|
Video
Accessibility Act and ATSC Standards
The President
signed the "Twenty-First Century Communications and Video
Accessibility Act of 2010" ("ACT") on October 8,
2010. The Act, among other things, requires the FCC to reinstate
its prior Video Description Rules not later than one year after
enactment of the Act (or approximately October 8, 2011). These
FCC rules were originally adopted in 2000 but were vacated by
a court in 2002 on grounds that the FCC exceeded its authority.
Those rules required TV stations affiliated with the big 4 networks
(ABC, CBS, Fox, NBC) in the top 25 DMAs (and the top 5 cable networks
nationwide) to provide 50 hours per quarter (approximately 4 hours
per week) of video described programming. The Act also gives the
FCC the authority to expand, over time (more than a ten year period),
the number of hours and the number of markets in which video descriptions
must be provided (see the
October 11, 2010 issue of TV TechCheck).
The technical
provisions in the ATSC Standards to deliver a Video Description
audio stream can be confusing. While the current versions of ATSC
A/53 referenced in the FCC rules enable the transport and signaling
of this stream, the newer versions of A/53 are clearer.
The first
point of potential confusion is that the audio stream originally
intended to be used for Video Description is labeled by the AC3
standard (A/52) as Visually Impaired (VI). The standards describe
this as an associated service, that is, one that the receiver
was intended to use (or mix) together with the main audio. This
"assembly-of-parts" concept dating from the system selection
days simply is no longer applicable; and no known ATSC receiver
has the two AC-3 decoders that would be needed to enable it. From
a production perspective, experts assert that Video Description
audio is a separately produced complete audio track with all parts
present including the descriptive narration interleaved with the
standard dialogue and effects.
Over the past
couple of years the ATSC Standards have been updated to reflect
this production and receiver reality, and while the label "associated
service" is retained, A/53 Part 5 now requires that associated
services contain complete program mixes containing all audio program
elements (dialogue, music, effects, etc.) that are intended to
be presented to a listener. This is signaled by setting the full_svc
bit in the AC-3_audio_stream_descriptor() to a value of "1."
ATSC is in
the process of revising A/52 to clearly be consistent with this
new wording; but there is nothing preventing setting the full_svc
bit to "1" in the current A/52 in the FCC rules.
The version
of A/53 referenced in the FCC rules requires that there must be
at least one Complete Main (CM) audio (unless there is no audio
at all). This is the case for new versions of A/53 as well.
For real time
transmission of a separate VI audio in an ATSC Transport Stream,
a separate AC-3 encoder would be needed. That encoder would embed
the metadata element 'bsmod' with the value of "010"
in the VI audio. Another AC-3 encoder provides the required Complete
Main stream with bsmod="000" embedded. The Program Map
Table would contain the PID values for the packets that deliver
the three elementary streams (video, CM, VI).
The Program
Map Table (PMT) is also required to have an AC-3 Descriptor associated
with each audio stream. More precisely, an AC-3 Audio Descriptor
shall be included in the descriptor loop immediately following
the ES_info_length field in the TS_program_map_section() describing
that Elementary Stream. This descriptor contains a corresponding
bsmod field to signal the type of audio service to the receivers.
The CM-associated descriptor has bsmod = "000" and the
VI-associated descriptor has bsmod = "010", both need
to have the full_svc bit set to "1." The CM audio's
descriptor should have the descriptor field priority set to "01",
so that cable head-end equipment selects that as the main audio
if an NTSC signal is being generated. Setting the priority field
to "10" is recommended to signal the audio is intended
to be the logical replacement for the SAP channel when NTSC is
generated from the digital signal (See A/79). However, use of
that value is not required and the priority field could
be set to "other audio" = "10" or "not
specified" ="11", which might be the case where
a second language service is being sent which is intended for
the SAP channel.
The AC-3 descriptors
in the EITs (for each event) are constructed in the same manner
as the ones in the PMT, enabling digital recorders to offer the
same choices to consumers before the recording session.
Receivers
should be able to use the AC-3 descriptor from either the PMT
or the EIT to inform consumers that a VI audio is available, and
then to select it for decoding.
In the event
the only audio that is available is a track that contains Video
Description, then it would have to be labeled as a Complete Main
service in the bsmod fields. It could not have its codes set to
signal it was actually Video Description (VI). As there would
be only one audio, receivers would be expected to render it independent
of this slight disconnect in signaling.
Note that
the language of the VI track can be the same as, or different
from, the language of the Complete Main, as that may be set independently.
Based on conversations
with Dolby about the test process for AC-3 decoders, receivers
should be able to decode streams independent of the value of the
bsmod field, although selection of specific tracks using bsmod
is not part of the testing, so there is no assurance that all
receivers can locate and select this particular track.
CEA is in
the final stages of developing a recommended practice (CEB21)
to provide guidance on selection of streams based on this and
other metadata about the audio streams.
|
