Having written two long articles for Personal Computer World on one of my great passions, aviation (here and here), it was perhaps inevitable that I would sell them a piece on an even greater obsession of mine – photography.
The article mentions the Sony Mavica. This wasn’t a digital camera – it was a still video camera, with an analog NTSC video image being recorded onto 2in disks (like smaller versions of the familiar 3.5in floppy). This was at the stunning resolution of 570 × 490. You needed a TV to view the images.
When the Mavica first appeared, in 1981, I was working as a section editor on a photography magazine. We were visited by a delegation from Sony who gave us a pre-announcement preview of the technology. I can still recall thinking, “Who would possibly want this?”.
Not many people, as it turned out. It would be some considerable time before true digital photography hit the mainstream.
COMPUTERS IN PHOTOGRAPHY
Photographers are a curious lot. Perhaps it’s the artist in them that makes serious photographers – amateur or professional – so reluctant to embrace new technology. Automatic exposure control, electronic shutters and other marvels of the technological age were frequently greeted with suspicion and derision – usually to be quietly adopted when the fuss died down. There seems to be a pride in doing things the hard way, but it is starting to give way in the face of a rapidly accelerating assault on photography by technology that would seem more at home in the computer world.
Some of the changes are happening inside the cameras themselves as they become increasingly sophisticated. For a while it was simple stuff like automatic exposure control that, even if it used a chip, would hardly tax the brain of a credit card-size calculator. Then manufacturers started to cram in more: a choice of exposure modes or different versions of the same mode. Program mode, for example, sets both the aperture and shutter speed of a camera, but you might have a choice of programs, using different algorithms to give priority to high shutter speeds (for sports, perhaps) or small apertures (to make the picture as sharp throughout as possible).
It wasn’t long before they started doing really clever things, thanks to chips. With the FA, Nikon was the first to introduce multi-pattern metering. This split the image area into five sections, the light level in each being measured separately. A special algorithm, developed by analysing around 100,000 photographs, used the intensity and pattern of the light levels to decide on the best overall exposure. The camera was remarkably good at dealing with difficult exposure situations.
Now autofocus has made it big in the serious camera market. And just as with metering, the camera may offer a choice of focusing modes. The Nikon F4, for example, will focus once per press of the shutter button, or it can be made to continuously focus as the subject moves around. Moreover, with a subject that is moving towards or away from the camera, it will effectively compute the speed and make an allowance for subject movement during the brief time lag between pressing the button and the shutter actually opening. No wonder the F4 needs two eight-bit and one four- bit microprocessors.
Introducing computers into cameras didn’t just mean overcoming technological problems, as Canon discovered when it was developing the T70. This was one of the first to go for the full hi-tech approach, using buttons to switch modes and set functions, and a multi-function LCD panel to tell you what’s going on. An optional data back provided features that had never been built into a camera before – like an intervalometer (which takes pictures at set intervals) and extremely long, electronically timed shutter speeds. At the launch, in 1983, I marvelled at how wonderfully advanced it all was, but the Canon spokesman admitted that they’d had the ideas and the technology hanging around for a while.
The reason they hadn’t launched the camera before was that they were waiting for the buying public to become sufficiently sophisticated. At that time, people were just getting used to the idea of programming their VCRs and other domestic appliances. That paved the way for the T70’s push- button approach.
Behind all this high-tech wizardry, however, the basic photographic technology, which means the film, is not significantly altered since its earliest days over 150 years ago. Yet even that could be on the verge of changing – at least a little – as solid state threatens to replace silver halides.
Sony fired the first warning shot in 1981. That’s when it announced the Mavica, a camera that replaced film with a floppy disk. It was the first of a breed known as the still video camera (SVC). Like many of today’s camcorders, SVCs are built around at least one CCD (charged coupled device) – a chip with rows of pits that react to light.
SVCs typically offer two resolutions, depending on how they form the image. A TV picture is created by lines running across the frame: in fact, the full picture is normally two sets of lines. The first set of horizontal lines forms the ‘field’ image. A second set, interlaced with the first (effectively filling in the ‘gaps’ between the first set of lines) completes the picture creating a ‘frame’ image. A frame picture thus has twice the vertical resolution of a field image.
The images are recorded on to a 2-inch disk – the same beast found in a few laptop computers. Although this disk has an unformatted capacity of 1Mb, this isn’t nearly enough to record the images digitally, so – at least with the first batch of cameras – the recording is analogue. Each disk can hold 50 field or 25 frame images.
Still video cameras have some obvious applications in desktop publishing. If you’re producing a company newsletter, say, or you’re an estate agent creating a property list, you’ll often need to get a photograph of a person or place quickly. The conventional route is to snap a photograph, get it printed as fast as possible (or use a Polaroid if you need real speed) and then digitise it with a scanner. A still video camera provides a much more direct route, even though you need to convert from analogue to digital formats, and the quality – though nothing like what you can achieve with a decent and film stock – is good enough. It’s certainly better than a scanned Polaroid.
The firm’s main product is the two-chip NVC 5000. One chip is used for luminance, the other for chrominance and colour difference. The camera gives 500 line resolution, which is good enough for some DTP applications, although as the colour is all dealt with by one chip, colour reproduction is a little flaky (reds are reputed to be weak). A choice of field or frame recording is provided. The camera offers interchangeable lenses, unlike its single-chip predecessor, and lens mounts matching most of the popular Japanese manufacturers are available. It has one drawback, though: it costs L5000.
A price wasn’t available for the new three-chip camera, but Sony was keen to point out its improved colour handling. Like the two-chip machine, one CCD handles luminance, but the other two share out the job of producing colour, being responsible for r-y and b-y (red minus yellow, blue minus yellow). Essentially it’s the same high-band format as the two-chip model, but the extra CCD gives a resolution in excess of 500 lines.
Just as important as the camera is the playback machine (the MVR5600). You use this like a VCR to view your happy snaps on a TV set. But its video output doesn’t have to go to a monitor: instead you can hook it to a digital image handler which, just like a modem, is capable of sending the image data over telephone and ISDN lines to another of its ilk. That’s when machines like the Mavica start becoming interesting to news and sports photographers. Without any messing around with chemicals, they can wire their pictures back to a newspaper or magazine. It’s then a simple matter to get the image files into a computer system for page make-up. The Daily Mail, for example, used the Mavica for Wimbledon, and it’s frequently used in the US for pictures of awards ceremonies.
Sony has produced a program for 24-bit video card-equipped Apple Macintoshes that communicates between the digital image handler and the Mac, converting image data to TIFF files. The image handler uses Adaptive Discrete Cosine Transform (ADCT) compression, storing the images internally in this format. The program’s job is to convert that format into TIFF.
In the UK, Canon was the first to make a real stab at the amateur market with the RC251, better known as the ION (which is supposed to stand for Image Online Network – well it would, wouldn’t it?). The L500 price tag of the camera was achieved in part by making it simple: it has a non- interchangeable, fixed-focus lens and a single CCD producing 50 frame images per disk. According to a Canon spokesman, the images are ‘higher quality than VGA – there’s a noticeable drop in quality when going from TV to a VGA screen’, although the picture is not as good as a frame image. The camera has a PAL output, but Canon also sells a playback unit that offers slightly better image quality.
In contrast, Nikon’s entry in the SVC scene reflects its position in the 35mm camera market – that is as the professional camera, particularly among press photographers. The two year-old QV-1000C is remarkably like a distorted version of the firm’s flagship professional camera, the F4 – indeed, the two cameras have several parts in common. Functioning is very similar too: the QV-1000C has interchangeable lenses, an optical, through-the-lens viewing system, a shutter speed dial running from 1/8th to 1/2000th sec, several automatic exposure modes, a hotshoe for a flashgun and even a pseudo-motordrive firing at up to 20 frames per second. Three sensitivity levels are provided that correspond to film speeds of ISO 400, 800 and 1600. Resolution is 450 TV lines (600 pixels per line) using NTSC.
Two zoom lenses are available for the system – a 10-40mm f1.4 and a 11-120mm f2, plus an adapter to use ordinary Nikkor lenses. Because of the image area of the CCD, these lenses are equivalent to 40-160mm and 44-480mm on conventional 35mm cameras. Given their relatively high speeds (which means ability to work in low light) they are ideal for photojournalism – precisely the area they’re aimed at. The system has been extensively used by the UK newspaper industry.
For the most part, the Nikon machine shoots frame pictures, although in the very high-speed continuous shooting mode it switches to field images. The surprise, perhaps, is that it’s black-and-white only. This permits higher resolution and sensitivity than colour systems – even ones with nominally the same TV resolution – and if you’re shooting for newspapers these are generally more important than colour. Ironically, while photographers and picture desks have shown great interest in the camera and want to acquire them alongside or instead of their conventional 35mm cameras, the managers of newspapers – the people who sign the cheques – have complained about the lack of colour and are holding back. But it’s a natural for DTP applications, which are overwhelmingly monochrome.
Nikon is planning a high-resolution, digital colour camera for 1992. This will use JPEG/ADCT compression, which is essential to get enough images on to the disk. Digital has the advantage that it suffers no distortion during recording: images recorded in analogue fashion on a floppy disk suffer the image equivalent of wow and flutter on a hi-fi record. This leads to distortion such as jagged edges, similar to the effect that pixels have when enlarged. Averaging of adjacent tones helps, but the best solution is digital recording. Digital images can also send their images straight on to a computer disk, with no need for extra equipment, and there’s no image ‘noise’. Such cameras already exist in Japan.
A crucial part of the system is the QV-1010T transmitter. It’s designed to transmit to conventional receiving systems – the wire machines used by newspapers for years that produce a hard copy print at the receiving end or which store the image on computer for page make-up systems. It conforms to CCITT 120, 60 and 240, AP, UPI and Japanese standards for transmission. Working like a small, desktop fax, the transmitter includes a CRT monitor for picture selection, works off a rechargeable battery or 12V DC and has modular jack outputs to connect direct to the phone system plus audio output for using an acoustic coupler. It takes about five minutes to transmit one picture which is slightly faster than transmitting an 10 x 8inch print using similar machines for conventional picture wiring.
Although Nikon considered having a digital output from the machine (and there’s a blanked-off hole to prove it), that was abandoned. To get the images on to your computer you’ll need a frame grabber capable of handling NTSC signals connected to the transmitter’s monitor output. Each picture occupies about 300K of memory. The transmitter uses an averaging algorithm to smooth tones and eradicate problems caused by noise, but this is not available through the monitor output.
The price of the system reflects its professional status. The camera body will impoverish you by a cool L2400, while lenses are L2665 and L535 and the transmitter L6400. You have to be pretty serious about getting into electronic imaging to pay those prices – or you must have a serious need.
This isn’t Nikon’s only still video product (although it is the only portable one). The firm also has a system based on a high definition TV (HDTV) camera. There are two HDTV systems in the world: the Japanese went for a resolution of 1125 TV lines, while a new European standard gives 1250 lines. The US hasn’t made up its mind.
Nikon’s camera, which is designed to be rostrum- or microscope-mounted, provides a screen resolution of 1920 x 1035 pixels. Nikon also sells a frame store (each image occupies about 6Mbytes), and the pictures can be sent to a computer or through a GPIB interface to a printer.
It was developed initially for Quantel – a manufacturer of video image manipulation equipment – as an image input device for large pictures or 3D objects. Now it’s used primarily by the semiconductor industry. Attached to a microscope, firms use the camera to check chips for faults. Everything can be done in near real time, and with very good colour resolution, because the camera uses only one monochrome tube. Three images are taken sequentially through different filters to give full colour. Given that chip factories aren’t exactly thick on the ground in the UK, there have been few sold over here, but in Japan the system has been successful.
Other applications include scanning film images: it’s a lot faster than normal film scanners – 1/3sec rather than many seconds per frame. An auction firm is interested in using it to send images of lots to overseas bidders. And one Japanese firm – Nomura – uses it to produce pictures of visiting clients, which must make it the world’s most expensive alternative to a Polaroid camera.
You don’t have to invest in expensive hi-tech equipment to feel the impact of new technology. In about a year or so’s time, even the most amateur of snapshooters with the simplest cameras will be able to reap the benefits. That’s when Kodak’s Photo CD product will become available.
Kodak, of course, has something of a vested interest in traditional silver-based photo technology. But it can’t ignore the fact that high technology makes for attractive consumer products. However much it may protest that film offers far greater quality than electronic imaging (as Kodak proclaimed very loudly when Sony first announced the Mavica), the firm knows that the snapshot market is one where ease of use and snazzy gimmicks often mean more than obscure debates about quality – what else could explain the success of 110 cartridge film? So it’s now getting into electronic imaging. Well, if you can’t beat ‘em…
Photo CD, first shown at Comdex Fall in 1990, is a service that runs alongside the normal developing and printing (D&P) business – or it will when it’s made available in 1992. You take your film along to a participating chemist or D&P shop: what you get back are your prints or slides, as normal, plus a CD with the images scanned on to it.
The D&P shop uses a workstation comprising a scanner developed by Kodak, a Sun Microsystems computer, a Kodak colour thermal printer and a Philips discwriter. The scanner uses three parallel linear CCDs, each giving a resolution of 2048 pixels across the frame. Kodak says the image quality exceeds HDTV standards. The resulting image file contains 18 Mbytes of data which is piped to the Sun Sparcstation for colour and density correction before being compressed using proprietary Kodak algorithms and hardware. With a storage capacity of over 600 Mbytes, each disc is capable of holding up to 100 images, scanned from any 35mm image source – black and white or colour negatives and slides.
As well as writing the disc, the workstation is capable of outputting prints using a Kodak thermal printer. Kodak claims a quality close to or matching conventional colour enprints, but with greater flexibility, as the workstation allows you to crop and zoom pictures easily. Other image manipulation features provided by the software include panning, ‘airbrushing’, edge enhancement, cut and paste, montage and colour editing. The workstation also prints a cover sheet for the CD showing miniature versions of all the pictures it contains.
Initial cost for the CD with 24 pictures will be around L11. The cost of adding further pictures is lower, as you already have the disc. The CD is effectively like a WORM (write once, read many) drive: you can go on adding pictures to it, up to the limit of 100, but you can’t delete or overwrite any.
Kodak has joined with Philips to produce a special L300 player for the discs. Scheduled for June 1992, this will play ordinary audio CDs, but also has an RF output for a standard household TV. The disc is also compatible with CD-I and standard CD-Rom or CD-Rom XA players. That means you should be able to use the pictures in interactive video, multimedia and desktop publishing applications.
From Kodak’s point of view the system has one important advantage. While enabling the company to move into the world of electronic imaging, it can do so through its existing chain of photofinishers, whose business is expanded, rather than stolen, by the new technology. Similarly, end users can get into the world on on-disc images without buying a new camera.