A CCD camera buzzword primer.
Flat fields? Histograms? Dark frames? Welcome to the world of digital imaging. Every technology is rife with jargon that is bound to baffle beginners, and astro imaging is no exception. Take, for example, the exchange I overheard at one amateur gathering.
"Hey Robin, that's a great eclipse photo!"
"Thanks. It was easy. I used a fast lens, fast film, and fast exposure."
Considering the different meanings of "fast" in that one sentence, there's plenty of room for confusion. However, now that I've made the jump from conventional astrophotography to CCD imaging, I believe that the latter's lexicon can actually be easier to learn. The lingo of digital imaging seems intimidating only because, unlike photography, most of us have had little exposure to it.
Nearly everyone will grasp the basic buzzwords after a single night out with a CCD camera. But therein lies the catch - getting to that point often requires wading through advertisements and company literature, which are typically awash in technospeak.
What follows is a simple guide for the first-time CCD buyer. Far from being a comprehensive glossary, it covers the basic words that should be understood when you are considering a purchase. Some camera features, such as the number of brightness levels recorded in an image, are determined by the manufacturer, while others, such as a CCD's sensitivity, are intrinsic to the detector itself. Although many companies make cameras, most use CCDs from the same source. For example, manufacturers such as Apogee, Meade, and Santa Barbara Instrument Group (SBIG) all offer cameras based on Kodak CCDs.
Pixel size. It can't be said too many times that getting the best results requires matching the size of a CCD's individual picture elements (pixels) to a telescope's focal length. This fundamental topic was covered in detail in the June issue, pages 38-42, so I won't repeat it here other than to say that many options exist. A telescope's effective focal length can be varied with Barlow lenses and focal reducers. The pixel size can also be increased by electronically combining adjacent pixels in a process called binning.
All the cameras I know of allow binning. But for technical reasons, I don't recommend more than 2x2 binning (combining four pixels into one) for "serious" imaging. Binning 3x3 (nine into one) or more pixels may be useful for finding objects, centering them in the field, and focusing but is likely to compromise a CCD's performance.
Bits per pixel. CCD images are often described as having a certain number of computer bits per pixel. This relates to the discrete levels of brightness (shades of gray) you can use to display each pixel. Cameras aimed at the amateur market typically produce 8-, 12-, or 16-bit images, which can resolve 256, 4,096, or 65,536 levels, respectively. Since the human eye can distinguish only about 40 shades of gray, it might seem that even 8-bit images are overkill. But taking advantage of a CCD's large dynamic range (the span between the darkest and brightest readings that record useful information) requires a good deal more than 256 discrete steps.
The uncertainty you create by dividing pixel brightness into only 256 steps (called quantization noise) can actually be the factor that limits a CCD's sensitivity to subtle differences of brightness. Quantization noise becomes insignificant for images that have at least 12 bits resolution, though 14 bits or more may be needed to fully exploit the dynamic range of today's popular CCDs. Having more bits than the minimum required is not a problem, and in some circumstances is desirable.
Shutters. No one would dream of owning a conventional camera without a shutter. But a shutter isn't a necessity with a CCD camera intended for long-exposure, low-light applications. In general, a shutter is important only for lunar, planetary, and solar imaging. While all shutters are operated electronically, some cameras have a so-called electronic shutter rather than (or in addition to) a mechanical one. Electronic shutters are usually found on cameras having frame-transfer chips. These chips have active imaging areas and separate frame-storage areas that are masked from light. When an exposure ends, the image is electronically shifted in a few milliseconds to the storage area, where it is unaffected by light during the more leisurely readout process.
Some CCD cameras can emulate an electronic shutter by specifying part of the active imaging area as a storage site. This works only as long as no light falls on the designated storage region. While this is often the case when recording planets surrounded by a dark sky, it's unlikely to work for large, bright objects like the Moon and Sun. These targets call for some form of mechanical shutter.
Getting an image from the CCD to your computer typically takes anywhere from a few seconds to a few tens of seconds depending on the camera and how it is connected to your computer. While this download time seems almost instantaneous compared to that needed to develop a conventional photograph, long downloads slow the process of locating objects, centering them, and focusing the camera.
Cooling. Even when left in total darkness every CCD generates electrons spontaneously that accumulate within the pixels. Fewer electrons are generated when the chip is cool. In an ideal world this dark current wouldn't exist, but there are effective ways of dealing with it during image processing. This being so, you want to keep dark current to a minimum, and all popular CCD cameras have some form of cooling to reduce it.
It might seem that if some cooling is good, more would be better, but you shouldn't compare two chips solely on the basis of their operating temperature. With modest cooling, modern CCDs that operate in the MPP (multi-pinned-phase) mode generate very little dark current compared to older chips, even if the latter operate at much colder temperatures. It is the amount of dark current, not the temperature, that is important. Furthermore, dark current must be factored in with other CCD parameters when you are comparing the performance of one chip with that of another.
Of greater concern to most amateurs is having a camera with regulated cooling, which maintains a constant chip temperature in spite of changes in the air. Except for the most demanding applications (such as high-precision photometry), a camera with regulated cooling will greatly reduce the amount of time you'll need to spend taking dark frames. These are the "lightless" exposures that allow you to subtract dark current from your astronomical exposures. Most of us find all too little time to observe, and wasting it taking dark frames because of an unregulated cooling system is not my idea of fun.
Antiblooming. Pixels have a maximum capacity to accumulate electrons (see full-well capacity below), beyond which they become saturated. If they continue to be exposed to light, they end up with excess electrons that generally spill into adjacent pixels. This is usually what creates the annoying streaks from bright stars in CCD images. Even the CCD cameras aboard the Hubble Space Telescope are not immune to this effect.
Some chips, such as those in Kodak's popular KAF series, can be ordered with an antiblooming feature. These chips are especially appealing to people interested in producing photograph-like images. But there is a downside. In the case of KAF chips, antiblooming reduces the detector's overall sensitivity by as much as 30 percent. There is also a general consensus that images from these chips are not suitable for photometric work, where total brightnesses need to be reliably quantified. Some chips have antiblooming protection that can be controlled electronically and thus reduced when you are taking images for photometry.
For chips that don't have antiblooming protection, you can reduce the effects of blooming by taking several short exposures and combining them later during image processing. The result is an image with a long effective exposure but without saturated pixels.
In addition to the camera terms mentioned above, there are a handful of buzzwords that turn up in advertisements and in the manufacturers' technical literature. While they often refer to very important aspects of a CCD's performance, they usually pertain to features inherent in the detector itself regardless of the camera brand in which the chip is used.
There are also parameters of CCD operation that are fine-tuned by the manufacturer to achieve a desired level of performance. As camera users we may find these particulars interesting, but they are usually not the deciding factor in what we buy. It's like a car with a double-overhead-cam engine - it sounds impressive in advertisements, but the bottom line to the owner is performance, not how it is achieved. Below is a sampling of some terms.
Quantum efficiency. The typical CCD in today's amateur cameras records between 30 and 50 percent of the light falling on it. This makes them 10 to 20 times more efficient than photographic emulsions and is one of the major reasons CCDs are such superb detectors for astronomy. Special chip variations can push the efficiency even higher. So-called thinned, back-illuminated chips can approach an astounding 90 percent efficiency in some parts of the spectrum. If you are particularly interested in photometry at blue wavelengths, or if you have a special bent for tricolor imaging, give some consideration to a chip's quantum efficiency at blue wavelengths, since CCDs typically are relatively insensitive to blue light.
Read noise. Camera electronics as well as random statistical variations introduce a small uncertainty in the amount of signal collected by each pixel as it is read out of the chip. Read noise is inherent in a chip, and it ultimately limits the performance that can be obtained. But you must weigh read noise with other factors when assessing a chip's performance. In other words, one type of CCD with 20 electrons of read noise may still offer better performance than another type of chip with only 15 electrons' worth.
Today most cameras use a process called double-correlated sampling to significantly reduce read noise for a given detector.
Full-well capacity. As mentioned earlier, pixels have a maximum capacity to accumulate electrons. It stands to reason that larger pixels have a greater full-well capacity (sometimes called well depth). The dynamic range of a CCD - the range of brightness levels it can distinguish - is essentially set by the full-well capacity divided by the read noise. This means that a large full-well capacity, while desirable, does not tell the whole story. For example, a chip with a full-well capacity of 200,000 electrons and a read noise of 50 electrons has a dynamic range of 4,000 to 1. On the other hand, a chip with a full-well capacity of 85,000 electrons but a read noise of only 15 electrons will have a greater dynamic range (almost 5,700 to 1) and be able to record more usable information.
Gain. This is one of digital imaging's tricky terms, in part because its meaning here is very different from that typically encountered in the literature of amplifiers and electronics.
The electrons accumulated in a pixel are read out in the form of an analog voltage, which must be converted to a number. This is done with an analog-to-digital converter, which is responsible for determining the bit-resolution of the camera's images. (As mentioned earlier, A/D converters are typically 8-, 12-, or 16-bit devices in today's popular astronomical cameras.)
In CCD parlance, gain is usually specified as the number of electrons in each step of the A/D conversion. These steps are often called analog-to-digital units, or ADUs for short. If each ADU contains two electrons, the camera is said to have a gain of 2; five electrons, a gain of 5; and so on. Typically, a manufacturer selects a gain that matches a pixel's full-well capacity to the resolution of the A/D converter. For example, if a pixel can hold 150,000 electrons and the camera has a 16-bit A/D converter (65,536 steps), then a gain of 2.3 electrons is about optimum.
So, which is better, high gain or low gain? Well, that depends. If our intent is to resolve subtle brightness differences in faint nebulae, then a camera with a low gain is desirable, since a pixel differing by a few electrons from its neighbor will be recorded as a different brightness after the A/D conversion. A low-gain system may not, however, exploit a pixel's full-well capacity. This would be a disadvantage for photometric work when the target and reference stars span a wide range of brightnesses. There is also the issue of binning, which increases the effective full-well capacity of a pixel, and this is often taken into account when manufacturers set a camera's gain. Rarely, however, is the gain optimized for use with more than 2 x 2 binning.
Certainly there are many other issues to consider when buying a CCD camera. One involves camera-control software. Foremost is whether the program runs on your chosen computer platform. Most cameras are designed for DOS or Windows machines, but recently SBIG introduced software written specifically for Macintosh computers.
All camera-control programs include functions for acquiring and displaying images and for saving them on a computer's disk. They also include routines for basic image calibrations such as dark-frame subtraction (to remove dark current) and flat fielding (to correct for individual differences in pixel sensitivity and for vignetting). Beyond that, however, further image-processing routines vary from manufacturer to manufacturer. Some offer very little, while others include a wide variety of sometimes sophisticated functions.
Nevertheless, at some point you will likely want to use your images with other display and analysis programs. This brings up the issue of compatibility. Most cameras have their own unique format for storing images. The closest thing to a universal format in the world of astronomy is FITS, which stands for Flexible Image Transport System. Almost all specialized astronomy programs can read FITS images, so it's worth verifying that the camera you select can save its images in FITS format. Another important format is TIFF, which stands for Tagged Image File Format and is used by many image-manipulation programs like Adobe Photoshop that are of interest to amateur astronomers.
The CCD cameras intended for today's astronomical market are all of high quality, and generally there are few pitfalls awaiting the buyer. Nevertheless, understanding some of the basic buzzwords will help make camera selection easier.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||charge coupled device|
|Author:||Cicco, Dennis di|
|Publication:||Sky & Telescope|
|Date:||Aug 1, 1997|
|Previous Article:||A garage-rooftop observatory: here's a telescope housing that combines relatively simple construction with a low-key design.|
|Next Article:||A matter of degree.|