Forgotten old movie tech

Last night, Scarlet and I went to Davies Symphony Hall and watched the 1925 film "The Phantom of the Opera" (starring Lon Chaney) with the original accompaniment score performed on solo organ (and Foley board) by Dennis James. Mr. James played the renowned Ruffati organ installed in Davies 25 years ago this year. It is the largest concert hall organ in North America. It was a spectacular evening, but the real highlight for me was a re-introduction to early motion picture technology.

Advances in the state of the art in any field always obsolete prior technology, resulting in the loss of specialized techniques used to optimize the technology of the time. Case in point: the iambic keyer. It represents the state of the art in optimizing the transmission of morse code. Morse code is an obsolete technology, relegated largely to the amateur radio bands nowadays. With the advent of satellite based search and rescue beacon technology, the last non-amateur use for morse code (namely the maritime service) has gone by the wayside. Apart from Amateurs, nobody therefore has a use for an iambic keyer. And I think it won't be too long before you'll need to go to a museum to see an Iambic keyer - the furthest development of a technological cul-de-sac.

Thus is it with film. When we think of early films, we think of black-and-white silent films. We think of that because we're used to television, and before color television, TV was itself monochromatic (black-and-white is a misnomer: both TV and film offer a continuous greyscale). But, as I discovered last night, early cinema was not monochromatic. Early TV was monochrome because the actual color that the viewer saw depended on the color of the phosphors that were built into his own TV set. Thus, everybody saw exactly one color - which tried to be as close to a neutral grey as possible.

This was not the case for film, however. While the actual photographic process was greyscale, the film stock itself could be tinted. Within the single film we saw last night, I counted at least 3 different film-stock tints. These differing tints were used by the film's creators to change the tone of the scenes. This is something that was impossible for television before the advent of full color broadcasting in the 1960s. When you saw a film on TV, it was greyscale, period (unless you put colored films or other such trickery in front of the tube).

Not only that, but certain scenes in the film we saw last night were actually in full Technicolor! Color photography was in its infancy in the 1920s. It was nightmarishly expensive, but it could be done. In addition, it was possible for much less money to highlight a single color - the Phantom's red cape, for instance - in a particular scene. This was also done in the film we saw last night.

In addition, it was not unheard of for some filmmakers to have certain elements of their films hand tinted. An example of this is still preserved today in the Chriterion Collection edition of Jacques Tati's film Jour de Fête, where the french flag is tinted red and blue.

Lastly, before the advent of synchronized soundtracks, it was customary for the projection frame rate for films to be variable. Usually, instructions were provided to the projectionist along with the reels of film for what speed various scenes were to be shown. Sometimes the projectionist would ignore those instructions and do whatever they felt was right (or perhaps they were just lazy and set one speed at the start). Because of that, individual experiences in viewing a single film could actually vary. Essentially, there is a tradeoff between slower speeds that flicker a bit less and use less feet-per-minute of film, versus faster speeds that make fast action less blurry.

With the advent of synchronized soundtracks, it was necessary to stick with a standardized frame rate (the industry chose 24 fps) to insure that the pitch of the sound didn't vary, but at the same time, one of the tools used for decades to customize the performance was lost.

With modern digital video technology, resolution and frame rate are, once again, adjustable. There's nothing that would prevent someone from varying the frame rate by scene. But the problem is that it likely wouldn't do any good, since most displays simply adapt the incoming material's frame rate to the native refresh rate of the display. Making matters worse, some displays either do a lousy job of this, or fail if faced with non-standard refresh rates. For instance, most cartoons are animated at only a maximum rate of 12 fps (with adjacent frames of 24 fps film being identical), and that's only during action sequences. The reason for this is the enormous cost of animation. It would make sense, therefore, to MPEG encode such cartoons at 12 fps. But this typically isn't done. Instead, the encoding is done at 24 FPS and redundant, empty I frames are sent in the extra time.

With the advent of television and full color movies, these techniques were rendered obsolete. In the case of television, the viewer could only see monochrome anyway, and in the case of movies, full color made the other tricks unnecessary. It is only in experiencing what must be characterized as an early cinema museum performance that we in the audience were privileged to get a glimpse of the highest state of the art of early cinema.

A quick tour of bay area TV via spectrum analyzer

Playing around with the new spectrum analyzer.

Solving the CableCard / OpenCable conundrum

One issue I've always had with satellite TV is that before recently, every tuner has required so-called "home run" wiring. That is, a coax run all the way from the receiver to either the dish or a multiplexer, with no splitters in-between. The reason for this is that there needs to be a communication path from the tuner back to the dish (or multiplexer) so that it can select which set of transponders it wants to tune at that moment. If you've got a couple of dual-tuner DVRs, that's quite a lot of coax. The latest little innovation from DirecTV is the SWM-8, which neatly solves this problem. It allows you to connect up to 8 tuners using as many splitters as you need (with the caveat that you need to have at least one branch of the splitters pass DC power from an indoor mounted power supply module). The way it works is very similar to SDV cable TV systems. There are 9 channels set aside - 8 of them provide one tuner's worth of signal down from the dish, and one is a shared upstream request channel from the tuners back to the SWM. Each tuner requests a channel to be allocated to it from the SWM at startup, and as the channels are changed, it requests a different satellite transponder be sent down its allocated SWM channel.

With this system, the satellite system itself can become arbitrarily complex, but the complexity can be hidden behind the SWM. Already the SWM can handle a single 5 LNB Ka/Ku dish plus two auxiliary dishes (one for international programming and one for carrying local channels for some markets).

As I've previously noted, this system is somewhat similar to SDV systems for cable - the only difference being that SDV channels are shared by all subscribers, not allocated solely to a single downstream tuner.

Verizon's FIOS offering already includes a box that transitions from fiber to coax for the household. In principle, that box could also provide addressable decryption so that the house simply sees all of the subscribed channels as clear QAM - no need for CableCard at all.

This same idea could apply in the world of cable TV as well. Each household would have an addressable box that would perform pre-tuning and decryption for a set of tuners within. There would be a spec for the channelization and for the upstream communication protocol. The channel data would be clear QAM. The same box could also have an Ethernet jack and provide IP service as well, for subscribers to cable-delivered Internet. Since the box obscures the cable infrastructure, it offers greater flexibility to the cable company in laying out their networks.

Wii would like to play

We bought a Wii last night. The concept of position and motion sensitive controls is truly revolutionary. Nothing that has come before has been as immersive. Our Wii came with the Wii Sports title, which includes Tennis, Baseball, Golf, Bowling and Boxing.

The baseball game is sort of one dimensional and boring. All you do is pitch and bat. It's a lot more like Over The Line than Baseball. I don't like it so much. Tennis is a bit more challenging. Tennis is played as doubles, and in addition to swinging the controller at the right time, you need to pick whether to swing forehand or backhand correctly depending on where the ball is. Golf is kind of fun. You can't swing too hard or it will just accentuate any hook or slice you've imparted on the ball, so controlling your swing is what it's all about. Boxing is played with the numchuck, which is a second controller wired to the first one (the primary controllers are wireless). The numchuck also has a position sensor in it, so with the controller in one hand and the numchuck in the other, you throw your fists around like real boxing.

But the pièce de résistance is Bowling. You swing the controller just like it was a bowling ball, and let go of the trigger button to let go of the ball. The english you put on your swing gets imparted to the ball. It has to be the single most accurate sporting simulation since the treadmill.

Scarlet and I really got some exercise last night. We bought a Wii because we decided to see if it really was true that you could get into playing the games as an exercise program. Well, the hype is true. I think this is going to be very good for us.

The only issue now is that I'm not sure what other games we'll want to buy. I'm sure we'll get tired of Bowling some day, but I can't really see that day from here.