilot from CCDWare make the taking of sky flats much easier. Diffuser flats - A diffuser can be anything that spreads light out evenly at the front of the telescope. A T-shirt, a sheet of milk plastic—just make sure that whatever you use doesn’t leave bright spots. Also make sure that the diffuser is evenly illuminated. Things like shadows from moonlight will create an uneven flat, as will uneven thickness in the diffuser. Light box flats - Plans for various light boxes are available on the internet. The general idea is to put some very weak light bulbs behind a diffusion screen (some designs use reflection and/or multiple diffusers). I’ve never used a light box because it is extremely challenging to get even illumination. I use strictly sky flats. The test of any flat is in how well it works. If a flat doesn’t correct shadows from dust motes, leaves dark holes in the image, or is otherwise flawed, you’ll have to put on your detective hat and try to figure out what’s wrong. The problem is usually some form of uneven illumination. The exception: if you make flats in the daytime, light may be entering the back of the telescope and ruining your flats. Sometimes you need to make a set of flats for every filter. Take a test flat with each filter. If you can see differences in dust mote shadows or vignetting, then it’s a good idea to take separate flats for each filter. If you see little or no difference, then it’s OK to take just a single set of flats with the Clear or Luminance filter. Generally speaking, if your color filters are farther from your CCD chip, the dust shadows will be softer and less likely to require separate flats. Tip: You can make binned flats from your unbinned flats using the Resize feature of your camera control program (e.g., CCDSoft or MaxIm DL). In some cases, you’ll need to adjust parameters in the FITS header to tell the program that this is really a 2x2 flat, not 1x1. For example, change the XBINNING and YBINNING keywords from 1 to 2. Combine methods How you combine darks, flats, and bias frames is very important. Reduction frames have noise just like light frames do. If you don’t use the right combine method, that noise will get into your images. While averaging yields the best overall S/N, it does nothing for outliers (cosmic ray hits, for example) and so I recommend taking large numbers of bias, dark, and flat-field frames so you can use statistical combine techniques to reduce both noise and outliers. That means taking at least eight of each type for best results. For bias and dark frames, use at least a median combine to get rid of cosmic ray hits and stray hot/cold pixels. Do not normalize bias and dark frames; that could lead to invalid results. For flat-field frames, on the other hand, normalization is nearly always desirable. Depending on the method you use for making the flats, there may be variation in brightness levels. Normalization will bring all flats to the same level and allow 32 you to combine effectively. Combining with median or fancy statistical methods without normalization can lead to invalid results. Some programs have a “Normalize” checkbox for flats that will make your job easier. The image at right is a frame from the Hubble telescope. It is a raw frame with color added to emphasize the high level of cosmic ray hits. Many of the statistical combine methods were pioneered to deal with such extreme cases as this. Color Exposure Times If you are looking for one thing you can do to make better color images, look at your color exposure times. Most imagers are using color exposures times that are not just too short, but way too short. The result is that even if the main object has good color, the rest of the image may not. 33 Most color imaging is done using the LRGB technique, and with good reason. Although LRGB is a compromise, it’s a very good compromise. Typically, the luminance information is captured unbinned, and the color information is captured binned 2x2. This matches the way that the human eye works: we see details in luminance at a certain resolution, and we see color at a lower resolution. There are many image storage techniques that take advantage of this difference, of which one is LRGB imaging with CCD chips. Tip: The LRGB technique works best when the resolution you image at is well matched to the seeing conditions. Otherwise, after upsizing the color data, you can wind up with stars that are too large. Such large stars will not match the resolution of the luminance data. The better your seeing, the less likely you are to encounter this difficulty. If it does occur, try selecting just the larger stars and applying a Minimum filter. You can also try imaging the L data 2x2 and the RGB data 3x3 so there is less of a size difference between the data sets. Whether color or luminance, a good image is one with low noise. Noise creates a grainy appearance. By binning the color frames, you increase the signal-to-noise ratio (there is just one dose of readout noise for the combined pixel, as opposed to four doses of readout noise for the four unbinned pixels). Noise is noticeable in the luminance image as graininess, but noise has a different effect on LRGB color. Noise in the color channels weakens the intensity of the color relative to the luminance channel. This can be overcome by matching the noise levels in the luminance and color channels more closely (by taking a longer total exposure in each color channel). In simplest terms, if you want richer colors, you need longer total color exposure times than you may be used to taking. For example, I commonly see images where the luminance data might have low noise due to long total exposure time (e.g., 30 or 60 minutes). But the color data may have total exposure times of 5-10 minutes. This results in a mismatch in noise levels, and the color looks washed out. Here are some simple guidelines for matching color and luminance data. For more details, please see chapter 6 in the printed book. In all examples, only luminance and red exposure times are shown. I am assuming that the green and blue exposure times will be proportional as recommended for the filter set and camera you are using. For the sake of argument, let’s assume that you are taking four 10-minute luminance images unbinned and that color images are taken binned 2x2. Weak color: Luminance 4x10min, color 1x10 min. Adequate color: Luminance 4x10min, color 2x10 min. Good color: Luminance 4x10min, color 4x10 min. Rich color: Luminance 4x10min, color 6x10 min. Superb, deep color: Luminance 4x10min, color 8x10 min. My choice of adjectives should give a good idea of the differences, but it is also useful to consider how each level of color richness affects the appearance of the image. Weak color: Dim objects will have no color at all, or if you stretch the image too far, dim objects will show unpleasant 34 and extremely grainy color effects. You will not be able to show all of the luminance details because the dim areas will look bad because of the color noise. In bright areas, you’ll need to hold back on the luminance processing or the color will appear very weak and washed out. Weak color inhibits your luminance processing to a frustrating degree. Adequate color: Color in the major bright objects will be acceptable, but dim objects will have only faint or noisy color. You will need to be careful in processing the color data not to reveal too much noise in the dim areas, and luminance data must not be pushed as hard as it could be to prevent washing out the color. Good color: Even dim objects will have at least a bit of clean color. Only the brightest areas will have washed out color, so you can process the luminance almost as you would for a luminance-only presentation. Rich color: Dim objects will have clear color, and only the very darkest areas will have noisy color or lack color. Pay careful attention to background color balancing and color bias removal with rich color. Superb, deep color: Every object, no matter how dim, will have clearly recognizable color. Bright areas will have eye-popping richness of color. The luminance layer can be processed to the fullest extent possible, and still retain good, solid color throughout in the combined LRGB image. You’ll note that these color imaging times are much, much longer than you typically see. Rich color requires a lot of data! It is low noise in the color layer that allows you to get rich colors throughout the image, and to obtain clear colors for dim portions of the image. Tip: Although 4x10 minutes luminance and 8x10 minutes color will yield good results, doubling these exposures times will do even better. You shouldn’t assume that just because the binned color images are twice as long as the luminance you will automatically get exceptional color results. Depending on the brightness of the objects in the image, much longer luminance exposure times (and therefore much longer color exposure times) may be required. You can’t go too long! Of course, the darker your skies, the more productive your results will be with longer exposures. Accessories The accessories you choose impact your imaging capabilities. Things like filters and focus wheels are not automatic choices, and they add complexities to the imaging process that need to be taken into account. Filter Wheel When I first started out CCD imaging, a filter wheel was one of those things that didn’t even come into play. The whole idea of color seemed like something for another day. I imaged for almost a year before I got a filter wheel. These days, it’s much more common for imagers to start with both camera and filter wheel (or with a one-shot color camera), ready to do some color imaging. Unless you are using a one-shot color camera, you can simplify your learning process 35 by taking on color only when you have a good grasp of monochrome imaging techniques. As simple as filter wheels seem at first glance, they still increase complexity. And there are a few key decisions to make before you choose one. In some cases, the filter wheels are integrated into the camera. Such cameras range from the tiny (and no longer made) ST-237 to the 11-megapixel STL11000. For many cameras, the filter wheel is an add-on choice. There isn’t a single filter wheel that has all the ideal options, but here’s a list of key points to consider when choosing a filter wheel for your camera: Number of filters: I would recommend always getting a wheel with at least five filter positions. This allows for an LRGB filter set plus a clear filter, a useful combination if you are imaging with both refractive and reflective telescopes. Or you might stick with a straight LRGB set plus a narrowband filter such as a hydrogen-alpha filter. Wheels with eight positions are useful if you plan to do both LRGB and serious emission-nebula imaging with multiple narrowband filters. I like an LRGB set plus h-alpha, SII, and OIII. The eighth position can be used for clear, or maybe a continuum filter to match the h-alpha filter. (A continuum filter is used to take images that allow you to subtract continuum data from the narrowband data, providing higher contrast and greater detail—with significant sacrifice of “pretty picture” qualities.) Size of filters: Most filter wheels are set up for 1.25” filters, but this is changing as more cameras come with larger CCD sensors. If you are using, or plan to use, a camera with a larger chip, then you’ll need 2” filters. There is a huge price difference in both the wheels and the filters between these two sizes. And an eight-slot wheel plus all eight filters raises the ante significantly. Interchangeable wheels: Some filter wheels allow you to pop the filter carousels in and out quickly, without removing the wheel from the telescope. This is handy if you are using a larger number of different filters. If you are imaging remotely, and can’t change carousels easily, then a filter wheel large enough to hold all of your essential filters is the best choice. Quality of included filters: Don’t neglect the quality of the filters included with the wheel. If you aren’t happy with filter quality, but like the wheel, consider purchasing just the empty wheel. You can get quality filters from several sources. Filter set design: Quality isn’t the only consideration—different filter sets have different designs with respect to the amount and range of wavelengths that the individual filters pass. Some filters are better suited to emission objects (e.g., planetary nebulae, HII regions) and others are better suited to full-spectrum objects (e.g., galaxies, globular clusters). The differences are usually small, however, and you are unlikely to experience insurmountable problems if you choose a filter set that doesn’t match all of your imaging goals. If color imaging is very important, spend some time learning about the different philosophies and designs of the major filter sets available to amateurs to see which one best fits your plans. Position sensing: This might seem like an obscure topic, but it’s anything but. Some filter wheels use infrared light to determine the position of the filter wheel, and you’ll need to be 36 careful about timing issues to make sure the light is out when the imaging starts. Other wheels use magnetic (e.g., Hall effect) sensors to determine the position of the wheel. Cleaning Filters typically sit right in front of your camera, and any dust on the filters will cast shadows on your images. While a flat will clean up most such shadows, it’s still better to keep your filters reasonably clean. Large dust motes can cast shadows so strong that a flat will not fully correct them, so an occasional cleaning will keep you out of trouble. Custom Scientific sells a complete optical cleaning kit, but lens cleaner or even a light breath on the lens plus a Q-tip does a reasonable job. It’s really the big chunks of dust that you want to keep off your filters. Choice of Filters First and foremost, please recognize that there are visual filters and photographic filters. Visual filters are often made to a relaxed standard that is not sufficient for photographic use. If you already have color or other filters that you use visually, do not assume that you can also use them for CCD imaging. Quite the contrary—it is unlikely you will get good results using visual filters for photography. Common flaws in visual filters include rough surfaces (scatters light, reducing contrast); non-parallel surfaces (creates astigmatism, so that not all portions of the field will be in correct focus at the same focus position); and leakage of nonvisual wavelengths (especially infrared, which causes colors to wash out in RGB imaging). Visual filters can also suffer from poor optical quality, creating a variety of aberrations in your images. The most common filter set purchased for CCD imaging is the four-filter LRGB set. But even this obvious choice isn’t nearly as obvious as it seems. Consider how the two different types of optical systems (refractive and reflective) bring an image to focus. Refractive systems, those that bend light with a lens or corrector plate, do not bring all colors of light to the exact same focus point. The quality of the refractive optic determines how different the focus points are for different colors. Reflective systems, on the other hand, have no such effect. The effect has a name: chromatic aberration (see illustration below). The result is that when blue is in focus, red will not be in focus. No matter how carefully you focus, the image will always be soft. You can compensate for this to some degree with color filters by refocusing for each color. 37 If your telescope has chromatic aberration, focus is always a compromise through a clear filter (or without a filter). Even with high-quality refractive optics, the non-visual portions of the spectrum can be poorly corrected. Since CCD chips are sensitive to these non-visual wavelengths, non-visual wavelengths that strike the chip could be out of focus. For this reason, many of the makers of LRGB filters offer two different versions of the white-light filter. By common usage, these are called Clear filters and Luminance filters. A Clear filter passes visual and non-visual wavelengths, and is ideal for reflectors. A Luminance filter blocks IR (and sometimes UV as well) and is ideal for refractive systems. What, you may ask, about systems that combine reflective and refractive optics, such as a Schmidt-Cassegrain (SCT)? The chromatic aberration of such telescope is intermediate between fully refractive and fully reflective systems. Typically, the stronger the curvature of the refractive elements, or the smaller the number of the refractive elements, the more likely you will be dealing with noticeable chromatic effects. (Multiple refractive elements can correct for this effect, such as on the newer versions of the Takahashi Epsilons.) Focuser After the mount is working well, nothing matters more to CCD imaging than focusing. If you are on a budget, make sure you allocate enough of your available funds to a quality (and convenient) focusing setup. Focusing manually is challenging. But even if you invest heavily in motorized, automated focusing, you still need to have good skills for evaluating the quality of focus. Software makes focusing mistakes now and then, and only the operator can make sure that the focus is as good as it can be. Focusing is hard because the closer you get to accurate focus, the more challenging it is to see the small differences in focus quality. Some software, such as FocusMax, attempt to get past this by measuring the change in focus quality on both sides of focus, creating a graph that points right at the best focus position. But in the end, I always check visually to confirm that focus is as good as it can be. Manual Focusing Manual focusing seems to be on the way out. Even so, you can spend anywhere from $300–1000 on motorized focusers, so manual focusing is a requirement for many budgets. It is extremely useful to have good manual focusing skills. The best way to check focus visually is to look for the dimmest stars in the image, and to watch for the focus position where they pop into view. Whatever the quality of focus, the dimmest stars simply disappear when you are not in focus. This allows you to know more surely than by any other method when focus is at its best. Someone, someday, may write a computerized focus routine that simply counts stars: when you get the largest number of stars, you are focused. (See the example on the next page.) If you do plan or prefer to use manual focus, consider a dial indicator as an aid. The plunger on the indicator bears against a moving portion of the focuser, and you can get accuracy down to a thousandth of an inch without spending a fortune. 38 A dial indicator is invaluable for making the small tweaks that get you to an accurate focus position. In fact, the hardest part of manual focusing is those small tweaks! Motorized Focusing As valuable as manual focusing skills are, there’s simply no question that motorizing focus makes things much simpler. Even if you are not using computer-assisted focusing, simply motorizing the focuser does two things. It makes it much easier to make those small tweaks, and it allows you to focus from a distance using cables to control the motorized focuser. Some motorized focusers include an encoder that reads out the exact position of the focuser. This is an aid to making small critical-focus adjustments. An encoder also takes you one step closer to computer-assisted focusing. Without an encoder, you can make small adjustments by making very quick button presses and hoping for the best. Computer-Assisted Focusing The easiest way to focus by far is with a computer-controlled motorized focuser. Both you and the computer are evaluating the quality of focus, but the computer does all the work. 39 The software that controls the focuser typically uses one of two methods to identify the best focus position. Some software measures focus quality at best focus, but as noted earlier FocusMax uses the rate of change in focus quality away from best focus to predict the best focus position. Whatever the method, you’ll find that computer-assisted focusing is reliable and fast. There are now several different choices available, both in hardware and in the software that does the focusing. In some cases, you can even automate the imaging process with regular re-focusing between exposures. This allows you to get some sleep while your telescope and camera collect images through the night. Computer-assisted focusing is good enough at this point that you will lose very few images to poor focus, if any, in a given night. Non-Focusing There are some telescopes that need little to no focusing, even when the temperature changes. Such telescopes are made of materials that don’t stretch or shrink very much as the temperature changes. Such materials are said to have a low CTE (coefficient of thermal expansion). The most commonly used material in amateur telescopes that has low CTE is carbon fiber. It is used in truss rods and in tubing. It has such a low CTE that even a temperature shift of ten or more degrees will have little or no impact on the quality of focus. When working with such telescopes, you can focus and forget about refocusing. I use an RCOS 20” RitcheyChretien truss telescope and I focus a few times a month. Another material, used more commonly (but not exclusively) in professional instruments, is called Invar. It’s a metal alloy that also has an extremely small CTE. Instrument rotator Also known as a camera rotator, this device changes the rotation of the camera without affecting focus. It’s wonderful to be able to recompose a shot without having to refocus. Rotators come in both manual and motorized versions. Some can be controlled remotely from a computer. Manual versions, such as the rotator that comes built into the Takahashi FSQ-106, are not as convenient as motorized rotators but when they are built this well they are still a very valuable tool to have in your arsenal. As I write this, motorized rotators are becoming more common. RCOS offers a very expensive rotator for their RC scopes, and two sizes of rotator are being offered by Optec. Keep an eye out for new entries in this niche—a rotator is a wonderful toy because it allows you to rotate guide stars into optimal position. When combined with an active otpics device (e.g., the AO-L from SBIG), high-speed guiding becomes common, enabling you to consistently get round stars. Setup The process of setting up your equipment for imaging is often overlooked as a way to improve the quality of your results. But small things add up, and following good practices during setup can both avoid disappointment and improve your ultimate image quality. 40 There are two types of setups - temporary (one or several nights), and permanent (observatories). The key difference between them is that when you are setting up an observatory, it’s worthwhile to spend a lot of time on refining and perfecting alignment. For temporary use, the goal is to strike a balance between the time spent on setup and the quality of the setup. Setup - Temporary By far the biggest factor in the quality of a temporary setup is the accuracy of polar alignment. Guiding, especially, will be better with a careful and accurate polar alignment. Polar alignment When I was just starting out imaging, I was very impatient. I wanted to start imaging as quickly as possible. So I would not polar align very carefully, and I would begrudge every minute spent aligning instead of imaging. The consequences of those quick polar alignments turned out to be severe. Images were lost due to excessive guiding errors, and I was spending way too much time finding objects and putting them on the chip. As things turned out, I actually took more images and was able to image more objects by spending more of my time at the start polar aligning. When I spent an extra 10-20 minutes polar aligning, I stopped wasting an hour or more each night on poorly guided images and lost objects. The lesson here is a simple one: time spent polar aligning is time well spent. I’m not saying you should spend hours polar aligning for a single night of imaging. I’m just saying that you should take the time needed to achieve and confirm a quality polar alignment. For most telescopes and cameras, a polar alignment that is within 2 arcminutes of the pole is going to deliver very good results. For very long focal lengths (>3000mm), an alignment within 1 arcminute is desirable. I like to use products such as TPoint to accurately measure polar alignment, and it is perfect for critical applications. Products like PoleAlignMax are excellent for quick but reasonably accurate alignment. Attaching an External Guidescope This is a huge item to pay attention to. If you are going to use an external guide scope, be prepared for one of the more challenging experiences of the CCD imager. The longer your focal length, the greater the challenge will be. Many imagers expect to simply attach a second small telescope, slap on a guide camera, and get good results. That may be true if your imaging telescope is a 4-5” refractor and your guider is a small 2-3” refractor. But as the focal length of your imaging telescope gets longer, the issues with external guidescopes start to show up in your images as trailed stars. The source of the trouble is both simple and complex at the same time. It is simple to describe—the two optical systems can flex by different amounts. When they do, guiding errors are the result. The flexure comes in many forms. It might be as simple as the weight of the guide scope causing a slight misalignment, or 41 it might be that the focusing mechanism on one or the other scope has some play in it. Your mirror or optics might slip or move during the exposure. The weight of the camera may bend the focuser tube on either or both scopes. Any and all movements, flexures, shifts, and misalignments can result in differential flexure between the two optical systems. The cures for this are as diverse as the causes. If the focuser tube has some play in it, add some set screws or otherwise find a way to lock down the focuser. If the mirror or optics shift, find a way to lock them down. Placing the Tripod Placing the tripod legs or pier base is very important. You want to situate the legs or base so that they will not move. There are some non-obvious ways in which one can invite trouble, however. You might expect that putting the legs of a tripod on concrete would be a great way to prevent losing polar alignment over the course of one or more nights. Not so! Concrete may be smooth enough to allow the legs to shift when you are making adjustments to the equipment. Worse, concrete can act like a drumhead and transmit your footsteps to the telescope and create strangely shaped stars from the transmitted vibrations. Putting a tripod or base into grass, loose soil, or gravel can all lead to trouble. Pay attention to how well anchored you are. Pick up a set of refractory bricks at a fireplace store—they are light, and work well for isolating your equipment from the perils of soft ground. For lighter setups, anti-vibration pads (readily available from most astronomy outlets) help stabilize your equipment during imaging sessions. Setup - Permanent (Observatory) A permanent observatory not only frees you from nightly setup duties, it gives you an opportunity to perfect your equipment setup in ways that are often not practical for temporary setups. Polar Alignment For example, a permanent observatory makes it worthwhile to spend multiple nights adjusting and checking your polar alignment. One of the under-appreciated nuances of a good polar alignment is that for heavy setups it takes time for the stresses to work themselves out after you make a polar alignment adjustment. For example, if you raise the elevation of the mount, it is ideal to go through at least one heating (day) and cooling (night) cycle to work out any tension remaining from the adjustment. If you are brave, you can use a wooden or plastic mallet on the pier to help release stresses, but I like the efficacy of a diurnal heating and cooling cycle to really work out the kinks. You can also pay attention to such things as how well-tightened the pier bolts are. It’s a good idea to tighten them down, then come back a day or a week later and tighten them again. Over the process of several days, you can fine tune your polar alignment to within 0.5 arcminutes or less if you work care- 42 fully and allow time for things to settle adequately. Like fine wine, a polar alignment requires time and experience. Supporting the Pier How much support does a pier really need? When you consider that the really big telescopes are designed with piers that weigh 500 to 1,000 times as much as the telescope and mount, it’s obvious that you cannot have too much support! Foundation for pier Support works from the bottom to the top. A broad base is always a good idea for the pier. For example, if your pier is three foot thick, then you might want to start with a four-byfour-foot base at least a foot or two thick below the pier. And of course the pier base should always be at least as deep into the ground as the local frost requirements. A broad, thick base supports the pier far better than simply drilling straight down. You should also compact the earth under the pier so it doesn’t settle and cause misalignment down the road. If you can connect the pier base with bedrock, that’s even better, but most locations do not offer that. Rods can be drilled into the bedrock to tie the concrete base to the rock. Physical separation of pier from observatory The pier should be isolated from the observatory. In other words, you’ll need to make the pier strong enough to stand completely on its own without support from any part of the observatory. This prevents transmission of vibrations from the observatory to the pier. If the lowest part of the observatory is open (a good idea; the less the observatory blocks the wind, the better), this exposes the pier to the wind. If the pier is very tall or thin, you may need a wind break around the pier. This also keeps solar heating of the pier to a minimum. If there is any wood in your observatory, keep in mind that it may bend or warp over time. Make sure that you allow enough clearance so that the pier won’t wind up eventually touching the observatory floor framing. It isn’t only critical to separate the pier physically; you should also separate it thermally. A concrete pier, especially, will retain heat and release it for a long time. Insulate any part of the pier that extends into the observatory space. Two inches of insulation is definitely not too much. Ventilation Temperature differences are the bane of the imager. If different surfaces have different temperatures, they transfer that heat to the air at different rates. This creates convection currents which create dome seeing problems. The cure is simple: ventilation allows natural air currents to push out the convection, and it also acts as a cooling fluid to carry off excess heat. Louvers, even without a fan, are very effective at allowing air to move heat out and break up convection cells in the observatory. But a fan that moves at least a few thousand cubic feet per minute (CFM) is a very good idea for just about any observatory. Start flushing out the heat a few hours before you plan to observe. 43 Software Camera Control Programs Once you have all of your hardware working properly, you’ll need to make sure that the software that controls it is also in good working order. Many problems with hardware aren’t really hardware problems at all. Proper setup and configuration of the software can eliminate problems with your hardware. Camera control programs like CCDSoft and MaxIm DL use the manufacturer’s software drivers to talk to the camera, and make it easier for you to take images and process them. It’s important to choose the correct camera driver (as well as the drivers for filter wheels, focusers, and other hardware) for each camera control program. The names of the drivers can be different from one camera control program to another. There is so much change in this area that your best bet is to join and monitor the Yahoo group for the software vendor, or join the vendor’s private support service. You can ask questions about properly setting up the control program for your hardware, and learn about driver and software updates as well. Drivers Most hardware talks to the computer through driver software. This is a bit of software (usually provided by the manufacturer) that allows other programs to talk to the hardware. For example, SBIG provides camera drivers that are used by MaxIm DL, CCDSoft, and other programs to control SBIG cameras. Other camera manufacturers also provide drivers that serve the same function. Mounts, focusers, and other hardware also often comes with drivers. With very few exceptions, you want to always have the latest drivers installed on your computer. Keeping up to date with drivers is one of the best ways to solve or prevent problems with your hardware. Failing to install the latest drivers is a very common source of trouble! Your camera manufacturer’s web site will have instructions on where and how to download and install the latest drivers. You should also keep an eye out for new drivers for other hardware - filter wheels, focusers, mounts, and so on. New drivers not only solve old problems, they also introduce new (and sometimes very useful) features. And they can create new problems, so keep the old driver handy, just in case. You can search for Yahoo groups at: http://groups.yahoo.com Telescope Control Programs Although some camera control programs can also control some telescopes directly (e.g., MaxIm DL), most imagers use a planetarium program as the primary telescope control program. My favorite is TheSky from Software Bisque, but there are others in common use as well. MaxIm DL can talk to the telescope through most of the planetarium programs, including TheSky. CCDSoft is designed specifically to work closely with TheSky, and there are some nice advantages to this arrangement. But if you have a favorite camera control program you can choose to keep using it if you are willing to handle the required setup and configuration. The software vendors provide the information you need to do this. 44 Telescope configuration is usually pretty simple, but occasionally serial port issues will crop up and make communication difficult. The software vendor will likely have seen and dealt with these issues in the past, and may well have web pages that describe how to troubleshoot serial communication problems. USB to serial devices are also a potential stumbling point. If you can’t solve the problem, consider getting a different brand of USB to serial converter. This simple change solves the problem more often than not. Pointing Software Telescopes and mounts have a limited ability to point accurately. Things like mount flexure, telescope flexure, lack of orthogonality between mount axes, polar misalignment, and other factors interfere with accurate pointing—even for some very expensive, high-quality equipment. If your camera has a large enough chip, or if your telescope has a short enough focal length, you may still be able to slew to objects and get them on the chip on the first try. But if you can’t find objects readily, or if you simply want objects landing closer to the center of the CCD chip, there is software available to increase pointing accuracy. I have used TPoint from Software Bisque for this purpose for many years. MaxPoint from Diffraction Limited is also available. TPoint has a deep set of features that allow high precision pointing, but MaxPoint is easier to use. The longer your focal length, the smaller your chip, or the greater your desire for perfection, the more likely it is that TPoint is the better choice. Image Processing Software Many camera control programs include at least a few image processing features. Some, like MaxIm DL, include a wide range of processing features that are designed specifically for astronomical image processing. Others, like CCDSoft, include some image processing features and will require you to finish the processing in another program, like Photoshop®. Although you can do some really nice processing in MaxIm DL, at some point you may find that you need more image processing power or convenience than MaxIm DL offers. The answer for most users is Photoshop® CS/CS2. That is why Photoshop® is the primary focus of this book. It has not only the processing tools you need, but they are implemented in ways that combine both power and accessibility. There are some clones of Photoshop® that are far cheaper, such as Picture Window Pro and Paint Shop Pro. While such programs put a fair amount of power into the astro imager’s hands, they are not yet up to the level of power or convenience found in Photoshop® for the most part. Photoshop®, specifically Photoshop® CS/CS2, is by far the best fit for astronomical image processing—as long as you can afford the high price. If your budget won’t stretch to Photoshop® CS, consider either picking up an older version of Photoshop® on eBay, or give one of the Photoshop® clones a try. Special mention goes to Mike Unsold’s Images Plus software. It is most appropriate for digital camera astro imagers, but its focus on astronomical image processing makes it a good choice for all types of imagers, including film scans and CCD as well as digital cameras.