Technology in Transition: Video storage

Video storage

By John Luff

When commercial broadcasting began, arguably at the 1939 World's Fair in New York, there was no effective means of routinely recording the signal for later playback. The most common method of the era was to point a film camera at a special picture tube and make a “Kinescope” recording on film. The electronic media captured film images. Isn't it interesting that film is now shot on video cameras and printed to moving chemistry?

Figure 1. Side view of a magnetic recording head. In both tape and disk recording, the signal passing through a magnetic field affects the arrangement of magnetic particles on the recording medium.

Television is a medium of considerable immediacy, but it is steeped in technology that often shapes how the media is used. When Ampex demonstrated practical commercial video recording in 1956, television was changed in material ways. Archival copies on film need not be created, for an all-electronic medium could produce results that were higher in quality. Moreover, they could be played back as fast as the tape they were recorded on was rewound. Reruns could be almost as good as the original broadcast, and time zone delay could be achieved without recreating a live event a second time. To broadcast engineers of the time it must have been quite a marvel. Heads spinning at 14,400 RPM and sophisticated signal systems were required. Those who might faint at high costs would have been well advised to take a large aspirin. Twenty years later quadruplex recording reached the height of its sophistication with the Ampex AVR-1 recorder, which virtually eliminated the artifacts that were inherent parts of video recording to that point.

Today we largely take video storage for granted. Home recordings can rival the quality of the first professional ENG recordings done on Umatic tape. The technology of tape was well understood when operators had to make routine adjustments to record drivers and playback amplifiers. Both tape and disk recording use the same principals from high school physics. If you pass a narrow gap wrapped by coils that induce a magnetic field in the area of the gap you can create alternating patterns of “north” and “south” poles in the thin recording medium. The recording medium is a material with magnetic particles either deposited in a vacuum on the substrate (typically a Mylar base layer), or painted onto the surface in a suspension of magnetic particles in a binder layer. (See Figure 1.) If the wavelengths to be recorded get shorter the gap must be narrower, and generally the coating thinner on the tape.

Similar basic physics govern disk recording. The limits of modern storage technology are mostly driven by evolving methods of placing the recorded patterns on the medium and then reading them back accurately. Quadruplex and other early tape formats separated the tracks by guard bands to ensure they could be uniquely identified. A control track was used to synchronize the movement of the head across the tape with the location of the tracks made as the tape moved through the machine. Over time the guard bands got narrower and eventually disappeared as some formats adopted polarizing techniques to allow tracks that are essentially overlapping to be read back. In addition to narrow track pitch, the gap width, one of the key factors determining the minimum wavelength of the recorded signal, got narrower. The width of the pole piece also got narrower to allow even narrower tracks.

Storage media density

The net effect of all factors is the steady improvement in packing density on the recorded medium. Tape formulations have improved, to a large degree because of the research dollar poured into consumer electronics research where payback can be huge. The storage capacity of modern digital formats varies from about 11Gb/hour for DV up to about 140GB/hour for D-5. This would not be possible without all of the research into the fundamental physics of recording. It is worth noting that the same facts that make video recording a lucrative market for manufacturers make hauling digital video over WAN networks very appealing to data carriers. We have a dense medium that gobbles resources.

All of this leads to a density per cubic inch that varies from under 1GB/in3 to about 2Gb/in3. Contrasted with disk recording, tape is far from being the king of the hill. The disk in the laptop I used to write this article stores a relatively low 12GB, but in only .5 cubic inches. The drives being delivered in many servers today hold 181GB in about 22 cubic inches, for a whopping 8GB/in3. A DVD is only about 80 percent of that density when removed from the machine and considerably lower when counted with the playback drive.

But the trend is clear. The 181GB drives are the same form factor that 20MB drives were only a few years ago. That 90 times improvement is pretty incredible, but experts in disk drive technology predict we may soon see the leveling of the curve. The cost of drives drops like stones, with each new generation of drives achieving the same sub-$200 price point in a time period that approaches a year. In terms of cost of storage though, digital tape, and especially videotape, is an amazing bargain. A $125 D-5 tape manages under a dollar per GB, while the modern high-speed 181GB drive costs about five times as much per GB.

Storage options

The result of this admittedly back-of-the-envelope, order-of-magnitude evaluation, is a simple fact. The death of videotape, though long predicted and some think feared by manufacturers of videotape recorders, is still a long way into the future. Disk drives get better every year, and closing the gap is now possible in the next few years. But if you own thousands of hours of programming it is still far cheaper to store the content on tape, even if you must access it under control of a media asset management system. When will tape die? Perhaps in a decade when optical storage reaches high enough capacity and access speeds approaching modern disk drives. That will not be soon either.

Researchers have their sights set on all kinds of technologies that may well take off this decade. Volumetric optical storage and a host of other rather complex but promising technologies could provide highly reliable and cost-effective storage for video.

I left out one medium that is critical to the future of television. It holds great promise of the highest image quality and reasonable storage density. It interfaces reasonably well with HDTV applications and has very long demonstrated shelf life. It is also ubiquitous on all continents, available for recording, archiving and display use. In fact most (or at least much) of television has been using it for decades. Of course it is the medium this article first reflected on, film. Film has a vibrant and important place in current and future recording and production, and it may outlive all current videotape and optical technologies. What goes around comes around.

John Luff is senior vice president of business development for AZCAR. To reach him,

Home |Back to the top|Write us

Send questions and comments