DIGITAL VCR FORMATS
For most people, the word format conjures up the image of a little rug near the entryway, or a message addressed to someone named Matthew. For someone choosing a digital VCR, format is perhaps the most important and far-reaching decision a buyer can make.
Before progressing in inch further, let's clarify that we're not discussing DTV (Digital Television) here. DTV, the new method of transmitting TV to our homes sometime between now and year 2006, also has a number of standards and formats. They are all devised to deliver 19.2Mbps (million bits per second) of MPEG-2 compressed data via satellite/cable/broadcast venues, resulting in one channel of HDTV, or maybe five channels of SDTV (Standard Definition TV) plus maybe some Internet data. DTV and digital VCRs may someday merge into one subject, but not quite yet. For this article were talking about VCRs that start with analog or digital video and audio, and record the signal digitally (as streams of ones and zeros), and can play back the data, converting it into analog video and audio.
Let's start calling digital videocassette recorders DVCRs. Just as analog VCRs have different formats to describe their cassette size and method of recording, so do DVCRs. One difference is that when an analog VCR puts out composite video, component video, or Y/C video, any other VCR with the appropriate inputs can use the signal; the signals are standardized.
DV signals are also standardized, but there are quite a number of standards, making it likely that the digital output of one type of DVCR can't be understood by the digital input of another.
To make sure one's output is compatible with another's input; make sure:
1. The inputs and outputs match, i.e., both are serial (all the data flows down one wire) or are parallel (the data flows down several wires at once).
2. The connectors match. To name a few, there are IEEE P1394 (Firewire), ATM, DS3, ESCON, FIDDI, Fiber channel, HiPPI, OC-3, SSA, SDI, SDTI (CSDI), SDDI, 4 forms of SCSI, and 5 forms of ethernet. SDI is an industry standard serial digital interface (noncompressed). D1, Digital-S, Digital Betacam, DVCAM, DVC PRO and PRO-50, and Beta SX all have SDI digital input/outputs. SDTI is a compressed digital interface allowing high-speed lossless data transfers to servers and DVCRs. Digital-S and Beta SX also use SDDI. DVCAM and some DV use IEEE P1394 (Firewire).
3. Both are using the same country standard (i.e., NTSC or PAL).
Outside of these standards complexities,
DVCRs behave much the same as their analog brothers. For one DVCR's
cassette to play in another DVCR, the two must have the same format.
Some DVCRs are even compatible with analog VCRs (see Table 1).
They rewind, play, record, and erase tapes just like in the analog
world. DVCRs have many of the same features as analog VCRs. Most
of the buttons have the same names. Their analog inputs and outputs
obey all of the video laws you've learned already. The analog
cables, connectors, and termination regimens are the same. Tapes
still need to be labeled to keep track of them. The more things
change, the more they stay the same.
Table 2 summarizes the main differences between DVCR formats. Table 3 goes into more detail for the technophiles among us.
Digital Formats Compared [See also Digital VCR Formats Compared: Tabular Specifications]
A few generalizations which apply to both analog and digital realms: The more tape you use per second, the heartier the signal. Also, as you'll gather from Table 3, the wider the track width (width of the magnetic path made by the spinning video heads), the more robust the signal. Both of these factors use up tape faster, increasing tape costs and reducing the time you get on a single cassette. But they make a better picture.
The highest quality tape formats presently use wider, faster moving tape than the lower quality formats. In the digital world, the D1, D2, and D6 formats use ¾" tape. As manufacturers endeavor to squeeze more and more into less and less space, these numbers come down. There was a time when 2" tape was the best, 1" tape second best, and ½" was just for amateurs. Presently ¼" tape can do a fine to job, but ½" digital formats still pack more wallop and ¾" more wallop still.
The only composite formats in Table 2 are D-2 and D-3; the others are component. The fact that they are composite doesn't make D-2 and D-3 inferior; the signal is sampled so finely that no discernable harm comes to the composite signal by digitizing it. These are fine machines especially when you start with a composite signal. If, however, your signal starts out component, there's no sense kicking it between the eyes encoding it into composite before recording it; here a component recorder like D1, D5, D6, D7, D9, DV, etc. is the best bet.
D1, D2, D3, and D5 are not compressed, which means their signals introduce no compression-related artifacts (picture flaws). The price for not compressing is high. High data rates of 142-270Mbps, require top-of-the-line circuitry to handle, and lots of tape to store those digits.
The other digital formats in the table use some form of compression to reduce the data flow and tape consumption. As a general rule, the more compression is used, the more artifacts become visible. JVC's Digital-S DVCRs compress only a gentle 3.3:1, and Ampex's DCT and Sony's Digital Betacam are 2:1. At these low compression rates,it is hard to detect any difference between them and no compression at all.
The next step down in quality (and cost, both to buy and to feed) are the DV formats. DV (Digital Video) describes the consumer DV format and its upscale (and compatible) brothers DVC PRO, and DVCAM. Both consumer and industrial DV machines share the same color sampling, ( 4:1:1) and the same compression ratio (5:1), the same compression method (Discrete Cosine Transform-DCT),the same cassette type, and many of the same standards. Digital8 is similar to DV as far as the signals and data are concerned, but this format is recorded on high quality 8mm videocassettes at double the normal 8mm speed. In other words, a 2 hour 8mm tape will last only 1 hour when used in a Digital8 camcorder.
Betacam SX uses an ultrahigh compression (MPEG-2) to squeeze its data rate down to 18Mbps. But this serves a purpose; the rate is low enough to pass through many phone and satellite TV services, making it perfect for instant news gathering.
Data can be stored on disk as well as tape, and there are digital disk recorders (DDRs) to do this. Advantage: If you're a news reporter, you can shoot the story, then slip the disk out of the camcorder and into the editor and begin editing the story immediately; no waiting to download the data from your tape into the editor's hard drive. Disks cost about 100 times as much as tape to store the same data, but accessing data on a disk is 100 times faster than winding a tape to find it.
One special feature on DDRs is retroloop, the ability to record about 20 minutes, then without stopping, continuing to record while erasing what was recorded 20 minutes earlier. This is great when shooting the calving of glaciers, for instance. Nothing happens for the longest time, then "whoomp," a chunk of ice the size of a building breaks away and plunges into the ocean. You always get the shot, you don't waste tape on motionless snow, and you're not caught changing tapes when the action strikes.
Nearly all DVCRs slice the picture into 13.5 million luminance samples per second. This is enough to make a sharp picture. What about color? Taking 13.5 million samples of each of the color components would make a dazzling picture, but also would overburden the machinery, using up too much bandwidth and tape. Thus a 4:4:4 sampling ratio-4 samples of Y, 4 of (R-Y) and 4 of (B-Y)-where the colors have the same sharpness as the luminance- is seldom done. The 4:4:4 sampling ratio is used primarily for high resolution RGB graphics. Incidentally, there are super high sampling systems like 8:4:4 which has twice the luminance sharpness plus excellent color. True high definition TV (HDTV) which makes its picture with more detail than NTSC, could be said to use a sampling of 22:11:11. That's a lot of data.
Back to the real world, because 4:4:4 yields excess color quality, most industrial VCRs (digital and analog) settle for 4:2:2, or 4 samples of Y, 2 of (R-Y), 2 of (B-Y). The colors still look great and are sharp enough for clean chroma keying and graphics compositing. All the professional DVCRs such as DVC PRO 50, Digital-S, and Digital Betacam use this sampling. (Exception: Professional composite recorders use 4:0:0, because they sample the composite picture-imbedded colors and all-4 times, then 4 more times, etc. Since there are no color components, to sample, the other numbers are zero).
One step down from 4:2:2 is 4:1:1 used by consumer DV equipment (including Digital8) and some professional DVCRs such as DVCAM and DVC PRO. Here Y is sampled 4 times, (R-Y) once and (B-Y) once. Thus the luminance remains as sharp as the better DVCRs, just the color is degraded. Still, the loss is hard to see. In fact, 4:1:1 is technically better than the NTSC video sent to our homes, which if it were component video, would score a dismal 3:1:.5. The inferiority of 4:1:1 appears when text, graphics, and special effects call for very sharp colors. They're not there. Thus 4:1:1 may look good, but doesn't hold up as well to chroma keying and other post production manipulation as 4:2:2.
Because the luminance sampling rates for most DVCR formats is the same (13.5MHz), the luminance (Y) resolution is the same, 720x480 pixels. With 4:2:2 sampling, the Y horizontal resolution stays the same and color (C) horizontal resolution is cut in half to 360x480. With 4:1:1 sampling, the color sharpness is halved again to 180x480 pixels. All of this is pretty straightforward, but here's a surprise: sometimes the signal must be converted to 4:2:0 used by DTV, DVD, digital satellite broadcasts, and devices outputting Y/C. A 4:1:1 signal does not gracefully convert to 4:2:0. A 4:2:2 signal does.
This happens because a 4:2:0 sampling puts all the (R-Y) color components on the odd scanning lines, and the (B-Y) components on the even scanning lines. This effectively cuts the ( R-Y) vertical sharpness in half (it's no longer on all lines), and does the same for the (B-Y) component. If the 4:2:0 converter looks for color data and it is all there (as when you start out with 4:2:2), the data gets used. Vertical resolution is halved, but c'est la vie. If, however, the 4:2:0 converter looks for color data and it's not there, which is half the time when you start with just 4:1:1, the data doesn't get used, and you end up with half the horizontal resolution. Then, because of the even/odd machinations of 4:2:0, the vertical color resolution gets halved. Half of a half is a quarter and the result is 4:1:0 sampling. Thus 4:1:1 when converted to 4:2:0 makes twice as fuzzy color as when 4:2:2 changes to 4:2:0. This is one reason to prefer the 4:2:2 formats like Digital-S, DVC PRO 50, and Digital Betacam.
Audio and the consumer DV format
If you plan to edit audio on your DVCRs, beware that DV (the consumer 25 Mbps format) does not precisely lock its audio to its images. Audio edits may have brief silences or clicks as the sound catches up to the picture. The PRO formats (DVCAM, DVC PRO, and PRO 50, etc.) lock the audio precisely to the pictures. Their audio edits are perfect and accurate.
Recording formats that spread their magnetism over a larger area are more robust. They stand up to the rigors of editing and are less sensitive to tape dropouts. It's simply a law physics: the more magnetic particles you involve the process, the stronger the magnetism can be. Manufacturers, in effort to pack the most data into the tiniest space, have reduced the detectable magnetism to just whisker above nonexistent. Serious pro users aim for formats with the wider track widths (See Table 3). Naturally, 1/2 inch tape provides more recording real estate on which to record a wider track than does 1/4 inch tape.
This doesn't mean you should avoid the 1/4 inch formats. When trekking in the Himalayas, every ounce counts, and trade-offs are appropriate. You can always dub your camera masters (digitally, without loss, and sometimes at four times the normal speed) onto a more robust format back home.
Bits per sample
Most common DV formats use eight bits per sample which results in 256 levels of image brightness. This is quite satisfactory, but it's worth noting that a few of the higher end formats use 10 bits per sample, permitting far more gradations in brightness yielding smoother pictures with a greater signal-to-noise ratio.
The HD Future
The 4:2:2 formats, DVC PRO 50, Digital-S, and Digital Betacam are poised for high definition. No, they are not HD, but when converted to HD, because the 4:2:2 image is so detailed, they survive nicely. Further, there are extensions to the formats (DVC PRO 100 and Digital-S 100) that raise the resolution ante in preparation for the big HD-Day. The "100" in their designations reflect, as you might guess, the 100 Mbps rate that makes 1080i or 720-60p hi def possible. The process is done by doubling the tape speed, which of course halves the recording length.
Recording HD requires a lot of data. Making an affordable digital HD recorder requires that the data torrent be reduced. By compressing the data to MPEG-2, and recording that onto DV type cassettes, it has been possible to get the price of an HD camcorder down to about $3500. The imagery is not TRULY HD (1920 pixels per line times 720 or 1080 lines per picture) because the lens is relatively cheap, the chips can't record that resolution, and the compression codec squashes some of the sharpness out of the picture. Still, the results are extraordinary, better than common SD from a DV recorder. MPEG-2, remember, is difficult to edit, because all the pictures are not there; some are "imaginary". Difficult does not mean impossible, and some software, teamed with sprightly dual-core computers, can reconstruct the "missing" pictures, making MPEG-2 editable. Better compression schemes (Wavelets, MPEG-4) may alleviate some of this data logjam, as well as faster computers.
Details are sketchy, but according to Dave Walton, Marketing Communications Manager at JVC Professional Products, there are Digital-S machines in broadcast facilities that are going strong with 5000 hours on their original heads. When the heads finally hit the dust (or become dust), drum replacements cost about $1000. The tape appears to survive 100 passes and a DS-104 104 minute tape costs only $42.50.
So what format do you invest in?
For high definition (HD) reproduction, film quality, or multilayer graphics compositing, where cost is not much of an object, D1, D2, D3, D5, and D6 are formats to consider.
For excellent quality pictures, amenable to chroma key and graphics layering, the choices seem to be Digital Betacam, Digital-S, and DVC-PRO 50.
For good quality ENG and industrial production at a moderate price, DVCAM, DVCPRO, and Betacam-SX seem to be the choices. Interestingly, Digital-S seems to price itself closer to DVCAM and DVC PRO but performs more like the higher priced Digital Betacam and DVC PRO 50.
For prosumers and budget minded professionals who won't be performing a lot of linear audio edits, or who expect to skydive with their camcorders, DV or Digital8 would seem most adequate.
1. The more samples you take of a picture the better the image quality will be. An industry standard is 13.5 million samples per second (described as a 13.5MHz sampling rate). The fewer samples you throw away, the better the image will remain.Color samples are sacrificed first. Thus a 4:2:2 sampling is preferable to a 4:1:1 sampling.
2. The less you compress the data, the better the quality. No compression, 2:1 compression and 3.3:1 compression look essentially identical; 5:1 is tolerably lossy.
3. The more tape you use to store the image (i.e., 20 micron track width on swiftly moving ½" tape vs. 10 micron tracks on slow moving ¼" tape) the more robust and reliable the data will be.
4. The more bits per sample you use (i.e., 10 vs. 8), the more accurate the recording.
|About the author||About Today's Video 4th. ed.||Return home|