Archaeology and Technology
Digital Still Cameras
I am prompted to write on the archaeological use of still cameras following my own recent conversion to them. In May 1998, I was still a decided skeptic about whether relatively moderate-cost digital photography was at all ready for serious professional use in archaeology. Now, 10 months later, I have taken more than 15,000 digital pictures for a variety of uses I would never have imagined. The utility of digital photography is breathtaking and demands your serious attention. I am writing to offer some perspectives to those new to this technology.
One of the initial shocks for the uninitiated is what seems to be relatively high price tags of digital cameras, especially compared with film cameras. Consider, however, that with digital cameras, film and processing costs drop to zero, and the media necessary to store large quantities of digital shots can be very inexpensive, as low as 1/20th to 1/30th of a cent per image. Thus, with digital cameras, the equipment is relatively expensive, while their use tends to compensate with low cost.
Digital cameras can be classified into a series of price cohorts, with ascending image quality and feature richness. The lowest ranks are cameras of VGA resolution (640 x 480 pixels) or less. These can still be purchased for less than $500, but their limited abilities compared with slightly more expensive cameras will soon leave them obsolescent.
Next up the line are the cameras primarily discussed in this articlewhat can be called quality digital point-and-shoot cameras. Prominent manufacturers of these compact and lightweight cameras are Epson, Kodak, Nikon, Olympus, and Sony. My experience derives from extensive use of a Nikon Coolpix 900 I purchased about 10 months ago for about $850, but many other similar cameras are available. This group of cameras generally have XGA resolution (1024 x 768 pixels) or better, an optical viewfinder like that of a simple film camera, and a small, 1.82.5 inch LCD display that serves as both viewfinder and a small display for reviewing shots already taken. Price ranges for these cameras range from $500 to $1000 for the camera itself, depending on features included; additional memory media ($100200) and sets of rechargeable batteries ($50100) will be necessary for extensive use. This group is often referred to as 1 or 2 megapixel cameras, depending on the product of their vertical and horizontal resolutions.
A great jump in price to the $5000 range adds a series of abilities, including manual control of exposure and interchangeable lensesoften the same lenses used with the manufacturer's film cameras. These cameras are generally beyond most archaeologists' budgets, and in fact now rarely offer greater image resolution than the better point-and-shoot type, although some increase in image quality is found.
Finally, very high-end cameras run into the $20,00030,000 range, offering very fast shooting capabilities and increased resolutions as high as 6 megapixel. The clear winner for archaeological fieldwork and simple object photography is the quality point-and-shoot camera.
Two of the greatest doubts that I encounter among my colleagues concerning digital photography are issues of image quality and capacity of the camera to hold images, especially as compared to 35 mm film format. Both are serious considerations, but an important perspective to maintain is that digital photography probably should not be considered a simple replacement for conventional film; in the immediate future, at least, film will continue to have resolution advantages. Instead, digital photography has characteristics that complement film photography, making a digital camera another tool for our kit, not necessarily a replacement of another one.
Image quality with digital cameras is to a fair degree reliant on resolution, i.e., the number of pixels, or color dots, in the image. But it is important to note in passing that the optics and non-resolution electronics of the camera also are issues in image quality, such that a higher resolution camera may have a poorer image by most standards than a lower one. Resolution varies in non-continuous jumps roughly corresponding to computer video resolutions of VGA (640 x 480 pixels, ~300,000 total) SVGA (600 x 800 pixels, ~500,000 total), XGA (1024 x 768 pixels, ~800,000 total), and on to megapixel (around 1260 x 980, ~1.25 million), two megapixel (1600 x 1200, ~2 million total), and higher resolutions. For comparison, 35 mm film can be considered to have the equivalent resolution of around 8 million pixels, roughly 3200 by 2400, but this is variable with different films and developments. Thus, affordable resolution at this time will be around half (in a dimensional sense) or one-quarter (in total pixels) of 35 mm film, although it is increasing constantly (25 percent per year is a reasonable estimate).
Resolution limits the degree of enlargement possible before pixelation, the point at which individual color pixels become apparent and obtrusive in the image. A rule of thumb is that good, photographic-quality images can be obtained at 4 x 5 inches for SVGA, while 5 x 7-inch images are obtainable from XGA; megapixel resolution will go as far as 8 x 10-inch images under ideal circumstances. For most of us, it is a very rare 35 mm image that will ever be enlarged above 8 x 10 inches, especially for professional images, and thus megapixel resolution is starting to encompass most common needs. Given the easily found ability to seamlessly stitch digital images together, it is possible to take a number of overlapping digital shots of a scene, and assemble a much larger image than the camera can intrinsically take.
Image-holding capacity is the other major issue, and for digital cameras it is increasing radically. Almost all moderate-cost digital cameras rely on either conventional floppy disks of 1.44 MB capacity, or one of two types of memory card. The 3 1/2-inch floppy disk, used heavily by Sony products, has the major advantage of being a low-cost, expendable medium that can be thought of somewhat like filmyou just carry a lot of disks, like you might carry many rolls of film. But like juggling hundreds of rolls of film, this also can be seen as emulation of a long-time inconvenient procedure. Also, in my experience, the 3 1/2-inch floppy is the most failure-prone digital medium, so reliance on them may prove a misplaced trust.
More common are the two tiny card media: the Compact Flash and Smartmedia formats. While they both involve tiny chip-like insertable memory, there are very real differences in compatibility and future promise. The Smartmedia was developed for other applications, primarily cell phones, and does not articulate easily with microcomputer inputs. Capacity of the Smartmedia, currently at 16 MB, will probably not go beyond 32 MBa severe limitation. The Compact Flash memory, developed specifically for digital photography, is less limited in size, with capacities now going up to 128 MB and beyond. Perhaps as important is that only a very simple, straight-through adapter is required to "size-up" and plug a Compact Flash card into the PCMCIA slot of a notebook computer. The computer will recognize the Compact Flash card as a disk drive, and transfer of images to hard disk is extremely fastI have transferred 30 MB of images from Compact Flash memory to a laptop computer (100200 photos) in under 45 seconds. While there are other ways to transfer photographs from the camera, such as cabling through serial, parallel, or USB ports, the PCMCIA slot on laptops, and equivalent ports that can be attached to desktop machines, remains a very attractive and functional procedure.
A variant of the small cards are the more standard ATA PCMCIA memory cardsthe larger, credit-card-sized cards often used with handheld computers. Although more bulky, these cards are available in quite large capacities, already exceeding 200 MB. This form of memory is effectively the same as the Compact Flash, and in fact cameras that accept the ATA cards can also use Compact Flash memory within a PCMCIA adapter.
Image capacity of the various media is dependent on two primary factors: the resolution of the image, and the degree to which the image is compressed (the color depth also affects size, but cannot usually be regulated in the camera). To give an example, if an imaginary image is 500 x 500 pixels, it will have an uncompressed size of about 733 KB, while a 1000 x 1000 pixel image will be about 2930 KB as a bitmapbasically an image in which every pixel is accounted for within the file. By compressing the file, the 500 x 500 image can be greatly reduced, typically to a file in the range of 2080 KB, or 510 percent of the uncompressed image size, using the fairly standard JPEG compression. The more extreme the compression, the greater is the loss, typically of detail and color, in the image. A file size reduction of 90 percent for most images will still yield a photograph effectively undistinguishable to the eye from its bitmap equivalent. Most digital cameras utilize compression, and often give different "quality" levels: This typically refers to degree of compression, not image resolution. In practice, I find that degree of compression, in the range used by most cameras, has relatively little effect on the image, while a reduction in resolution has a very visible impact on image quality. Thus, if there are strong limits on storage space for images, it is far better to save space by compression than by reducing resolution. Figure 1 shows a 1260 x 980 photograph that is substantially compressed to 140 KB by the digital camera (it would have been 3600 KB as an uncompressed bitmap). Three closeups of a small area of that image show the original JPEG compressed format (23 K for this 280 x 280 pixel closeup), an extreme JPEG compression (7 K), and the same area with the original JPEG compression but with half the resolution (also 7 K for what becomes 140 x 140 pixels in reduced resolution). It should be apparent that the recompressed JPEG image, although suffering some degradation, is much superior to the lowered resolution image, although the file size reduction is exactly the same.
We can now consider the image-holding capacity of current digital cameras. Taking 1280 x 960 as a standard, reasonable quality image size, a floppy disk could hold somewhere in the range of 1020 shots under strong to extreme compression. A 48 MB Compact Flash module will hold about 400 images under strong compression. My Nikon Coolpix 900 camera, which has three quality settings, the lowest of which ("basic") is equivalent to the strong compression mentioned above. The "normal" setting doubles the image file size to around 250 KB, and "fine" images are again double sized at 500 KB (but all have exactly the same pixel resolution, just different degrees of compression). I have extensively examined shots taken across the quality range for different subject matters, and have a very difficult time perceiving any difference in the images, suggesting that strong compression can be
Figure 1a. Stone tool from Kenya, shot with Nikon Coolpix 900 camera, 960 x 1280 resolution, "basic" quality setting (strong JPEG reduction); file size = 140KB.
Figure 1b. Closeups of lower right area of stone tool in Figure 1a. Left image is enlargement of original digital image of Figure 1a (280 x 280 pixels, file size = 23KB), middle image is extreme JPEG compression (280 x 280 pixels, file size = 7KB), right image is reduced resolution with same JPEG compression as left image and Figure 1a (140 x 140 pixels, file size-7KB).
used without noticeable loss. Thus, current cameras can store well in excess of 300 good quality images prior to downloading, and this figure probably can be doubled by the time this article is printed. On the horizon is a newly developed, tiny hard disk meant to plug into a Compact Flash socket, which will have a capacity of around 300 MB, and would be capable of holding around 2,000 images of the resolution and quality under consideration. It is obvious that digital cameras are taking on a very different type of storage character than film cameras, in which weeks or months of continuous shooting will be possible, with sporadic downloading to a computer.
For the field archaeologist, this offers both advantages and challenges. Digital cameras take the cost out of shooting great quantities of photographs, give instantaneous feedback about the outcome of a shot on their LCD screens, but require access to a computer for storage. In a seven-week field season in Chavín de Huántar, Perú, in 1998, I experimented widely with the Nikon Coolpix 900 camera mentioned above. I took more than 10,000 photographs, discarding about a quarter of them as duplicate or of inadequate quality. My primary field computer was a middle-of-the-road pentium laptop, with only 2 GB of hard disk space shared for many functions, but an active matrix XGA display useful for viewing images. The volatility of digital media, such as hard disk failures, and the theft potential for computer equipment make some sort of backup capacity for digital photographs not only a necessity, but actually a great advantage. I use a SCSI PCMCIA card in my laptop computer, which allows me to link to a number of exterior devices, including an external hard disk, a Zip drive, and most importantly, a CD-Rom writer. With a 650 MB capacity, and with writeable CD media costing under a dollar each, the CD is an ideal media. I have written several hundred CDs under a variety of conditions, and have never had a CD fail once it has been written. The capacity of a CD is about 2,5005,000 images of the type discussed above. Perhaps most important is that multiple copies of all images can be maintained, minimizing the possibility of loss. Carried in compact wallet-like carrying cases, CDs occupy very little space, have negligible weight, and can be shared widely among computers. While the CD medium will probably be in use for a long time, it is true that CD archives will eventually have to be converted to future formatsbut this should not be an arduous task as a digital transfer.
The strongest advantages and disadvantages of digital cameras come from their intrinsic linkage to computers. A film camera is a very stand-alone operation, excepting perhaps the eventual linkage to a slide projector. In contrast, a digital camera outside of the computer world is a fish out of water. The end product of a digital camera is better considered to be a digital image, rather than a digital path to a conventional print or slide. While digital images can be printed on high-quality printer paper, it is a slow process, and yields results no better and often worse than film at about the same cost. The less said about making slides from digital files, the betterusually this involves slide-making services charging a hefty fee. The real forte of digital images is in display on a computer screen or from an LCD or video projector. As archaeologists continue converting to computer use and digital presentations, this issue will lessen with time.
I like to think about digital photography as more of an imaging tool than photography per se. With film we usually attempt to make accurate images of what we see, and often run into problems when our eyes see better than the camera, such as in dim light. With digital imaging, we are not at all limited to replicating the eye. As I wrote in a previous Bulletin article [1998, 15(5): 1419], making continuous, seamless panoramas for virtual reality is a digital specialty. In this case, the digital camera can allow us to look all around ourselves. Digital images can be manipulated endlesslyincreasing or decreasing contrast, eliminating noise, altering or correcting color, orone of my favoritesinstantly and accurately converting color images to black-and-white. How often have we been frustrated at publication time when beautiful color slide images are converted to lifeless, murky black-and-white illustrations?
Inexpensive imaging with digital photography allows rapid documentation of things we typically do not photograph. In Chavín in summer 1998, I photographed almost all excavated objects, often in multiple views. The shots are taken rapidly, bracketed in their sequence by pictures of the tags or bag labels. Since the pictures are dated to the minute, automatically, in the form of the file creation time, and given a sequence number by the camera, it should always be possible to reassociate the objects with their labels if they should lose their order. I also took a great number of field documentation photographs of surface and excavated subjects. I am surprised by how useful it is to have 10 or 20 shots of a given subject, rather than the two-to-five frames I would normally take with film. I also found that the digital camera, with very effective automatic settings, will take pictures in tight situations that I cannot even get my head into, and give instant feedbacksort of a detachable eye!
In another situation, we found a large (~200 KG) piece of decorated stone cornice in one of our excavations (Figure 2), unfortunately broken in antiquity. I knew that various pieces of previously recovered lithic art were stored throughout the internal
Figure 2. Composite photograph showing cornice
fragment excavated at Chavín de Huantar
in 1998 (left half of image) matched to cornice fragment excavated decades ago and now
stored in the Gallery of the Labyrinth (right half of image). Left segment of cornice approximately
62 cm wide.
"Digital photography has characteristics that complement film photography, making a digital camera another tool for our kit not necessarily a replacement of another one. "
gallery system of Chavín, and wondered about the seemingly remote possibility that one might be the missing segment of our find. A couple of hours, and two dozen digital photographs later, the answer was cleara perfectly apparent match was found when the digital photographs were scaled and brought together, edge-on, in the computer (Figure 2). This same result could be obtained through rubbings or standard photography, but nowhere near as efficiently.
Another use I have found particularly effective is documenting field records, while still in the field. I shot all field notes, field drawings, and even borrowed publications under a single incandescent light bulb at night, with very good results. In effect, the digital camera and CD writer combination can serve as a very inexpensive photocopy machinewhen was the last time you got 2,000 + photocopies for less than one dollar (the cost of CD storage)? With the camera on a tripod, the speed of "copying" is easily that of a photocopy machine, and compared to the nearest copiers4 hours from Chavínmuch faster, cheaper, and usable. At 1280 x 960 resolution, full-page documents are quite readable, and much better than photocopies for legibility of light pencil or otherwise indistinct content. Digitally-shot profile drawings can easily be traced in a computer drawing program, if not auto-traced, yielding quick but accurate field products. By the time I left the field, I had four complete backup copies of all photographs and field records at negligible cost. I still use the digital records much more frequently than the originals; I now have all years of Chavín field records on a single CD, which I can carry to any work locationquite a difference from carting around 2,400 pages of original or xeroxed records! Indexed into directories, these records can rapidly be reviewed.
These same qualities allow for rapid and convenient reproduction of other images, such as illustrations in publications. For producing digital slides for lectures and presentations, I have found the camera much more effective than a scannerit is faster by far, and produces an image close to the XGA resolution I project with available equipment. I can easily shoot and transfer worthwhile graphs or photographs into a digital slide lecture sequence in just a few minutes. Similarly, in the Peruvian highlands I was able to give digitally projected lectures including shots of excavation surfaces and objects revealed less than a day before.
Another unusual, but very helpful use I found for the camera involves a collection of about 500 film slides I maintain in Lima, Peru, for giving talks, and as a security backup of some important images. I have had no efficient way of documenting which slides are in this collection, and many years I end up taking slides to Peru that duplicate images already there. A listing of slide catalog numbers only partially solves the problem because of the time involved in looking up and examining the originals of any given slide number. The digital camera provided a solution, simply by rapidly shooting the projected images of the Lima-stored slides at the low resolution setting of the camera (640 x 480 pixels). The entire shoot took two hours, and the resulting 500 images occupy about 10 MB of disk space1/60th of a compact disk. From these images I have made thumbnail sheets, and thus have a series of eight very compact index images of the slide collection that summarize its contents at a glance.
Other impressive qualities of digital cameras are their abilities to take closeups and their low light capabilities. Fairly recent vintage digital cameras are capable of producing full-frame images of areas as small as an inch square, and models appearing at the time of this writing are reducing that size by halfwell into the range of macrophotography, but without the need for specialized lenses. Like their video camera cousins, digital cameras also are capable of taking very useful photographs under very low-light situations in which most film cameras would be highly disadvantaged, and certainly require a film change. And, most important in such difficult conditions, with the digital camera you can immediately observe the quality of the obtained image on the camera's LCD screen. Another light advantagemost digital cameras automatically adjust for the type of light available and many can be set specifically for incandescent, florescent, cloudy, or bright sun, etc., with excellent white balance. Similarly, brightness and contrast can be adjusted prior to shooting, and newer point-and-shoot cameras have shutter and aperture priority settings, along with a number of metering options. In the relatively near future, these cameras will have manual control of exposure and features typical of the best film cameras.
Another consideration with digital cameras is that they require a fair amount of battery power. Alkaline batteries are drained in just a few dozen shots; however, the new nickel metal hydride (NiMH) rechargeable batteries are capable of powering the cameras much longer. My camera requires 4 AA batteries; a set of NiMH rechargeables will provide 200 + shots, somewhat less using flash. These batteries charge in under three hours and do not have the memory problems that require complete discharge of NiCad batteries before recharging. I maintain three sets of rechargeable batteries and have never exceeded their capacity, even on days of shooting well over 600 pictures.
The things to keep in mind when selecting a digital camera depend on the end use, but I suggest that most serious archaeological applications will require a minimum 1280 x 960 resolution: sufficient for high-quality digital presentations or for publication. Just as with film cameras, however, the optical quality of the lens system is very important, especially in the higher resolution cameras. There is no substitute for trying a camera before buying it, but reviews in magazines or on the web are helpful. Next comes the important issue of zoom capabilitiesmost quality point-and-shoot cameras come with a 3 X optical zoom, giving the equivalent of about 36110 mm zoom in a 35 mm film camera. This optical-mechanical feature adds considerable expense to a camera, so less expensive but good resolution cameras typically have no zoom or a digital zoom feature. Digital zoom is next to useless, because it effectively only enlarges and crops the native resolution format; thus, digitally zooming 2 X with a 1280 x 960 resolution camera will produce a zoom image with an effective resolution of about 640 x 480the user could do just as well shooting without the zoom, and cropping the image. A number of cameras have add-on lenses that will give fish-eye, very wide angle, and telephoto lenses beyond the range of the optical zoom.
Most cameras in the 1280 x 960 + resolution range have a bewildering range of features accessed by various menu systems and LCD panels. Some features that are not universal, but could be important in particular situations include a macro or closeup setting, the ability to shoot a rapid series of shots, or a connection to an external flash for sophisticated low light photography. Almost all cameras allow examination of the images in memory, deleting, protecting, or even hiding them from casual observation. One particularly important feature, especially for those interested in panoramic virtual reality, is the ability to lock the exposure settings on one shot, and then continue to use those settings until reset. This step is necessary to get equivalent, stitchable exposures, and may have applicability in other archaeological situations. This is as close as the point-and-shoot variety of digital cameras come to manual settings, and allows some additional control.
Relatively new features combined into a number of different camera models are sound and video recording abilities. Sound "bites" of short duration could be very valuable in labeling photographs as they are shot, potentially speeding the photographic process. The utility of sound attachments needs to be explored, for they would be most useful when large number of photographs are being taken, and sound takes up a relatively large amount of space. Video, naturally enough, is even more space expensive, and video of significant quality (defined by resolution and frames-per-second) will quickly fill most media. Undoubtedly archaeologists with sound and low-level motion needs will find these hybrid cameras worth considering.
A final word of advice on the timing of digital camera purchasesbecause of the camera market, new digital cameras are generally introduced in late spring and late fall, prior to vacation and holiday seasons. New features will appear at those times, with older models discounted; accessories for the cameras are introduced between these peaks of new camera models. April and May will see a flood of new cameras; particularly interesting are models with resolutions in the 1600 x 1200 range for about $800 to 1000giving image quality superior to $15,000 digital cameras of just a few years ago. There are always arguments for postponing the use of a new technology, including its rapid changing character. My suggestion is that digital still photography is now at a point of great utility for many of us, and we ignore it at the peril of our time, effectiveness, and data quality. ·
My thanks go to Wayne Llano and Richard Olen in the Digital Photography area of Keeble and Shuchat Photography in Palo Alto, California, who have provided me with immensely helpful and updated perspectives on digital cameras.John W. Rick is in the Department of Anthropological Sciences at Stanford University.
back to top of page