MPEG and Nonlinear editing..........
Can you say "new markets"? Firewire is mentioned.
Beyond Broadcast:
Extending MPEG into
New Markets
The first products based on the architecture described in this paper will be
designed for the existing broadcast and DVD authoring MPEG-2 markets.
However, as users of analog video and other forms of digitally compressed
video begin to realize the inherent benefits of using MPEG-2, the market will
expand rapidly to include many new applications such as nonlinear editing and
personal authoring.
MPEG and Nonlinear Editing
MPEG-2 has been the compression format of choice for transmission and
storage of digital video, but it has not yet penetrated the "upstream"
applications of editing and acquisition.
The JPEG still image standard is the most popular choice for many video
editing applications. Video is captured in analog format and Motion-JPEG
(M-JPEG) is used to compress individual frames in real time without
prediction at video rates of 30 frames per second. The compressed video can
then be manipulated as desired using sophisticated editing tools, with
step-by-step instructions describing each editing step stored in edit control
lists on a hard disk. The final step involves performing the edits on the original
source material using the edit control lists and very sophisticated video effects
systems.
In this market, the primary factor dictating the use of M-JPEG is the
availability of inexpensive, single-chip codecs. Real-time video editing systems
use two or more simultaneous streams of source video to produce transition
effects between scenes. Multiple codec chips decompress the multiple streams
of data coming from the hard disks. Though MPEG-2 provides significantly
better compression than JPEG for the same video quality, the cost of the
multichip MPEG-2 encoders and their relatively large form factor have been
prohibitive for this market.
As technology has progressed, new digital acquisition technologies are driving
new standards such as DV, also known as DVC, a format that is now being
adopted rapidly by the editing community. Using the DV format in editing
eliminates the need for analog capture and JPEG compression devices. Two
current formats for DV include DV consumer, which uses 4:1:1 chroma
sampling and supports a 25 Mbps bit rate, and DV prosumer, which employs
4:2:2 chroma sampling and a 50 Mbps bit rate. Like M-JPEG, DV is
significantly inferior to MPEG-2 in compression efficiency, but there are
single-chip codec solutions that are suitable for cost-effective video editing
stations.
Because signals used in the nonlinear editing process undergo compression,
decompression, and recompression several times before they are finally
transmitted, high-quality 4:2:2 chroma sampling is preferable for high-end
video editing applications. However, until now the highest quality MPEG-2
encoders did not typically offer 4:2:2 encoding as an option. Now that a
single-chip 4:2:2 MPEG-2 codec is available for high-end applications and
single-chip 4:2:0 encoding and decoding is available for larger volume editing
applications, video editors at all levels can enjoy all the inherent benefits of
MPEG-2. The higher compression efficiency of MPEG-2 reduces the disk
space needed for video editing, saving system cost. Compression efficiency
also allows the use of inexpensive disks with lower throughput capabilities than
are required when using DV or M-JPEG. Furthermore, MPEG-2 is already
the format of choice for video storage and distribution. This compatibility
between editing and storage/transmission allows users of editing equipment to
access tools for such tasks as archiving and video manipulation that have
already been developed for the broadcast industry. It also permits the editing
solution to serve as the final encoding solution for storage or transmission to
another site.
To ease the transition of the editing market from M-JPEG and DV encoding
to MPEG-2 encoding, tools will be required that facilitate straightforward
transference of JPEG or DV encoded material into MPEG-2. This process is
called transcoding. Transcoding features can be easily created on systems
using both the MPEG-2 codec architecture and Firewire technology.
Firewire, also known as the IEEE 1394 digital connection system, is a
bi-directional link for compressed digital video, digital audio, digital time code,
and other data, plus digital control messages. Using a high-bandwidth 200
Mbps link, Firewire transfers JPEG or DV from tape or disk into the encoder
system. Software on the main system CPU board, usually a PC motherboard,
can decode the compressed video, with the resulting video then compressed
in real-time using the MPEG-2 codec.
The Personal Encoding Market
Implementation of the new codec architecture in personal computers
represents the first high-volume application for MPEG-2 encoders, and the
first step toward one-to-one encoding.
Today, all high-end PCs are designed to include DVD-ROM drives and
MPEG-2 decoders. This technology is expected to move rapidly down into
medium- and low-end PCs within the next few years. Moreover, the next
generation of DVD storage devices will include the ability to record. With the
availability of a cost-effective solution, MPEG-2 encoding will be a standard
requirement for PCs along with MPEG-2 decoding. Because of its technical
and cost advantages, a codec is the preferred architecture for this solution.
Employing a PCI standard bus in the codec architecture allows easy
integration of the encoder onto a PC motherboard. The 33 MHz bus interface
may act as either bus master or slave. Though the architecture does not
support audio encoding directly, it can take the audio stream as an input,
timestamp it, and send it to the host CPU for encoding. Multiplexing, which is
also done on the host CPU, can now be done easily without running into
difficult A/V synchronization issues.
The way in which encoding tasks are partitioned by this architecture is
especially conducive to the capabilities of the PC. Today, even the most
sophisticated PCs struggle to keep up with MPEG-2 decoding and suffer
quality degradation as a result-it is simply not feasible to do real-time,
high-quality MPEG-2 encoding using the host CPU. However, audio
encoding and multiplexing are relatively simple tasks that use only a portion of
the CPU capacity, thus freeing up cycles for the user to perform other
operations along with encoding.
As MPEG-2 encoders find a home in PCs, the benefits of digital video will
become evident to a spectrum of PC users. With a Firewire hookup and a
camcorder, every PC can now become a powerful video capture and editing
station. Recording edited video to recordable DVD eventually allows
low-cost video distribution in a format that anyone with a DVD player can
view, with dramatically better quality than with VHS tape. Because the format
is digital, data in the form of World Wide Web pages, updated price lists, and
other applications can be multiplexed with the video. Catalogs, magazines, and
high-impact promotional sales material may now be distributed in a
cost-effective manner.
Video conferencing is also a potential high-volume PC application for this
architecture. The codec design supports simultaneous encoding and decoding
of Simple Profile MPEG-2, while the flexible architecture supports a "low
delay" encoding algorithm which minimizes encoding latency. This is an
important consideration for communications.
The new codec architecture will finally position video as a tool to be used in
everyday life-as a sales tool, communication device, or to archive home
videos and send then to relatives. |
