Файл: Audio and Radio Engineering.pdf

Serial Digital Video/Audio Systems 8-23

Audio data packets are multiplexed into the horizontal ancillary data space of the C

B

/C

R

 par-

allel data stream, and audio control packets are multiplexed into the horizontal ancillary data
space of the Y parallel data stream.

MPEG-2 Audio Transport

SMPTE 302M specifies the transport of AES3 data in an MPEG-2 transport stream for televi-
sion applications at a sampling rate of 48 ksamples/s [7]. The MPEG audio standard itself
defines compressed audio carriage, but does not define uncompressed audio for carriage in an
MPEG-2 transport system. SMPTE 302M augments the MPEG standard to address the require-
ment to carry AES3 streams, which may consist of linear PCM audio or other data. MPEG-2
transport streams convey one or more programs of coded data, and may be constructed from one
or more elementary coded data streams, program streams, or other transport streams.

The specifications are described in terms of a model that starts with AES3 data, constructs

elementary streams (ES) from the AES3 data, then constructs packetized elementary streams
(PES) from the elementary streams, and finally constructs MPEG-2 transport streams (MTS)
from the packetized elementary streams. Although this model is used to describe the transport of
AES3 streams in MPEG-2 transport streams, the model is not mandatory. MPEG-2 transport
streams may be constructed by any method that results in a valid stream.

The SMPTE audio data elementary streams consists of audio sample words, which may be

derived from AES3 digital audio subframes, together with validity, user, and channel status (V, U,
C) bits and a framing (F) bit. There may be 2, 4, 6, or 8 channels of audio data conveyed in a sin-
gle audio elementary stream and corresponding packetized elementary stream. Multiple pack-
etized elementary streams may be used in applications requiring more channels.

8.2.3c

Data Services

SMPTE 334M (proposed at this writing) defines a method of coding that allows data services to
be carried in the vertical ancillary data space of a bit-serial component television signal con-
forming with SMPTE 292M or ANSI/SMPTE 259M [8]. This includes data broadcast services
intended for the public as well as broadcaster internal control and communications. Despite the
reference to the bit-serial interface, nothing in the specification precludes its use in a parallel
digital interface for component digital video signals. The data described in the standard can also
be transported in K-L-V format according to SMPTE 336M, or via other means.

The data packets are located in the active line portion of one or more lines in the vertical

ancillary space. Data can be located in any lines in the area from the second line after the line
specified for switching to the last line before active video, inclusively. Individual data services
are not assigned to any specific data lines; receiving equipment should identify and select ser-
vices on the basis of their ANC DID and SDID fields.

Because ANC data may be located in the lines immediately preceding active video, manufac-

turers of video compression equipment must ensure that these data bits are not included in video
compression calculations.

The chrominance (C

b

/C

r

) and luminance (Y) data are carried in two separate streams within

the 292M signal, complete with their own ANC data flags and CRCs. Defined data services are
carried in the Y stream. Other data services can be inserted into either one of these streams with-
out restrictions.

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

Any use is subject to the Terms of Use as given at the website.

Serial Digital Video/Audio Systems

8-24 Audio Networking

In the 259M/125M signal, the chrominance and luminance data are carried in a single stream.

In this case, all data services are carried in this stream with a single ANC data flag and CRC.

There is no specific provision in SMPTE 334M for ensuring that the relative timing between

the video and its embedded VANC data is correct. The only timing relationship that exists is cre-
ated when the data are embedded in the video. Once that relationship is established, the deter-
ministic nature of 292M or 259M transport ensures that the relationship is preserved.

8.2.3d

Time Division Multiplexing on SMPTE 292M

Given the wide variety of commercially accepted standards for both standard and high-definition
television systems used in studio production, post-production, and distribution facilities, a large
number of incompatible interfaces exists for the interconnection and routing of the various
4:2:2I, 4:4:4I, 4:2:2:4I, 4:4:4:4I, and 4:2:2P component and composite video signals throughout
a given plant [9]. In addition, high-definition video signals used in a number of television facili-
ties operate at a bandwidth of 1.485 Gbits/s.

A higher bandwidth interface such as that used to carry high-definition component video

could carry several lower-bandwidth component or composite video signals over a single physi-
cal link, reducing the number of cable or fiber routing signals throughout the plant and simplify-
ing the overall routing requirements. In fact, the lower-bandwidth signals could be formatted into
digital active line areas of existing high-definition interfaces in such a way that the resulting sig-
nal would appear to most existing pieces of equipment as a regular high-definition video signal.

In addition, the low-bandwidth signals could be any generic data stream, including com-

pressed standard- and high-definition video, not just component video signals. The obvious ben-
efit of this multiplexing scheme is that existing equipment for serializing and deserializing high-
definition bit-parallel data signals for distribution throughout the plant could also be used to seri-
alize and deserialize the high-bandwidth multiplexed data stream without any additional hard-
ware.

All that is needed to implement this system is hardware for multiplexing and demultiplexing

between bit-parallel high- and low-bandwidth formats.

With these real-world needs in mind, the SMPTE developed an innovative solution aimed at

bridging the video hardware of the present with that of the future. SMPTE 346M (proposed at
this writing) defines the time division multiplexing (TDM) of various standard-definition digital
video and generic 8-bit data signals over the high-definition serial digital interface specified in
SMPTE 292M. The objective of this high-definition multiplexing interface is to use a single
physical link to transmit, distribute, route, and switch a complete family of existing 10-bit video
formats and various data formats.

Active video and vertical blanking areas in the HD SDI stream are time divided into 19 inter-

leaved channels. The word in the first channel is used to indicate the data validation of the
remaining 18 channels. A single video or data stream is multiplexed into one or multiple chan-
nels of the total 18 data channels. A control packet is multiplexed into the horizontal ancillary
data space of the luminance parallel data area after the switching point of the HD stream for each
video or data stream. The control packet indicates how and which channels are used for this
video or data stream. It also contains stream clock reference information for clock recovery of
the original clock signal.

Multiple standard-definition video or data streams can be multiplexed in and demultiplexed

from a single HD SDI stream with a total delay of a fraction of a horizontal line. By dividing the

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

Any use is subject to the Terms of Use as given at the website.

Serial Digital Video/Audio Systems

Serial Digital Video/Audio Systems 8-25

payload into segments, which are subdivided into channels, time-division multiplexing can be
applied to several data sources in such a way that very low latency is generated during the demul-
tiplexing process. A fixed number of channels in each segment provides an efficient implemen-
tation of different data by reducing the complexity of the multiplexing and demultiplexing
processes.

SMPTE 346M specifies the format for multiplexing multiple asynchronous standard-defini-

tion video streams and generic data into the high-definition system interfaces as defined in
SMPTE 274M and ANSI/SMPTE 296M in the bit-parallel source format for the bit-serial inter-
face defined in SMPTE 292M. The major standard-definition system interfaces are defined by
ANSI/SMPTE 259M and ITU-R BT.601-5. They are the 4:2:2 digital component video signal
interfaces defined in ANSI/SMPTE 125M, ANSI/SMPTE 267M, and ANSI/SMPTE 293M.
These standards specify interfaces for the 270-, 360-, and 540-Mbits/s bit-parallel formats. The
standard also covers other serial video standards such as 143 Mbits/s 525-line and 177 Mbits/s
625-line composite digital signals.

SMPTE 346M can be used for a 270-Mbits/s SDI interface to carry multiple streams of data

and a single 8-bit or 10-bit composite digital video signal. In a similar manner, a 540-Mbits/s
SDI system can carry multiple streams of video and data. SMPTE 346M can also be extended to
future higher bit rate serial digital standards.

The standard definition video formats referred to in SMPTE 346M are listed in Table 8.2.4.

8.2.3e

Packet Transport

SMPTE 348M (proposed at this writing) provides the mechanisms necessary to facilitate the
transport of packetized data over a synchronous data carrier [10]. The HD-SDTI data packets
and synchronizing signals provide a data transport interface that is compatible with SMPTE
292M such that it can be readily used by the infrastructure provided by the standard. SMPTE
348M uses a dual-channel technique where each line carries two data channels, each forming an
independent HD-SDTI data transport mechanism. The two channels are word-multiplexed onto

Table 8.2.4 Summary of SD Video Formats Referenced in SMPTE 346M 

(After [9]

SD System/Sampling Structure

525

×  60 or 625 ×  50, 4 ×  3

13.5 MHz

525

×  59.94 or 625 ×  50, 16 ×  9 18 

MHz

4:0:0I

135 Mbits/s

180 Mbits/s

4:2:2I

270 Mbits/s ANSI/SMPTE 267M

360 Mbits/s ANSI/SMPTE 267M

4:2:2:4I

 360 Mbits/s

540 Mbits/s

4:4:4I

360 Mbits/s ITU-R BT.601-5

540 Mb/s ITU-R BT.601-5

4:2:0P

360 Mbits/s

540 Mbits/s

4:4:4:4I

540 Mbits/s SMPTE RP 174

720 Mbits/s

8:4:4I

540 Mbits/s

720 Mbits/s

4:2:2P

540 Mbits/s ANSI/SMPTE 293M

720 Mbits/s

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

Any use is subject to the Terms of Use as given at the website.

Serial Digital Video/Audio Systems

8-26 Audio Networking

the HD-SDI stream such that one line-channel occupies the C data space and the other line-chan-
nel occupies the Y data space.

The standard provides for a baseline operation that supports a constant payload length per

line-channel having a maximum payload data rate up to approximately 1.0 Gbits/s. It further pro-
vides for an extended operation mode that supports a variable payload length through the
advancement of the SAV sequence to ensure a constant payload data rate regardless of the HD-
SDI frame rate. The HD-SDTI protocol is compatible with SMPTE 305M.

SMPTE 348M describes the assembly of two channels of 10-bit words multiplexed into one

HD-SDI line for the purpose of transporting the data streams in a structured framework. The
HD-SDTI data blocks and synchronizing signals provide a data transport protocol that can
readily be added to the infrastructure provided by SMPTE 292M.

SMPTE 292M requires a sequence of 10-bit words that define a television horizontal line

comprising five areas in the following sequence (the first two areas are often described together):

•

EAV: a four-word unique timing sequence defining the end of active video (of the previous
line)

•

LN/CRC: two words defining the line number followed by a two-word CRC error detection
code

•

Digital line blanking

•

SAV: a four-word unique timing sequence defining the start of active video

•

Digital active line

An associated television source format standard defines the rate of television horizontal lines

by specifying the following parameters:

•

The number of words per line

•

The number of words in the digital active line (and hence the number of words in the digital
line blanking period)

•

The number of lines per frame

•

The number of frames per second

SMPTE 292M currently defines four source format standards (1152, 1035, 1080, and 720

active lines per frame). SMPTE 125M describes the meaning of the EAV and SAV word
sequences that can be applied to all relevant source formats. 

A decoder operating under the SMPTE 348M standard is not be required to decode all the

source formats available to SMPTE 292M. The source formats that must be supported by the
decoder are specified in the application document.

8.2.3f

540 Mbits/s Interface

SMPTE 344M (proposed at this writing) specifies a serial digital interface that operates at a
nominal rate of 540 Mbits/s [11]. The standard is intended for applications in television studios
over specified lengths of coaxial cable. Separate SMPTE documents specify the mapping of
source image formats onto the 540 Mbits/s serial interface.

The connector has mechanical characteristics conforming to the 50-

Ω BNC type. Mechanical

dimensions of the connector may produce either a nominal 50- or 75-

Ω impedance and are usable

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

Any use is subject to the Terms of Use as given at the website.

Serial Digital Video/Audio Systems

Serial Digital Video/Audio Systems 8-27

at frequencies up to 540 MHz, assuming a return loss that is greater than 15 dB. However, the
electrical characteristics of the connector and its associated interface circuitry must provide a
resistive impedance of 75 

Ω. Where a 75-Ω connector is used, its mechanical characteristics must

reliably interface with the nominal 50-

Ω BNC type defined by IEC 60169-8.

The application of SMPTE 344M does not require a particular type of coax. It is necessary,

however, for the coax to be a 75-

Ω type and for the frequency response of the coax, in dB, to be

approximately proportional to 1/

 from 1 MHz to 540 MHz to ensure correct operation of auto-

matic cable equalizers over moderate to maximum lengths.

The channel coding specified in SMPTE 344M is scrambled NRZI. The LSB of any data

word is transmitted first. To maintain synchronization and word alignment at the serial receiver,
EAV and SAV timing references are inserted into the data stream.

8.2.4

Serial Data Transport Interface

The serial data transport interface (SDTI) is a standard for transporting packetized audio, video,
and data between cameras, VTRs, editing/compositing systems, video servers, and transmitters
in professional and broadcast video environments [12]. SDTI builds on the familiar SDI standard
that is now widely used in studios and production centers to transfer uncompressed digital video
between video devices. SDTI provides for faster-than-real-time video transfers and a reduction in
the number of decompression/compression generations required during the video production
process, while utilizing the existing SDI infrastructure.

The SMPTE 305M SDTI specification evolved from a collaborative effort on the part of

equipment vendors and interested parties, under the auspices of the SMPTE PT20.04 Workgroup
on Packetized Television Interconnections, to define a common interchange interface for com-
pressed audio and video.

Because SDTI is built upon the SMPTE 259M SDI specification, it shares the same mechan-

ical, electrical, and transport mechanisms. BNC connectors and coaxial cables establish the
mechanical link.

SDI transports uncompressed digital video using 10-bit words in the 4:2:2 Y, U, V component

mode for 525- and 625-line applications. Words are serialized, scrambled, and coded into a 270-
Mbits/s or 360-Mbits/s serial stream. In order to synchronize video timing between the transmit-
ter and the receiver, SDI defines words in the bitstream called end of active video (EAV) and
start of active video (SAV), as illustrated in Figure 8.2.9. At 270 Mbits/s, the active portion of
each video line is 1440 words and at 360 Mbits/s, the active portion is 1920 words. The area
between EAV and SAV can be used to transmit ancillary data such as digital audio and time code.
The ancillary data space is defined by SMPTE 291M-1998.

SDI and SDTI can co-exist in a facility using the same cabling, distribution amplifiers, and

routers. Cable lengths of more than 300 meters are supported. SDI repeaters can be used to reach
longer distances. A versatile studio configuration that supports all the required point-to-point
connections can be established using an SDI router.

8.2.4a

SDTI Data Structure

SDTI uses the ancillary data space in SDI to identify that a specific video line carries SDTI
information [12]. The packetized video is transported within the active video area, providing 200

f

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

Any use is subject to the Terms of Use as given at the website.

Serial Digital Video/Audio Systems