There’s an exciting upheaval in the video world: Digital Video. In the same way that CDs replaced vinyl records with superior quality and the promise of perfect copies, DV will soon make analog video seem like a quaint backwater.

A new computer connector called FireWire is completing this digital utopia. When DV is combined with FireWire, you have everything you need to edit great quality videos. Of course, as with all technological marvels, there is a certain amount of trepidation involved. With this in mind, I asked my know-it-all friend Dr. Cathode a few questions about the technology:

What is DV?

Q: What does it mean to digitize video?

A: You know how video converts light into electricity, and then uses that to drive an electromagnet that records on tape? Digital is like that, but instead of using, say 3.69 volts to represent a particular spot on the screen, it uses digital numbers.

Q: How does that improve quality?

A: Analog’s fluctuating voltages tend to build up noise after several copies are made. DV stores numbers as a series of ones and zeros (called bits). The voltage difference between one and zero is big enough that noise can’t swamp it. The result is a pristine signal that looks good and is easy to work with. For the same reason, it’s also an excellent medium for archiving video.

Q: I’m still not clear about how you attach numbers to a video signal.

A: A timer is set to look at, or sample, the video stream 13.5 million times per second. With each sample, it converts voltage level to an 8-bit number called the luminance signal (or Y signal) which represents the intensity of the light. Every fourth count, it also samples two color components (called R-Y and B-Y ).

Q: So that’s what they mean by component video?

A: Exactly. Intensity and color are separate components of the signal.

Q: So that’s why DV looks so much better than VHS?

A: That, plus 500 lines of resolution: better than VHS, Hi-8, S-VHS or even Laserdisc. It’s actually comparable to Betacam – the format of the pros.

Q: But it must use more tape to record three digital numbers than a simple little voltage signal.

A: That’s true. There’s simply a lot of data in a video image. A 500*525 screen with full color sucks up three-quarters of a million bytes! Multiply that by thirty frames per second to get about twenty million bytes every second, uncompressed.

Q: Uncompressed?

A: Yes, to fit, the data gets compressed with an algorithm called DCT [see sidebar]. It brings the data rate down to about 4 million bytes per second (MBps). It’s a lot, but Sony and Matsushita have figured out how to shoe-horn it onto a tiny tape.

Q: And do they digitize the audio too?

A: You bet. Digital video sports a stereo track of 16-bit, 48kHz audio – better than CD quality! There’s even an option to use four sound channels instead, but the quality suffers a bit: 12-bit at 32kHz. That’s not as good as CD sound, it beats the heck out of ordinary VHS.

Q: How do they fit all that onto the tape?

A: Not easily. First they use very tiny recording heads – about a tenth the width of a human hair — and then they spin them at about 9000 rpm’s. As the head rotates, it lays down ten diagonal tracks per frame of video, with each track divided into four sections. These machines are miracles of engineering precision and reliability.

Q: What are the four sections you mentioned?

A: The two biggest sections are video and audio. Of the two, video is the largest. Then there is a section called subcode that has goodies like the track number, time code (for frame-accurate editing), date and time data, copyright information and more. The last section is called ITI for Insert and Tracking Information.

Q: It all sounds wonderful, but it seems like it’s pushing technology to the max. With such data densities, don’t you get problems like drop-out?

A: Um, yes, but being digital provides another plus. Error correcting codes can fix most localized drop-outs before you ever see them.

Q: Do these cameras also shoot stills?

A: Yes, with much better definition. A sixty minute tape can record 580 high-resolution images.

Q: Sixty minutes? I thought the tapes held hours of video.

A: The full-size professional tapes hold four and a half hours (fifty gigabytes), but consumer level mini cassettes can hold no more than an hour.

What is FireWire?

Q: How is FireWire related to DV?

A: FireWire and DV are two entirely separate technologies. However, they work together like magic. FireWire (or IEEE 1394) is a high-speed computer bus invented by Apple. Currently, it clocks in at 200 Megabits per second (Mbps) — more than enough bandwidth for video.

Q: You call it a bus. Is it like SCSI, with daisy-chains and terminators?

A: Yes, in a way. But you can connect 63 devices in a FireWire chain , not just seven, with no terminators. In fact, you can plug and unplug devices while the bus is running. That’s called hot-swapping, and it means you don’t have to power down your computer every time you want to hook in your camcorder.

Q: Is it easy to connect or will I have problems like I have with plug-n-pray?

A: FireWire does a great job of plug-n-play. When something is plugged in, it’s automatically recognized. When you unplug it, FireWire compensates.

Q: What does it look like?

A: FireWire is a lightweight cable that has six wires inside: two apiece for dual bi-directional lines and two more for current. The cable ends in what looks like a game-boy connector.

Q: Is FireWire really going to catch on?

A: You bet! A few of the companies developing products for FireWire include Apple, Sony, Matsushita (Panasonic), Texas Instruments, IBM, Yamaha, Toshiba, Philips, Compaq, Motorola, NEC and the very important Wintel duo: Microsoft and Intel. There’s a feeling that FireWire could replace all of the computer connectors we know in the not-so-distant future.

Q: How does it work to edit video?

A: You can use one of the bi-directional lines to handle both recording and playback functions. That cuts the number of cables in half. All the DV data are available, including subcode and ITI (insert and track information). It gets saved to your hard disk for frame-accurate editing.

Q: I bet you’re talking about a large hard disk.

A: Oh yeah. At 3.5 MBps, you’ll fill a gigabyte drive in less than five minutes. And if you’re planning on a lot of special effects, the frames need to be de-compressed into RGB before they can be digitally massaged. Thank goodness the price of drives is going down even as their speed and capacity goes up.

Q: Can you dump data directly to a hard drive?

A: That’s right, you don’t even need a computer. You can also copy from one camcorder to another with zero loss.

Q: What will this do to the consumer video industry?

A: Since its introduction, editing video has been a pain for everyone except the pros with fancy equipment. The world is about to be inundated by a new wave of videographers who will flock to the simplicity of digital editing and its cool special effects.

Thank you, Dr. Cathode. I feel I know a lot more about how DV and FireWire are natural partners in the brave new world of digital video.

Scott Anderson is the author of an animation program and a book about digital special effects.

[Sidebar]

Compression
Without compression, it could have been another decade before digital video became practical. Fortunately, mathematics turns out to be useful, after all. A mathematician named Fourier said "Any signal can be represented as the sum of a bunch of simple sine waves with different amplitudes and frequencies" (a loose translation). This is the basis of the Discrete Cosine Transform (DCT). The DCT is at the heart of the JPEG compression algorithm, used in the computer world to squeeze still pictures into a fraction of their original size.

The DCT splits an image up by frequency components and then tosses out certain ranges. The converted and compacted data is what gets stored on the tape. When the data is fetched off the tape, it is reconstituted by an inverse transform – similar to the way division can undo multiplication. DCT typically yields a 5:1 compression ratio, so the data goes down to about 3.5 megabytes per second. Still hefty, but manageable by current recording technology.

MPEG takes compression another step: if part of the image (a still background, for instance) doesn’t change from frame to frame, it doesn’t need to be stored. That compresses well, but makes it hard to edit. The current frame may depend on a hundred frames before it, slowing things down. DV engineers took the middle road, and compressed each frame independently. It still saves room, but keeps it easy to edit.

             –SA

LEAVE A REPLY

Please enter your comment!
Please enter your name here