It was like the passing of an old friend.
Well, it actually happened and the world did not come to an end – there was no blood in the streets, no popular uprising. High-powered analog broadcast television got shut down. Perhaps doing it in two steps helped. Many of those who were unprepared or uninformed earlier this year got the message and were ready for June. Those who were oblivious to what was happening got little sympathy since they had two chances.
Personally, it was like the passing of an old friend. Television has been my whole professional career, and analog television formed the bulk of it. Thinking about it, the real fun and excitement came as digital technology made inroads into the analog world and made all sorts of new things possible.
One of the personal high points was the EnCamera experience. EnCamera is the technology that allowed embedding 4.5 Mbps of digital information into the analog television signal without damaging the video. It came about because three of us were old enough to thoroughly understand analog television, yet young enough to be into digital.
The intersection of the analog world with the digital world made a robust transport of a significant data rate possible. We built a UHF television station in Scottsdale, Ariz., and had an experimental license from the FCC. That was more exciting than getting an amateur radio tag.
The technology and the company got sold in 2000 and became the basis for the MovieBeam system originated by Disney. A set-top box was sold with 100 movies encrypted on a hard drive. Ten new movies were trickle-charged onto the hard drive each week using EnCamera technology riding on the PBS and ABC signals. Purchases of movies were reported by dial-up. MovieBeam passed on before analog television did because the world didn’t need another way to consume movies. But the technology worked well.
The analog NTSC signal is, and will always remain, a wonderful series of examples of the engineering method of technical compromise. When the system was being formulated, engineers understood that the scarce resource was in the limited number of affordable active devices. The active devices were vacuum tubes. They were expensive, large and consumed a lot of power. Only a handful could be afforded in a receiver. So the compromise was to design a signal that required only a few active devices, even if that meant that the signal used spectrum inefficiently. What good would an efficient use of television spectrum be if no one could afford to buy a receiver?
The transistor as the active device solved a lot of the vacuum tube’s problems. But massive progress came when transistors were used for digital purposes.
It’s worthwhile trying to understand the fundamental reasons for digital’s impact. There are two major factors: electronic memory and Moore’s Law.
The main thing analog circuits have great difficulty with is memory. There simply is no method for storing a lot of information in analog circuits in a manner that allows rapid access and low cost. And memory and storage are the elements that make so much of digital’s magic possible. Digital television squeezes 10 or more standard-definition signals into the same 6 MHz used for one analog signal because of digital processing. Redundancy is removed by comparing pictures to find elements that are the same. Then only the differences need to be transmitted. That is only possible if the pictures can be stored and rapidly accessed. And the more pictures that can be stored and compared, the more redundancy eliminated.
The other digital enabler is Moore’s Law. Gordon Moore of Intel observed that digital transistors keep getting smaller and faster. Every 12 to 18 months, the number of digital transistors that will fit on a given piece of silicon doubles. Cost is strongly related to the size of the silicon chip. So, for a given cost, the number of transistors available to do processing doubles every Moore cycle. By the same token, the cost of a given number of digital transistors is cut roughly in half every 12 to 18 months. So you can take your Moore dividend either way, or in some combination. And smaller transistors run faster. So processing speed increases, as well. This benefit does not accrue in the same way for analog transistors. Only digital transistors enjoy the full benefit of Moore’s Law.
The consequence of all of this is that we can afford to put billions of transistors in inexpensive consumer electronics products. So the scarce resource is no longer the number of affordable active devices.
And there is another deficiency with the old vacuum tubes. They have a built-in wear-out mechanism. Transistors do not have a similar wear out problem. An important aspect of the vacuum tube wear-out mechanism is its statistical nature. A system with thousands of vacuum tubes fails when just one, or a few, tubes wear out. Statistically, it means that it is not possible to run a large system of vacuum tubes for very long without having a continual massive maintenance problem.