The problem of audio loudness—dramatic shifts in volume when a program goes to a promo or commercial, or a viewer changes channels—has annoyed digital TV viewers for years, forcing them to frequently scramble for their remote to adjust the volume. Recently the issue has drawn the attention of Congress, with Rep. Anna Eshoo (D-Calif.) backing a bill, the Commercial Advertisement Loudness Mitigation (CALM) Act, that would require the broadcast and cable industries to regularize the volume of advertisements and the programming surrounding them.
But broadcasters say that much of the CALM Act, which has been approved by the House Communications Subcommittee and is now seeking support in the Senate, would simply reinforce steps that broadcasters are already voluntarily undertaking to solve the loudness problem. Those measures have been formalized in a new recommended-practice document—called “ATSC Recommended Practice: Techniques for Establishing and Maintaining Audio Loudness for Digital Television”—from the Advanced Television Systems Committee, the U.S. digital TV standards body. The document addresses the great flexibility, and inherent complexity, of the Dolby Digital multichannel audio system that is part of the ATSC standard.
The document, which was approved earlier this month, provides technical guidelines that program producers, broadcast networks and stations can follow to avoid wide variations in volume between different programs and excessive loudness in commercials. It focuses on audio measurement, production and post-production monitoring techniques, as well as methods to effectively control loudness for content delivery or exchange. It has a particular focus on how stations can ensure that their digital audio encoders are set to accurately reflect the actual loudness of programs and commercials, which then allows consumer DTV sets to automatically regulate the volume, as per the original design of the DTV system.
“What’s important is that everyone is starting to move to uniform practices,” says ATSC President Mark Richer. “We have momentum already.”
While only broadcasters are required by FCC rules to adhere to the audio requirements of the DTV standard, the ATSC’s recommendations on managing loudness are being supported and adopted by cable operators and programmers. The ATSC document also gives specific guidance to cable operators on how to downconvert broadcast DTV signals to serve analog TV viewers without causing loudness problems.
Cable’s cooperation is vital to ensure a consistent listening experience for the majority of TV viewers, says Richer, who notes that a recent ATSC seminar on loudness drew strong participation from the cable industry. John Eberhard, VP of technology and distribution for cable sports giant ESPN, says that ESPN is following “an essentially similar regimen” as the broadcast networks in managing loudness. He sees broad support for the ATSC guidelines among both cable programmers and operators, who have frequently drawn complaints from viewers over the loudness of their local commercials.
“Everyone’s circled the wagons and started talking again, and I expect this is going to get fixed pretty quickly,” says Eberhard, who adds that the NCTA has been working on the loudness issue for several years.
Flexibility, Not Ambiguity
In the analog world, it was standard practice to manipulate program audio to ensure efficient transmission and avoid big volume swings. But the Dolby Digital multichannel audio system for DTV (formally known as AC-3) was designed not only to provide a wider dynamic range, but also to preserve the original audio from the program creator throughout the transmission chain, all the way to the consumer’s HDTV set. To do so, the DTV signal includes metadata indicating the relative loudness of the program that the audio decoder in the consumer receiver relies on to maintain a uniform loudness.
“The decoder in the TV set is being driven by the metadata at the TV station,” explains Jim Starzynski, NBC Universal principal engineer and audio architect, and chairman of the ATSC’s subgroup on digital television loudness. “It’s like a hand on the volume.”
The FCC rule for DTV broadcasts, known as ATSC A/53, requires stations to make sure this metadata parameter of “dialnorm”—short for “dialog normalization”—accurately reflects the average loudness of the “anchor element” in the program audio, which is typically the dialog, so that DTV sets can effectively normalize the loudness. But it doesn’t give much detail beyond that. So some stations measured loudness differently, while others never changed the default dialnorm settings on their AC-3 encoders to match the actual loudness of what they were broadcasting. The end result was wildly fluctuating volume for viewers.
“It’s sort of a backhanded rule about loudness: ‘The transmitted dialnorm value must match the loudness of the program,’” says Tim Carroll, president of audio processing vendor Linear Acoustic. “But it doesn’t tell anybody how to go about doing that.”
That’s where the ATSC recommended practice comes in, with guidelines for how broadcasters should measure program loudness and a loudness target for the content they receive and transmit.
The networks have agreed to measure loudness using the International Telecommunications Union (ITU)’s BS.1770 algorithm, which is based on human hearing and provides a numerical value indicating the perceived loudness of the content being measured. BS.1770-compliant devices measure loudness in the unit LKFS, which is equivalent to a decibel (dB), in numbers ranging from -1 (loudest) to -31 (softest). For example, a -15 LKFS program can be made to match the loudness of a quieter -22 LKFS program by attenuating it by 7 dB.
As per A/53, a particular loudness value in LKFS is equivalent to the same dialnorm number that needs to be set in the ATSC encoder; a -15 LKFS loudness value, for example, has to be matched with a dialnorm number of 15 for a DTV receiver to properly decode the audio and normalize the loudness. While the loudness number could vary from program to program, the networks have recommended the target of -24 LKFS, which research has shown matches the average volume of the widest range of content. Most are shifting their broadcast delivery chain to meet that loudness target.
Audio that doesn’t hit this loudness target either has to be changed to -24 LKFS through processing, which can be done in both real time and the file-based domain, or the dialnorm number on the encoder has to be changed to match the loudness.
“The golden rule is to make sure the program loudness matches the dialnorm value,” says Jim DeFilippis, Fox senior VP of digital television technologies and standards.
This year, Fox shifted its program delivery specification for loudness from -23 to -24 for its programming; NBC is doing the same thing for its network programming and owned-station broadcasts. Stations are buying BS.1770-compliant monitoring devices to measure the loudness of commercials and syndicated programming, and both networks and stations are using a mix of hardware and software-based tools to adjust loudness.
Affiliates are free to set whatever loudness levels they want for their own programming, as long as they make sure the dialnorm number matches it. For example, an affiliate could choose to run its 11 p.m. newscast at a loudness of -27 LKFS. As long as it properly adjusted its encoder, either manually or through automation, viewers shifting from NBC primetime programming at -24 to the local news at -27 wouldn’t notice a difference in volume on their DTV sets.
“As long as the dialnorm matches their loudness, the system will automatically calibrate and smooth out the station’s volume,” Starzynski says.
However, some loudness problems are inevitable despite broadcasters’ best efforts, according to Bob Seidel, CBS VP of technology. CBS has been using a loudness control system on its commercial ingest since late 2007, but that hasn’t solved “boundary layer” problems that frequently occur with dramas.
“We have a lot of shows like Ghost Whisperer and CSI where you’re often dealing with dead people, and the actors are talking softly in muted tones,” Seidel notes. “If you have a soft passage leading up to a commercial break, the consumer might raise the volume a little bit, and then the commercial comes on and it’s blaring.”
Both Seidel and DeFilippis say that program providers have become aware of this boundary layer problem and are making subtle adjustments to the audio going into breaks to address it.
Carroll is hopeful that through voluntarily compliance with the ATSC guidelines, the industry can avoid legislation and program producers can retain all of the creative tools the DTV system affords. “Self-policing works better,” he says. “The film industry did it with trailers. You really don’t want the government telling you how to mix. But that’s where it could be headed.”