By Jim Taylor
The Crux Of The Story
High-definition DVD. The holy grail
for videophiles. At least until the next holy grail appears on the horizon.
We’ve been expecting HD-DVD for a long time. Even before DVD was released
in its current form, in 1997, companies were showing technology demonstrations
of improvements on DVD. In the last few months, the buzz around HD-DVD has gotten
much louder. Unfortunately it’s hard to distinguish real information from
rumors, hearsay, and sheer speculation. This article presents the crux of the
story and what it means to Widescreen Review readers.
First, what is HD-DVD? The “HD”
part can stand for high-density or for high-definition. High-density means the
physical format of the discs will be improved. The discs will be the same size
and look the same, but the microscopic pits on them will be smaller and closer
together to hold more data. High-definition means the video format will be improved
to carry at least 720 scan lines, probably 1080 scan lines, preferably in progressive-scan
format (1080p) rather than interlaced format (1080i). There are different ways
to carry the high-def signal. One way is to crank things up on the venerable
but aging MPEG-2 format. The other is to switch to a new compression-decompression
format (codec). See the “Codecs Demystified” sidebar. The audio
format will probably not change much if at all, although DTS has proposed that
its audio codec become a standard feature of the new format (rather than the
optional feature it is now), pointing out that DTS® Digital Surround™
can potentially be extended to 10.2 channels. Of course the same can be done
with Dolby® Digital.
There are currently three main contenders
for the HD throne, and each has its own combination of formats and codecs. Table
1 summarizes the differences. All three variations of HD-DVD will require new
players to read the discs. (Although there is a proposal for a backwards-compatible
disc—more details later in this article.)
The contenders are:
1) Blu-ray, a significantly changed
version of DVD that uses smaller pits and tracks to fit about 27 billion bytes
(about 5.5 times standard DVD capacity) on a single layer. A read-only (ROM)
version of a Blu-ray disc will use dual layers per side to fit about 50 billion
bytes on a single side. Blu-ray achieves this higher density by using a 0.1-mm
cover layer to move the data closer to the lens (standard DVD uses a 0.6-mm
cover). A blue laser at 405-nm wavelength is required to read the smaller pits.
Video will be encoded as MPEG-2 transport streams.
2) Next-generation HD-DVD. A minimal
extension of the current DVD format with smaller pits and tracks, but no change
to the data layer depth, although a blue laser is needed to read the disc. Video
can be encoded with MPEG-2 or a new codec still to be chosen.
3) HD-DVD-9. An adaptation of the
existing dual-layer DVD-9 format, requiring no changes to the physical disc,
only a new codec for the video.
Which of these contenders manage
to succeed depends on a variety of market pressures, technical development,
and political maneuverings among the various consumer electronics manufacturers.
When Is It Prime Time?
According to the Consumer Electronics
Association (CEA), just over 2 million DTV sets will have sold in the U.S. in
2002, bringing the installed base to 3.5 million. Since more than 80 percent
of these digital TVs have no tuner for off-the-air HD reception or set-top box
for HD cable or satellite reception, most of the digital TVs currently in homes
are being used to watch DVDs. While DVDs on DTVs look surprisingly better than
on regular analog TVs, the DTVs are not being used to their potential. If you’re
one of those 3.5 million DTV owners you probably can’t wait for HD-DVD.
But is it too early? Is HD-DVD being rushed to market so early that it won’t
be as good as it should be? The answer to that question is not black and white,
especially since a major driving force behind the introduction of HD-DVD is
the installed base of 4 million DTVs in Japan (a household penetration of about
9 percent, compared to 3 percent in the U.S.) plus a healthy digital transmission
system. DTV owners in Japan would love to have a version of DVD they can use
to record HD shows, even if they’re not interested in prerecorded movies.
Meanwhile, standard-definition (SD)
DVD is just hitting its stride. Player sales and disc shipments in the U.S.
and worldwide are still more than doubling every year. Hollywood studios are
making a boatload of money from DVD. The studios are interested in HD-DVD but
they’re loath to jeopardize the current success of DVD by releasing a
new format too soon.
Thus the catch 22 of HD-DVD: there’s
a growing market for recordable HD-DVD, but it’s dangerously early to
release a “ROM” version of HD-DVD for prerecorded movies. In other
words, more and more people are interested in HD-DVD recorders, but they assume
the recorders will also play HD movies. How can DVD hardware manufacturers come
out with something to feed the demand for HD recording, yet at the same time
be able to play movies in an HD format that hasn’t been defined and that
Hollywood generally doesn’t want to support for a few years yet?
D-VHS® and the D-Theater™
format (covered extensively in Widescreen Review) is something of a release
valve for the pressure. D-VHS allows the studios to start releasing limited
HD titles to the early adopter market without endangering the SD-DVD goose that’s
still busy laying golden eggs.
To Transition Or Not To Transition?
DVD player manufacturers in Japan
would love to have a new format sooner than later. They’ve already lost
most of their market worldwide to players made in Taiwan, Korea, and China that
are so cheap it’s a wonder anyone can make any money from selling them.
So they’re pushing ahead to come out with a new generation that gives
them a few years of high margins. Also, many of the engineers and planners in
these companies are just as much fans of high-quality video as the regular readers
of these pages are—they honestly want to come out with a new format that
moves us all into the world of high-definition video.
But again, the general market is
not ready, and neither are the content suppliers. For this reason there’s
lots of talk of a “transition” format. For example, an HD format
based on the existing DVD physical format would be doable today. The players
hardly need change at all: they could use the same red lasers, the same lenses,
the same drive units, only with the requirement to read data twice as fast.
The main difference is that the players would need to have new decoder chips
and video circuitry to handle high-definition video. The production and replication
facilities could stay essentially the same, only new encoders and authoring
software would be needed for encoding and formatting the HD content.
This is part of the reasoning behind
the HD-DVD-9 format: if you start with the goal of fitting an HD movie onto
a dual-layer DVD-9, what do you need? About 95 percent of all Hollywood movies
are 135 minutes or shorter. Since DVD-9 discs have a capacity of 8.5 billion
bytes, a 135-minute movie can fit on the disc at a data rate of just over 8
megabits per second (Mbps). About 60 percent of movies are 105 minutes long,
which translates to a data rate of 10 Mbps. And most TV shows, which have become
very successful on DVD, are only 60 minutes long, which gives you 18 Mbps to
play with for encoding HD versions of TV shows. Compare these numbers—8,
10, and 18—to current SD-DVD data rates of 4 to 8 Mbps. At first blush
it seems that there’s simply not enough of a jump in data rates to be
able to support a jump in video quality. But the recent advances in video encoding
might be the solution. New coding formats, using the same basic discrete cosine
transform (DCT) approach as MPEG-2 but making many minor tweaks to improve efficiency,
can achieve the same quality as MPEG-2 at only half the data rate. Or, put another
way, the new encoders can get more than twice the quality at the same data rate.
Obviously there are concerns about
video quality. There’s a strong “HD-DVD: One Format Only!”
campaign on the Internet, backed by people who are understandably concerned
that squeezing HD video into the existing DVD format could compromise quality.
Compared to the 19.4 Mbps data rate of DTV, or the 28.2 Mbps data rate of D-VHS,
8 to 10 Mbps seems inadequate. But consider that DTV’s 19.4 Mbps includes
extra information for transport and error correction that isn’t needed
by DVD, and that HD MPEG-2 encoding on D-VHS has lots of room for improvement.
If you accept the assertion that a new codec can get at least a 100 percent
improvement, then taking the DTV essential data rate of around 18 Mbps and cutting
it in half gives you 9 Mbps. Right in the ballpark. And that doesn’t take
into account that DTV uses constant bit rate (CBR) encoding rather than DVD’s
more efficient variable bit rate (VBR) encoding. Similarly cutting the D-VHS
data rate in half gives you 14 Mbps. Not too far off, especially since the D-VHS
data rate is probably higher than it needs to be. That is, encoding the same
material at 22 or 24 Mbps instead of 28 Mbps makes no discernable difference.
If claims that new codecs can be two to three times better than MPEG-2 are true,
then a data rate of 9 to 14 Mbps should give the same quality as D-VHS.
In a sense, it’s like the days before DVD came out. LaserDisc was the
reference standard, and LaserDisc owners were highly skeptical that a tiny little
disc with compressed digital video could ever look as good as a LaserDisc. There
was plenty of fear and complaining that video quality would be compromised.
Looking back, we can plainly see that DVD was a big improvement over LaserDisc.
Likewise, today there is plenty of fear and complaining that video quality will
be compromised with a HD-DVD data rate that’s too low. But how many of
the people doing the complaining have seen the new encoding formats? Reports
from Hollywood testing in the second half of 2002 indicate that the results
are most impressive and getting better every day. Considering how much better
MPEG-2 encoding has gotten since DVD was released, it’s reasonable to
expect that new encoding formats will also continue to improve. But at this
point it’s all still speculation. A healthy skepticism is good, to make
sure video quality is not compromised. What’s not good is a head-in-the-sand
attitude that lower data rates can’t possibly do the job.
Hurry Up And Wait
All that said, does it make sense
to introduce a transition format? Let’s assume that the alternative is
Blu-ray. Blu-ray could be considered the “true” HD format: it has
high capacity and high data rates (up to 36 Mbps), so you can easily get the
same HD quality as D-VHS using MPEG-2. How long until we can get movies on Blu-ray?
It will take at least one year for Blu-ray to make it out of the laboratory
and into homes, probably two years. The first version of Blu-ray might be for
home recording only, in which case it might take another two or three years
until there’s a prerecorded version for movies. So it could take until
2005 to 2007, perhaps longer. Many of us wish we had HD-DVD today. Do we want
to wait that long?
Interestingly, Time Warner, the major
proponent of “red-laser HD,” thinks we should wait until blue laser
is ready. Warner’s viewpoint has been widely misrepresented in the press
and public opinion. Warner is pushing for red-laser HD-DVD-9, but only as part
of a single HD-DVD format. Just as today there are DVD-5s and DVD-9s, and most
customers don’t know or care about the difference, there could be red-laser
HD and blue-laser HD. Red HD discs would have lower capacity but would be cheaper,
just like DVD-5 discs, while blue HD discs would have higher capacity but would
cost more to replicate, like DVD-9s. As time goes by the cost difference will
inevitably decrease, but for the first few years of the format there could be
a difference in replication cost of $2 or so.
Then there’s another twist
being proposed to the DVD Forum: layered encoding for backward compatibility.
This has been covered before in Widescreen Review (see Issues 43, November 2000;
Issue 35, November/December 1999; and Issue 34, September/October 1999) particularly
in interviews with Joe Kane and me. The basic idea is that instead of jumping
to a completely new format that won’t play on existing players, the SD
version of video can be encoded along with the “difference signal”
created by “subtracting” the HD version from the SD version. The
result is a video stream that could play in standard players, while the HD enhancement
layer could be read and decoded by HD players and added to the SD base layer
to create an HD signal. The nice thing about this approach is that discs could
be released today that would play in SD on the existing 250 million DVD playback
devices worldwide, but could also provide an HD signal for future players. Some
potential disadvantages is that this approach is probably less efficient than
a pure HD approach, and most attempts to make it work have resulted in degradation
of the SD signal. In other words, in order to get really nice HD quality tomorrow,
you might have to live with lower SD quality today. How many people are willing
to make that sacrifice?
Where Do We Go From Here?
So what’s next for HD-DVD?
It’s hard to say. Here’s the current state of things:
The Blu-ray group is busy working
independently from the DVD Forum to get their format done. It’s possible
that Blu-ray recorders could be out in time for Christmas of 2003, but it’s
more likely we won’t see them until 2004, at least in the U.S. Originally
the Blu-ray group was ignoring prerecorded movies and concentrating entirely
on introducing Blu-ray as a home recording format, but recently they realized
that content is what makes or breaks a format. (Without Hollywood movies, DVD
would be nowhere.) Work is now underway in the Blu-ray group to define a ROM
version and a specification for video navigation to provide menus and interactivity
ala DVD-Video. Even so, since Blu-ray requires significant changes across the
entire disc replication industry, it will be some time before movies could be
distributed on prerecorded Blu-ray discs.
In November 2002, the DVD Forum steering
committee approved work on the next-generation blue-laser DVD format proposed
by Toshiba and NEC (see “From The Editor’s Couch”). The goal
is to have a specification finished in 2003, which means recorders and players
could appear in 2004. In theory, this format doesn’t require changes to
the existing DVD replication process, so prerecorded movies could be released
in 2004 if the studios are willing.
The DVD Forum is also going ahead
with work on a red-laser HD format. New encoding formats are being evaluated
and work is being done to figure out the minimal changes needed to the DVD-Video
specification to accommodate HD video, such as higher resolution subpictures
and menus. A red-laser HD format could be introduced quickly, but it will probably
be held until the blue-laser HD format can be released at the same time.
As always, copy protection is lurking
in the background. Copy protection issues delayed the release of DVD by about
six months in 1996 and 1997. It could once again cause significant delays in
the release of HD-DVD, especially since studios are even more sensitive about
releasing “studio master quality” versions of their titles. The
copy protection scheme will probably be a variation of CPPM (copy protection
for prerecorded media), the improved version of CSS that’s used for DVD-Audio
discs. Again, any attempt to release HD-DVD recorders before things such as
copy protection for HD-DVD video are finalized means the recorders won’t
be able to play HD movies. We’ve already seen the results of this conundrum
in a million or so DVI displays that can’t show video protected with the
HDMI (HDCP) method.
The upshot of all this is that there
probably won’t be a “transition” format for HD-DVD. There
really isn’t an issue of red-laser HD-DVD “competing” with
blue-laser HD-DVD, since it looks like they will be released hand-in-hand as
variations of a single HD-DVD format approved by the DVD Forum. But there is
definitely a major train wreck shaping up between the Forum’s blue-laser
HD-DVD and Blu-ray, which is especially ironic since Blu-ray companies are all
members of the DVD Forum.
A few of the scenarios that might
play out:
-Realistic case: two formats (red + blue versus Blu-ray)
-Worst case: three incompatible formats (red versus blue versus Blu-ray)
-Really worst case: splits within the Blu-ray group resulting in even more competing
formats
-Utopian case: a single format (blue or red + blue), complete with flying pigs
The one thing we can be fairly certain
of is that backward compatibility will be preserved. Just as today’s DVD
players can play audio CDs, tomorrow’s HD-DVD players will play today’s
SD-DVDs (and audio CDs to boot). The players may have to get physically bigger
just to make room for all the little logos on the faceplate. Customers will
expect future DVD players to be able to play the following: CD (CD audio), CD-R
(CD audio, MP3), CD-RW (CD audio, MP3), DVD (DVD-Video), DVD-R (DVD-Video),
DVD-RW (DVD-Video), DVD+R (DVD-Video), DVD+RW (DVD-Video), HD-DVD-9 (red laser
HD), and blue laser HD.
The player manufacturers have their
work cut out for them. The next few years of DVD development should prove most
interesting. In succeeding articles, I will discuss more details of individual
HD formats as well as the latest developments both on the technical side and
the “political” side.