Digital Format Wars, Redux

As manufacturers pronounce the dawn of tapeless acquisition, are we ready for another revolution?

Sony's XDCAM optical system relies upon blue-laser technology and records both a high-resolution original and a lower-resolution, frame-accurate MPEG-4 proxy.

The transition from analog to digital certainly caused confusion in the video industry. Most of the confusion came from all of us having to understand a whole new technology. The revolution itself, thankfully, was relatively straightforward. There was one so-called consumer format — DV. It used 5:1 compression on 1/4in. tape. There were also four proprietary formats: DVCAM, DVCPRO, plus DVCPRO50, and Digital-S. The first two were essentially DV with a few twists designed to make the 25Mbps format more attractive to “pros.” More specifically, the enhancements made these formats better for those editing videotape. The latter two used a pair of DV25 codecs — plus two pairs of heads — to record 50Mbps to videotape.

Once you understood the similarities and differences among these five formats, and understood the peculiarities of locked and unlocked audio — you were ready to move forward. Moreover, despite Beta SP owners who continued to claim analog offered a better picture than digital, folks moved to digital in droves. Now, these folks face another stage in the digital revolution. And this stage of progress, it seems to me, is far less clear than the last one.

My confusion over the coming changes began at NAB 2003 when Sony and Panasonic both announced their new tapeless formats. While we all know that eventually we will be recording on something other than videotape, I certainly would not place “get rid of tape” at the top of my wish list. With Mini-DV tapes selling for less than $5 and tape quality at an all-time high, what's the big problem with tape?

As I look at the world of video problems, I would emphasize the following: inability to achieve clean lowlight capture, inability to handle high-contrast situations, inadequate spatial resolution, and high cost. Why the concern about cost? Although the stock market dramatically improved last year based upon increases in corporate profits, many small production companies are still hurting. A major reduction in the price of 1/2in. and 2/3in. CCD camcorders would be the greatest enhancement these companies could find at NAB 2004. Therefore, replacing a $5 tape with $10,000 of RAM hardly seems a giant step forward. And, with that point in mind, let's first look at a world without tape to see what it offers.

Both Sony and Panasonic argue that a tapeless world supports a dramatically faster “workflow.” Consequently, the cost-benefit analysis involves the cost of all new gear, plus a much higher media cost — versus a need for speed. I see only one market that clearly has such a need — ENG. And that, of course, is the market Sony and Panasonic have targeted with their new formats.

Panasonic, with its P2 technology (shown here with the AJ-SPX800), went the solid-state route. Each 4GB P2 card can hold 16 minutes of DVCPRO footage or 8 minutes of DVCPRO50.

Listening to the pre-NAB press conferences from both companies, one hears three scenarios. In the first scenario, all of the proxy video is transmitted back to an anxiously waiting studio. There, producers will assemble the storyboard that enables them to determine exactly which shots must be transmitted back from the remote location. In the second scenario, the cameraperson or field producer assembles a storyboard from proxy video and transmits it to the studio for immediate broadcast. In the third scenario, the media is rushed to the waiting hands of an editor who pops it into her NLE and quickly edits a story just before airtime. In this case, there is no need to copy the video to the NLE.

While the first two scenarios seem valid, the latter seems less so. I live in midtown Manhattan and so I see a lot of news stories being covered. In almost every case, the cameraperson arrives in a news van. Then, while the engineer points the microwave antenna, the camera operator mounts his camcorder on a tripod and feeds the camera's cable back to the van. It seems most news stations have a fetish about “live” reporting. Hence, many breaking stories are covered from a hovering helicopter.

Of course there are other advantages to eliminating tape. The near zero-maintenance for memory-based recording must be highly attractive to networks that field dozens upon dozens of camcorders nationwide. Blue-laser media's ability to act as both storage media and archive media must also be attractive to stations.

What stations and networks may not yet grasp, however, is how unattractive their news programs are when the network's HD programming comes to an end. If you are accustomed to HD, it's unpleasant to watch the image from the expensive studio cameras and downright painful to watch the on-location stories. In my opinion, given the accelerating sales of widescreen plasmas and RPTVs that are “HD Ready,” plus the nearly universal availability of HD OTA and cable HD, news groups should be spending money primarily on HD equipment.

Certainly I don't expect that news should be covered by either $100,000 HDCAM or DVCPRO HD equipment. Thankfully, there is an alternative — HD recorded using MPEG-2 (i.e., HDV). The most inexpensive way to build a newsgathering HDV camcorder is to use either a mini- or standard-sized DV tape-transport with a camera head that has three 1/2in. or three 2/3in. CCDs. Connecting one of these to a powerful laptop via FireWire creates a low-cost field editing system. If proxy material is required, MPEG-4 proxy can be stored on low-cost PCMCIA memory cards. JVC is expected to announce a camcorder of this type at NAB 2004. Naturally, such an HD camcorder will be very popular in markets beyond ENG and will support my slogan, “Friends don't let friends buy SD.”

Before looking further into HD newsgathering, let's take a closer look at the technologies being offered by Sony and Panasonic. Sony's XDCAM optical system records both a high-resolution original (either DVCAM or IMX) and a lower-resolution, frame-accurate, MPEG-4 proxy. Newsgathering teams can transfer the proxy information to Xpri laptop editors where it can be field edited and the EDL written back to the optical disc. Alternately, the proxy can be transferred to the studio at up to 30 times faster than realtime, so producers can immediately start writing scripts.

The optical media consists of a 12cm (5in.) rewritable disc in a protective cartridge. Blue-laser technology allows this new media to achieve far higher-recording capacity and bandwidth than with a red laser. A single disc holds 90 minutes of DVCAM material — or 45 minutes of IMX MPEG-2 material recorded at 50Mbps, 55 minutes at 40Mbps, and 75 minutes at 30Mbps.

Panasonic claims that a Blue-laser mechanism is not reliable enough to be used for field use. Therefore, Panasonic made an even bolder leap into the future. It's P2 solid-state memory card uses the PCMCIA form factor and incorporates four SD memory cards plus a controller IC. With a 4GB P2 card, respective record capacity is 16 and 8 minutes with DVCPRO and DVCPRO50. Panasonic's pro camcorder will have five slots, so up to 80 minutes of video can be recorded. A card can be rewritten a minimum of 100,000 times without performance degradation and offers a 20MB/s bandwidth that is more than adequate to support 100Mbps DVCPRO HD.

An intrinsic advantage of P2 card recording is that when a P2 card is inserted into the PCMCIA card slot of a laptop computer, it will be instantly recognized as a mounted disk drive. Not only can one record from a P2 module, but one can also edit to it. Panasonic has announced that P2 cards will initially be available in 2GB and 4GB capacities, with 8GB, 16GB, and 32GB capacities to follow. A 4GB module is expected to cost about $2,100. Over time, the cost per minute of recording will decrease following Moore's law. By following the curve of solid-state rather than mechanical technology, the cost of recording could move toward the vanishing point.

Examining both technologies reveals one advance that will change the video industry forever. Both technologies affect a near total divorce between the digital compression technology employed and the digital storage technology utilized. Video and audio are now only data, data that will be moved and stored as if it were our bank records or an Internet request for pizza. As long as the media supports the data bandwidth and has a short access latency, it can hold motion images, multi-channel sound, proxy media, and metadata.

The other confusion — OK, a concern — I had at NAB 2003 was inspired by introduction of radically different proprietary media. Panasonic and Sony have, over the last year, brought to market tape decks that can read each other's 1/4in. tapes. This is exactly what customers need. Moreover, such decks could easily evolve to read — and even write — HD MPEG-2. Unfortunately, we are moving into a world where a universal “deck” would have to have a 1/4in. tape transport, a Blue-laser cartridge slot, perhaps a Blu-ray slot, five P2 PCMCIA slots, and potentially several SD card slots. As a camcorder reviewer, I find this a nightmare scenario. Small post houses will also have a problem with a multitude of media types.

This raises a question: What role does the new random-access media have for those of us not doing ENG? All of us want a more efficient editing process. And clearly, any fast random-access media enables more efficient editing. That means at some point we will all support a move to a tapeless world. The questions, it seems to me, are when and at what cost. The issue of “cost” involves whether random access media will have a much higher cost than tape. At this point, it's hard to imagine that any media can be as cheap as tape. But, upon assessing all the advantages of tapeless production, we may feel the higher media cost is more than fully offset.

The “when” question pivots on the question we each must answer: Do we need to eliminate tape in what will likely be the last NTSC equipment we purchase, or can we wait until we buy HD? To answer that question, we need to know how long we will have to wait for affordable HD equipment that has an acceptable level of quality. Such equipment would have five characteristics: full manual control, three 1/3-, 1/2-, or 2/3in. CCDs, wide-word DSP, 720p60 and/or 1080i operation, and better than 4:2:0 chroma resolution. The first three are primarily a function of cost. For example, enhanced CCDs that are 16:9 native and have 1.5 million pixels (1440×1080) will not be cheap.

The latter capabilities require a new generation of MPEG-2 encoders. To support 720p60 or 1080i requires the encoder to process twice as much data as necessary for 720p30. Increasing chroma resolution requires — to keep the same level of image quality — higher encoder bandwidth. This, in turn, requires greater recording bandwidth. Sony's Blue-laser recording mechanism can record up to 80Mbps. (Currently it records 45 minutes of MPEG-2 IMX at 50Mbps.) With an enhanced HD encoder, up to an hour could be recorded at about 35Mbps. With a double-sided optical mechanism, an hour of 80Mbps “enhanced” HDV could be recorded.

Panasonic's P2 module can support recording up to 160Mbps. This can easily support DVCPRO HD and/or HDV. With today's 4GB modules, a five-slot P2 camcorder could record 40 minutes of 50Mbps enhanced HDV. The use of enhanced HDV instead of DVCPRO HD would mean an HD camcorder could ship before 8GB modules are available.

Sony has already indicated that it will show an XDCAM-based HDV camcorder at NAB 2004. It is likely it will use enhanced HDV. Possible format name? Perhaps HDV2, HDVCAM, or IMX-HD. Panasonic might also show a P2-based camcorder that uses enhanced HDV. Doing so, of course, would require Panasonic to add an inter-frame format to its line of intra-frame formats.

What will an enhanced HDV format mean? First, it would enable some manufacturers to try — as they did with DV — to position HDV as a “consumer-only” HD format. Second, it will eliminate concerns over chroma resolution and MPEG-2 artifacts that some have with HDV. This, in turn, would extend the market for the new random-access HDV camcorders from ENG to EFP.

Sony's XDCAM and Panasonic P2 units already have full manual control and 2/3in. CCDs, so such a camcorder will easily meet our control and quality requirements. Thus, our move to HD will be primarily a question of equipment cost. At this point, we have no idea how much such camcorders will cost, but one thing is obvious. These camcorders will not be priced as “prosumer” HDV camcorders. (This raises the serious question of how and when we will see more prosumer-priced HDV camcorders.)

Nevertheless, enhanced HD camcorders will be much cheaper than today's HDCAM and DVCPRO HD camcorders. They will be professional-quality camcorders with “pro” prices. Moreover, they will show the path from NTSC SD to MPEG-2 HD for those with the necessary budget.