By Mark Schubin, February 29, 2008
Are advanced video codecs twice as good as MPEG-2? To figure out the answer, it might be worth considering some other statements.
Bell did not invent the telephone. Edison did not invent electric light. And Sony didn't come out with the first camcorder. Or maybe they did. It all depends on absent adjectives.
The issue is not (or at least not entirely) about the technological development timeline. Consider Bell and the telephone.
Bell's earliest transmission of any form of sound via electrical apparatus was in 1875, but, at least as early as 1821 (long before Bell's birth in 1847), the young British scientist Charles Wheatstone was giving "telephone concerts" in London. These consisted of the sounds of such instruments as a piano, a harp, and a dulcimer emerging from an untouched "enchanted" lyre hanging from the ceiling.
The sounds were actually conducted to the lyre by the "cord" from which it was suspended, actually a solid rod connected to instruments played in an unseen room. Although Wheatstone called the lyre an "aconcryptophone," he noted that the same principle could be used to transmit sounds at a distance, even from London to Edinburgh. A device for delivering sound at a distance is a "telephone" (Wheatstone also invented a stethoscope-like device for amplifying sound and called it a "microphone"), and the term was commonly applied to all sorts of purely acoustic devices, from funnel horns for the hearing impaired to speaking tubes used on ships.
So Bell could not have invented the telephone. What about the electrical telephone?
Certainly, he was not the first to dabble in the field. In 1836, Charles Page, a U.S. patent examiner, published a paper on the transmission of sound over electrical wires. In 1849, Antonio Meucci, an Italian-American whom the U.S. House of Representatives recognized in 2002 for "his work in the invention of the telephone," demonstrated the "talking telegraph." Other inventors followed: Charles Bourseul in France, Johann Philipp Reis in Germany, Innocenzo Manzetti in Italy, Elisha Gray and Thomas Edison in the U.S., and, perhaps after Bell, Emile Berliner, Francis Blake, Amos Dolbear, and others.
In fact, one of the longest-running patent battles in U.S. history involved Bell, Berliner, Blake, Dolbear, Edison, Gray, C. H. Chinnock, A. G. Holcombe, J. H. Irwin, J. W. McDonough, G. M. Phelps, C. A. Randall, G. B. Richmond, W. L. Voelker, the American Bell Telephone Company, the Clay Commercial Telephone Company, the Molecular Telephone Company, the Overland Telephone Company, the People's Telephone Company, and the Unites States Telephone Company, among others.
Western Electric, long the manufacturing arm of the Bell System, was actually incorporated by competitive inventor Gray in 1872 and sold to telegraph-service provider Western Union in 1875 (which ran its own telephone business until losing a patent dispute in 1879). Bell acquired Western Electric in 1881.
So, why is Bell so often credited with the invention of the telephone? Interestingly, his work that led to his version of the telephone wasn't even intended to transmit speech; all he wanted to do was to create a "harmonic telegraph" that could carry more data down a single line than could Edison's quadruplex system. Nevertheless, he probably was the first to make a long-distance phone call (as early as 1876), and AT&T's gigantic Bell System was a direct result of his efforts. If Bell didn't come up with the first electrical telephone, he helped make telephony practical.
Edison invented many other things besides the quadruplex telegraph, but the first electric light was not among them. Edison's first patent for an electric light was filed in 1878, but, in 1849, at the opening of Giacomo Meyerebeer's opera Le Prophète, the Paris Opera used electric lighting to simulate a sunrise. And even that was not the first public demonstration of electric light (never mind nature's lightning or static-electrical sparks).
In 1802, Humphry Davy passed a current through a thin strip of platinum, causing it to glow. It was probably the first public demonstration of an incandescent electric light (and of an electric toaster element). Later, Davy demonstrated the arc light that was used at the opera.
So Edison didn't even invent the first incandescent electric light. The light bulb? No. That had been demonstrated by 1840 by British scientist Warren de la Rue.
What Edison came up with was the first practical electric light bulb, one that provided plenty of light, lasted a long time, and was relatively inexpensive and easy to use. And at least some of that was due to the work of Lewis Latimer, an African-American member of "Edison's Pioneers," who improved the filament, created the screw-in base, literally wrote the book on Incandescent Electric Lighting, and, coincidentally, also drafted the Bell telephone-patent drawings.
The history of technology is filled with improperly assigned "firsts." Marconi's "first" wireless transmission of 1895 was preceded by, among others, Bell's wireless transmission of clearly recognizable voice by light in 1880 and by other means in 1882. But it was Marconi's transatlantic radio transmission of 1901 that was honored by a Royal Mint coin in 2001 (although there's some dispute about even that in radio-historical circles). And it was concern by the U.S. Navy about Marconi's practical development of long-distance wireless communications in a British-based company that spurred the creation of the Radio Corporation of America (RCA).
Given those early wireless experiments, it is, of course, not true that KDKA in Pittsburgh was the world's first radio station. It was simply the first commercial broadcast station licensed by the United States.
RCA -- the old RCA -- is back in the news. A hot ticket on Broadway these days is The Farnsworth Invention, a play written by The West Wing creator Aaron Sorkin. The Farnsworth of the title is Philo Taylor Farnsworth, the invention is that of television, and the play covers imagined interaction between Farnsworth and David Sarnoff, head of the gigantic RCA
. Unfortunately, Farnsworth didn't invent television.
First, there's that absent adjective, again. Before Farnsworth's first television pictures, John Logie Baird had already demonstrated recognizable video images of a human face, and silhouettes had been transmitted before Baird's achievement. But Baird (and most who came before him) dealt with television cameras and displays that required spinning disks or drums to work. Today we refer to their developments as "mechanical television." At the time, however, they were just television, and Farnsworth dabbled in all-electronic television.
Unlike Bell, Edison, and Marconi, Farnsworth does seem to have been the first to achieve that for which he is famous. As best is known, before anyone else, he demonstrated an image captured by an all-electronic video camera on an all-electronic display. But he wasn't the first with the idea.
Many years before Farnsworth, Alan Archibald Campbell Swinton presented and published his concepts for all-electronic television -- quite similar to Farnsworth's -- in Britain, and Campbell Swinton's ideas were preceded by pictures actually displayed electronically by Boris Rosing in Russia. And, despite many useful improvements he developed for television, Farnsworth's camera, lacking the image-storage principle developed by Hungarian Kalman Tihanyi (another claimant to the concept of all-electronic television), wasn't practical.
If invention requires either being first with an idea or coming up with a practical solution, then Farnsworth, despite his earliest demonstration, isn't the inventor of even all-electronic television. But television probably has no single inventor, and Farnsworth made important contributions to the field.
Similarly, Sony made important contributions to the camcorder, including an early demonstration of a consumer prototype and the introduction of Betacam, the workhorse of the news industry for many years and the progenitor of Betacam SP, Digital Betacam, HDCAM, Betacam SX, MPEG IMX, and HDCAM SR (so far). But the first camcorders to be sold were the so-called Type M series (Panasonic's Recam and RCA's Hawkeye), introduced one year before Sony's Betacam. Neither caught on the way Betacam did, which is why Sony can be said to have created the first long-term successful camcorder, and Panasonic's term for a recording camera, Recam, never replaced camcorder, the earlier term introduced by consumer-electronics writer and editor Dave Lachenbruch (who also coined the term consumer electronics).
The preceding historical excursions might be entertaining, informative, or both. They're intended to offer a lesson on linguistic precision. Bell can be said to have invented the practical telephone system. Edison can be said to have invented the practical light bulb. And Sony can be said to have introduced the first long-term successful camcorder. Without the adjectives, the statements are wrong.
Consider some other statements. Canon's XH A1 camcorder has a minimum illumination of 0.4 lux. Discs are faster than tape. H.264 compression is twice as good as MPEG-2. Are they true? Or are they missing modifiers?
Minimum illumination, all by itself, is meaningless. What is being illuminated? A shiny white sequin or a cloak of black velvet? What video level is the minimum illumination supposed to achieve? At what signal-to-noise level?
To say that a camera-lens combination has a minimum illumination of 0.4 lux of 3200K creating 50-IRE video from an 18% reflective test chart with the shutter off at f/2 at minimum focal length at a signal-to-noise ratio of 56 dB conveys some useful information. Without the qualifiers, no useful information is conveyed; any camera can have any minimum illumination.
As for the second statement, it's certainly likely that one can randomly access information recorded on a disk faster than getting to the same information recorded linearly on tape. But a VHS recording is done and ready to be taken away as soon as the eject button is pushed. Some DVD recordings require a finalizing stage that can take quite some time.
So many television stations have multi-standard Betacam-based players that video news releases may be delivered on a Betacam-type tape with reasonable assurance that they can be played right away. Can the same be said of a P2, CompactFlash, or SxS card or a Rev Pro or XDCAM disk? There remain a number of situations in which tape is faster than disk, not even counting system crashes or format obsolescence. Again, however, for random access in an equipped playback system, disk is faster than tape.
Then there are those bit-rate-reduction (or compression) systems. Comparing JPEG2000 to MPEG-2 is like comparing apples and oranges; the two have many differences. H.264, also known as AVC and MPEG-4 Part 10, is easier to compare to MPEG-2.
That doesn't mean they're directly comparable. First, which form of MPEG-2? Which profile? Which level? Which form of H.264? What data rate? What source material?
Essentially any compression system can deal perfectly with a static signal like color bars at almost any data rate. And, with a high-enough data rate, essentially any compression system can deliver perfect pictures. How can one system be twice as good as another when both are perfect? Only by fully defining the source material, the data rate, the coding parameters, and the meaning of "twice as good" (a standardized average rating by expert viewers? non-expert viewers? some objective measurement? which?) can a statement comparing two compression systems be meaningful. That's why there's so much argument among compression-system proponents.
Precision might seem picky, but it leads to progress. Useful minimum illumination, for example, probably won't improve if we don't know what it is.
Would you rather shoot with Edison's light, Bell's microphone, and Farnsworth's camera? Hmm. Come to think of it, that could be a lot of fun!
.
|