It's been a hectic week:  we're getting part of the station's roof replaced, which should be a good thing.  The problem is, there are a number of satellite dishes sitting on that part of the roof -- so we've been hustling to relocate the dishes to better permanent locations on the ground.  It's involved digging a number of holes, hauling ten-foot sections of mast pipe around, mixing thirty-six bags of concrete (80 pounds apiece), and horsing the dishes down off the roof and setting them on the new masts.  One of the dishes is moved, three to go.  It would have been done by now, but we've been delayed by everything from unusually hard ground to having to repair several cables from our ABC network system.  Suffice it to say that Heliax transmission line is no match for a roofing screw.

Sigh.

I've gotten some interesting questions and comments relating to how DTV works in general, and about PSIP in particular... so for those who are curious, we'll take a look at the basics from a non-technical perspective.  Today, let's start by talking about packets.

Digital TV works by sending a constant flow of data -- over 19 million data bits every second.  This flow is organized into packets, each of which consists of 188 bits; each packet might contain a piece of picture, a piece of audio, a piece of a program guide, or some other service.  The first important point to catch is that these packets show up at your receiver at a regular rate.  It's rather like an escalator: it moves at a constant rate of so many steps per minute, no matter how many people happen to be riding it at the time.  You might have someone on most every step, or maybe some of the steps will be empty.

Just as the escalator has an upper limit to the number of passengers it can move during a given time, digital TV also has an upper limit to how much data it can transmit.

The second key idea about packets is that each one has an identifying number that tells your receiver where to send the data: to the video decoder, to the audio decoder, and so forth.  These packet ID numbers are called PIDs (original, huh?), and range from 0 through 8,191.  It's rather like your phone number:  it tells the phone company switching equipment where to route your calls.  Some PIDs are reserved for special purposes, and there are standards for how others should be organized.  We'll take a closer look at this at some other time, but right now I want you to meet PID 8,191:  it's called a "null" packet.  As you might guess, a null packet contains -- trumpet fanfare -- nothing.  It's just a space filler when there isn't something else to transmit.

Let's ride with this escalator notion a bit more... imagine that you have a bunch of buckets, and you stick one on every step as it goes up.  (You guessed it:  each bucket is a packet, and it has a marking on the outside that says what's in the bucket -- that's the PID number.)  Someone is grabbing buckets full of video data, or audio data, or electronic program guide data, and putting them on the next available step as it appears.  And if a step comes along but no filled bucket is ready to go, he sticks an empty bucket (yes, a null packet) onto the step.  As the escalator makes its way up, every step has a bucket on it:

That's about as far as I want to take this for now... for one thing, I've got satellite dishes to move today.  Also, someone is going to bring up what happens to the buckets when they get to the top:  do they pile up like impatient commuters behind a gawking tourist until they finally burst forth in an agitated blast of humanity, or do they fall off the end like candies at the end of Lucy's conveyor belt?  (Nothing like citing a TV show from the days of black-and-white as an illustration of DTV more than 50 years later!)

I've avoided real work long enough.  Have a good weekend!

-- Jeff