Mitigating loud commercials, part 2

This second part of our CALM Act tutorial addresses implementation. The law goes into effect on Dec. 13, and every broadcast engineer should be familiar with the law, its implementation and implications. However, the fact is that much of the law was implemented 17 years ago when the FCC adopted the ATSC standard, including A/52 and A/53. A/52 describes the audio codec, and A/53 describes the DTV format. Both identify the golden rule of loudness, which is that the loudness of normal spoken dialog and the metadata parameter called dialnorm shall match. If all content obeys this rule, the loudness will be consistent.

The CALM Act was passed to eliminate letters, phone calls and e-mails to the FCC or other federal government officials, elected or appointed, complaining about loud commercials. If no viewers complain about loud commercials on your station, then your station is in compliance. However, once the FCC identifies a “pattern or trend” of loud commercial complaints on your channel, be prepared to defend your station with documentation, because documentation is all you’ve got to prove your Innocence.

The great side benefit of the CALM Act is that it opens the possibility for less processing and better dynamic range, presenting the sound track as the content producer intended it to be heard. On the other hand, footnote 119 on page 18 of the FCC Report and Order says not necessarily.

It says, “A station or MVPD can install, utilize and maintain, in a commercially reasonable manner, a real-time or ‘conventional’ processor to ensure consistent loudness by limiting dynamic range, rather than by setting the dialnorm or meeting the Target Loudness. Conventional processing modifies the dynamic range of the decoded content by reducing the level of very loud portions of the content to avoid annoying the viewer and by raising the level of very quiet portions of the content so that they are better adapted to the listening environment.”

In other words, if traditional limiting and compression works for you, fine. Is it the best you can do? No.

Do the right thing

DTV allows broadcasters to abandon limiting or compression at the transmitter or STL because a digital transmitter can’t be overmodulated. Some broadcasters happily removed or reduced their downstream audio processing and limiting and depend on VU meters at ingest and unity gain throughout the facility. The unintended result was more loud commercial complaints. The key to transmitting the best audio viewers can hear is by implementing the practices recommended in ATSC A/85. The dialnorm value should equal an asset’s LKFS level (Loudness, K-weighted, relative to Full Scale), and in general terms Full Scale is 0dB plus headroom. A dialnorm value of 24 means normal spoken dialog is 24dB below full scale digital.

Depending on duration, content loudness is measured using one of two techniques. One is for short-form commercials running 2 minutes or less. The other is for long-form content over 2 minutes. For short-form content, all audio channels are measured over the entire length of the segment. This can be accomplished in the file domain from frame one to the last frame. Therefore, in the file domain, loudness measurements are repeatable to 0.1dB or better, well within the required +/- 2dB. Determining short-form loudness is pure math.

Measuring the loudness of long-form content can be more complex. This is because the A/85 recommended practice calls for measuring all audio channels but only during normally spoken dialog sections — the parts a viewer would typically focus on or use as an anchor. This can be done manually where an experienced mixer decides where in the program to measure, and it can also be done in an automated manner in the file domain.

In other words, the person making the measurement gets to choose what part or parts of the content to check. Unless measurements are performed in non-real time on exactly the same frames of audio, repeatability is nearly impossible to obtain. If commercials are included in long-form content, such as content from a network or syndicator, they are considered part of the program for measurement purposes. This means that numbers upstream and downstream can be different.

Products that measure loudness of pre-produced short- and long-form content provide only an after-the-fact a report card. Measuring the loudness of live television such as news and sports relies on live metering and a mix engineer who uses a loudness meter as a general guide. A loudness review can benefit audio mixing engineers by identifying loudness issues in their mix and training their ears to better mix for dialnorm loudness.

Loudness control of live television is then up to the mixing engineer, and there are a growing number of effective new loudness metering products that can be very helpful. The basic rule for mixing engineers is to keep hands off the audio control room monitor volume control. Keep in mind that all this loudness commotion is aimed to make happy viewers, who may ultimately be counted in ratings.

The FCC Report and Order calls measuring the loudness of content broadcast after-the-fact “spot checking,” and it is mentioned 159 times in the document and its footnotes. Spot checks are annual and unannounced. They can be continuous or for 24 hours. They are looking for ATSC compliance. Variations of more than +/-2 LKFS from dialnorm are not in compliance. This is an average measurement of audio across the entire segment. It is not instantaneous and is not the dynamic range.

While the “golden ears” community may argue about LKFS comfort ranges, spot checks are the best method available to engineers for identifying compliance issues before anyone complains.

The answer

The best long-term response to the CALM Act is to demand better loudness care at every stage. It’s not all that complicated to maintain compliance at the affiliate level once the inventory has been tested, but networks and MVPDs are swamped with original content and laboring to obtain, maintain and document loudness compliance. Typically, incoming content that doesn’t comply is run through an automatic scaler. A scaler adjusts the average audio to match a target. When performed in the file domain, scaling can be done much faster than real time. Scaling can also be done manually in real time, but this can be time consuming.

Networks and syndicators are all working on providing loudness certified content because affiliates and MVPDs can’t do it all alone. Without everyone working together on long-form and short-form loudness compliance, the CALM Act could become the PALM Act.

A great deal of information about the CALM Act and Report and Order can be found here.

The author wishes to thank Andrew Sachs with Volicon and Steve Smith with Broadcast Technology Consultants for their assistance and information they provided.