Digital transmission brings with it a host of new possibilities — and headaches for those unprepared to make the most of the opportunities. The flexibility of DTV means that the delivery medium can be tailored to suit the particular business model of each broadcaster, and the basis for this lies in multichannel transmission. This dynamically reconfigurable platform offers multiple programs within a single transmission, each with multichannel sound. Planned correctly, the service can vary on a daily basis, so it's important to know some of the tools available, both for systems, video and audio.
Major and minor channels
In 2004, the FCC incorporated by reference ATSC PSIP Standard A/65B into its rules. By this action, the following PSIP elements were made mandatory: the master guide table (MGT), the system time table (STT), the virtual channel table (VCT) and the first four event information tables (EITs). While information in the first two is usually of limited interest to the broadcaster, the next two are critical. The FCC regulations say that the VCT must contain the NTSC channel number for each major channel entry, as well as other stream and service information, and that correct program titles are expected. This forms the basis for TV receiver (and user) channel navigation.
In order to reduce viewer confusion, the VCT defines a lookup table that can point virtually to the old NTSC channel for which the station has already established an identity. It also provides a mechanism for automatic redirection, for those stations that transition their current DTV channel to a different DTV RF channel. The VCT also enables the concept of major and minor channels, i.e., the method by which multichannel transmission is realized.
The EIT contains the program schedule information that is needed to form a program guide and can cover events up to 16 days in the future. As the timing of some events may change even at air time (such as for live sports events), it is conceivable (and useful) that the schedule may change on a moment's notice. While A/65 does not require an update rate of the EIT, broadcasters asked the FCC for clarification on the transmission requirement.
The FCC Order on Clarification adopted May 29, 2008, responded, “While we encourage stations to update the EIT as rapidly as possible when overages or other circumstances result in changes to scheduled programs, our rules and policies do not require that updates be accomplished in real-time.” Of course, it is in the broadcaster's best interest to update this information in a timely manner, but practical operations may not always allow this.
PSIP makes the DTV world go 'round
The PSIP contains the following structure for each program, which is identified by a packet ID (PID):
VCT, with channel number (internal designation), service name, NTSC channel number, ATSC channel number and virtual channel number;
elementary stream audio descriptors; and
PCR PID, which addresses a unique program clock reference for each PID, and allows each program to have its own time base.
ATSC PSIP allows the use of 100 minor channel numbers for DTV video (or DTV audio-only) service. This means that as many as 100 DTV subchannels can be encapsulated within one ATSC transport stream (i.e. one transmission channel). Of course, efficient operation on a fixed-bandwidth transmission will usually mean a total number of channels far below this limit. At the same time, subchannels may come and go during the broadcast day, so it is useful to have an extensive range of channel numbers from which to identify the different program services.
Dynamic channel management becomes especially relevant if the broadcaster elects to use directed channel change (DCC). With DCC, the ATSC standard provides the capability for a program to push the receiver to a new subchannel.
Directed channel change
One interesting use of this could be for targeted advertising, where multiple versions of an advertisement are sent on different subchannels. (See Figure 1.) With a DCC-capable DTV, the receiver can be instructed to change to a different virtual channel if the user has enabled the feature and has provided individualization information during the setup process of the receiver. In fact, the receiver can be directed to change to an entirely different RF channel.
In order to define the behavior of the receiver, PSIP includes a DCC table. With this, the requested channel change can be unconditional, or can be based on geographic, demographic or other content selection criteria. In addition, the context of the channel change can tell the receiver whether it should remain on the new channel or return to the original channel after a set period of time or upon a return direction.
In order to remain transparent and avoid viewer confusion, such channel changes may be hidden from the viewer, and the temporary virtual channels may be hidden from the user's program guide display. Of course, switching bit streams is not a trivial task, especially if a seamless switch is desired. But ATSC (i.e. MPEG) includes splice point fields in order to accomplish switching without artifacts.
Multichannel sound's implementation challenges
While the subject of multichannel sound has been covered here extensively, some key elements bear repeating as we approach the February 2009 transition date. One fundamental issue is the way that audio will be reproduced in the viewer's home.
Of course, a 5.1-channel transmission will not always be listened to on a full 5.1 system. But the choice of reproduction depends on a combination of factors: the mix and setup at the encoding end, the viewer's equipment and the viewer's choice (or unintended default) of how to operate their equipment. This means that downmixing — the reduction of 5.1 transmitted (or original) channels to a fewer number of listened ones — can occur with unintended consequences, including missing material or the dreaded stereo “hole in the middle” where dialog (and other material) is either missing or spatially misplaced.
It is especially important to understand and monitor all downmix possibilities: mono, stereo (left-only/right-only, or Lo/Ro), downmix to left-total/right-total (Lt/Rt) followed by Pro Logic decode (as occurs on some home equipment), and of course a full 5.1 presentation. (See Figure 2.) The LFE channel is usually discarded during downmixing — wherever it occurs — so one must be careful about what is placed in this channel, as stereo and mono listeners will not hear it.
In the broadcast plant, both HD-SDI and SDI interfaces provide for 16 channels of embedded audio. This means that the entire multichannel audio package can be kept tightly integrated with the video — an important factor when trying to minimize audio/video lip-sync errors.
Also, keep in mind that two types of metadata (containing dialog level and dynamic range control information) can exist in the plant — one being a subset of the other. Consumer metadata contains information for one audio program of up to 5.1 channels of audio and is found only inside a Dolby Digital (AC-3) stream. The metadata connectors on Dolby equipment produce and accept only professional metadata, which contains up to eight consumer metadata streams.
Additional information about metadata and Dolby Digital (AC-3) can be found in ATSC A/54, available at www.atsc.org. Also, the ATSC Implementation Subcommittee IS-SEWG Audio Group has generated many documents covering the subject of multichannel audio. Joining the group would be a great way to get an in-depth understanding of the subject, and keep up with new developments and solutions.
Aldo Cugnini is a consultant in the digital television industry.
Send questions and comments to: firstname.lastname@example.org.
Editor's note: Regarding the September Transition to Digital: The MGT and the STT are essential. If the first is not correctly constructed, a receiver cannot locate the other tables. The STT contents are critical because sending the wrong time will impact any DVR tuning based on event start times and may impact receivers’ ability to actually use the EITs. Also, because there is a time accuracy requirement, allowing the clock to drift in the PSIP generator equipment can result in a violation of FCC rules.
The PSIP contains the listed structure for the virtual channel, not the program. Second, the video stream descriptor (which is carried in the PMT for each program) does not contain bit rate or aspect ratio information.
While the major channels in the TVCT (in the United States) can be 2-99, the total number of subchannels that can be signaled is more than 100,000. Early experiments to assign subchannels dynamically caused consumer confusion and few broadcasters to alter the lineup during the day.
The IS disbanded some years ago. Its public findings can be found in the IS Findings subsection under Standards on the ATSC Web page.
Art Allison, NAB
Aldo Cugnini responds: Art, thanks for making those important points. As a key contributor to the ATSC standards, your comments should always garner deference. Standards are, by necessity, written in very terse and (hopefully) exact language, and efforts to generate short abstracts are challenging to the rest of us. Bit rate and aspect ratio are, of course, carried in the video sequence header, and my final point was intended to generate involvement in the general activities of ATSC.