How DV Works: Technical Feature

Ladies and gentlemen, step right up! Inside this tent you’ll have a remarkable opportunity to get closer than ever before to digital video, otherwise known as DV. I’ll give each of you an unusual close-up look at the mechanics of DV, at the various DV formats on the market and at the reasons DV can do so much, so well. So follow me into the tent of wonders!

Look – the journey is already beginning. We’re now shrinking, small enough to penetrate the inner workings of a DV camcorder. Let’s enter through the lens housing and start looking around.

Light, Sound and Current

As we move through the zoom lens, note that at this point, digital video is a lot like analog video. Light and sound enter the camera through a lens and microphone and then a computer transforms the real world into electronic signals.

Digital and analog part ways fairly soon, however. The tiny silicon charge-coupled device (CCD) at the end of the lens barrel uses hundreds of thousands of pixels to make DV look incredibly sharp and clean, with around 500 lines of potential resolution (or more, in three-chip pro cameras).

From Analog to Digital

Next we come to the circuit boards, which do an enormous amount of the work of making your DV footage look and sound fantastic. The software coding and computer components contained in the boards produce a digital replica of each moment of video and audio in the analog-to-digital conversion process. There is also circuitry that works in reverse, for playback on your television. It’s the "digital" part of DV that puts this technology head and shoulders above consumer analog video formats. Digital video is pure data, not analog signals, allowing pristine and endlessly repeatable transmission of high-resolution data through an all-digital pathway.

Doing the Math

All consumer digital video formats (Mini DV, Digital8, DVCAM and DVCPro) utilize the same basic data format and data rate (25Mbps) to encode and decode 30fps NTSC video data.

  • Sampling DV encoding hardware samples each frame of video for luminance (brightness) and chrominance (color) information. It uses 4:1:1 (Y:U:V or YUV) sampling for this operation. The hardware scans each line of every 720×480 video frame, taking four pixel samples of luma information (Y) for every one pixel sample it takes of chroma information (U and V). That cuts down on extra data and also provides the right mix of luma and chroma detail to satisfy our eyes, which are more sensitive to brightness (luma) than color (chroma).
  • Compression The DV brain then mathematically compresses each resampled frame of video to speed throughput and save storage space on tapes and hard drives. This is accomplished with a 5:1 DCT (discrete cosine transform) mathematical algorithm that discards as much unnecessary image information as possible while retaining much of the quality of the original image.
  • Audio A separate sampling process takes the audio signal (after pre-amplification) and turns it into data as well. An audio sample rate of 48kHz (with a 16-bit depth per sample) produces a single track of high-fidelity digital stereo audio (2 channels). Alternately, a 32kHz sample rate with a 12-bit depth yields two stereo tracks (4 channels total), one of which can be used for voiceover narration.
  • Vital data All of this pristine but compressed digital information is bundled with additional vital pieces of generated data. This information includes time code, time/date information and digital pilot tone signals to replace the conventional control track of analog video, which the DV format lacks.
  • Error correction Also added to the data mix are error correction bits. Digital video data travels in tiny packets and the DV hardware adds unique codes that verify and correct corrupted data bits.

    Express delivery

    The whole package is finally bundled in data packets compliant with the DV standard. Every one of these packets – each the size of a single DV track – contains four independent regions: a subcode sector for time code and other data, a video sector, an audio sector and a sector for insert editing and track data. These packets move at a rate of 25Mbps (megabits per second), which translates to roughly 3.5MB of disk storage space per second of DV video.

    Where the action is

    We’ve seen the brains, but now we’ve come to the brawn – the spinning drums that record data onto the tape and read it off. The drum that houses the heads is a polished metal cylinder that’s angled in the cassette compartment and rotates at a very high rate. Rollers hold the tape against the drum’s grooved surface, where a number of electromagnetic heads make slanted swipes across the surface of the tape, recording tracks of data that correspond exactly to the DV packets described above.

    Everything about this system is microscopic and is measurements are in microns, or thousandths of a millimeter. In fact, the record heads are so small, the tracks are so narrow and the data they contain is so densely packed, that a minute of digital video – about 200MB of information – occupies less than two meters of tape. Put another way, a DV cassette can hold about 13GB of digital information.

    The Skinny on Digital Formats

    Everything up to this point is common to the 25Mbps Mini DV (DV25), Digital8, DVCAM and DVCPro formats. When it comes to recording to tape, however, manufacturers have developed several different ways to store the data.

    Mini DV tape comes in a 55mm wide plastic cassette to fit consumer camcorders. The tape itself is 6.35mm wide and is coated with metal that was deposited using an evaporated processing technique (ME). It moves at a rate of about 19mm per second, with a track width of 10 microns. The typical 60-minute Mini DV cassette is about 70m long and stores around 13GB of data. The closely-related standard DV tape (designed for use in VTRs) is the same tape format, but comes in a cassette that’s twice as big and holds as much as 180 minutes of tape.

    DVCAM (Sony) and DVCPro (Panasonic) formats are modified DV25 for the professional market. They use a wider track pitch for greater reliability and move the tape past the heads much faster. Both formats offer as much as three hours of running time on a single cassette.

    The DVCPro format has several other pro-level features. DVCPro tapes use a metal particulate (MP) process instead of ME. Unlike Mini DV and DVCAM, DVCPro can use optional linear tracks at the top and bottom edges of the DV tape to record analog time code and audio information.

    The Digital8 format also has idiosyncrasies. Larger than Mini DV tape, Digital8 records onto 8mm and Hi8 tape. The key difference is that in a Digital8 camcorder, the tape moves twice as fast as in its analog relatives, and the signal is digital. The Digital8 format is backward-compatible with the analog 8mm format. That’s a big plus if you have a closetful of legacy 8mm gear and tapes.


    Most of what you’ve seen here applies to digital VTRs as well. Decks use the same processing system as camcorders to understand analog and digital signals. Where decks differ from, and usually outshine, camcorders is their mechanical robustness and their multiple digital and analog input/output capabilities. These units also offer format cross-compatibility: many DV, DVCAM and DVCPro decks can play back all of the different tape formats.

    Output and Beyond

    We conclude our tour at the camcorder’s FireWire connection. And that’s appropriate, because FireWire is a big part of DV’s success. FireWire is a data transfer protocol like USB or Ethernet. FireWire moves dense packets of data at extremely high rates, and that makes it perfect for moving DV data between camcorder and computer. The DV/FireWire one-two punch has created a real revolution in consumer video, enabling all-digital desktop video production.

    Parting advice

    As you leave this tent of wonders, make sure to remember that all DV equipment, from the most economical camcorder to the most elaborate high-end VTR, makes use of basically the same computational brain. And as for the many differences, don’t worry about the underlying technology. Whether your DV comes in the form of a pro camera with a giant lens, a portable deck with an LCD screen or a palm-size camcorder for travel, there’s a digital heart in all of these devices.

    [Sidebar: Standards]
    FireWire, like DV, is an international standard (IEEE-1394) that technology manufacturers have agreed to abide by in the interest of compatibility. But that doesn’t mean that everyone agrees on what to call the standard. Apple Computer, which played a large role in developing the technology, named it FireWire and Sony dubbed their version i.Link.

    [Sidebar: Longevity]
    Anyone who’s used analog formats has seen dropouts and other signs of signal loss resulting from faults on the magnetic tape or from recording problems. When it comes to longevity, expect analog tapes to last at best 15 years before they start to degrade. Error correction built into DV eliminates many dropouts, but what about longevity? Since DV tapes use magenetic material to record data, you’ll see gradual deterioration of these signals, too, though error correction can make up for some loss. Fortunately, it’s easy to make perfect backup clones of DV tapes via FireWire.

  • Videomaker
    The Videomaker Editors are dedicated to bringing you the information you need to produce and share better video.

    Related Content