This article is taken from a paper called: Loudness Issues Are Dead, Long Live Dynamic Range Problems! The paper was presented at NAB 2016. The author of the original article is Ken Tankel, Telos Alliance, Cleveland, OH.
The abstract is updated and an addendum is added here.
Abstract
The CALM Act went into effect December 15, 2010. After 15 years Television broadcasters, for the most part, have a handle on loudness. Commercials are under control and loudness complaints are minimal. Does that mean we're all done with the business of managing television audio? As it turns out, the answer is, "Not quite!". True to Newton's Third Law, "For every action, there is an equal and opposite reaction", well-intentioned but misguided efforts at managing loudness have taken their toll on broadcast audio. In this presentation, we will explore why audio production had to change with digital delivery to the home and how TV audio can be both compliant with loudness requirements AND engaging.
Original Article
Around the world, complaints about loud commercials come from viewers soon after broadcasters transition from analog to digital television delivery. In the U.S., the transition occurred Friday, June 19, 2009. Loud commercials became enough of a priority for the public that, by December 13, 2011, the U.S. Congress adopted the Commercial Advertisement Loudness Mitigation (CALM) Act. Described briefly, the CALM Act requires the Federal Communications Commission (FCC) to enact rules that require commercials to have the same average volume as the programs they accompany. The FCC rules requiring broadcasters to be compliant with the CALM Act went into effect a year later on December 13, 2012. The number of commercial loudness complaints logged by the FCC, in the month following enactment of the CALM Act, was 4,7771.
There will always be complaints about loud commercials. The FCC received 192 complaints about commercial loudness in the third quarter of 2012. Given plenty of advertisements about the CALM Act and the opening of an FCC website for the public to register complaints, people voiced their displeasure. The FCC website was open for business the following month, December 2012, and nearly 5000 complaints were received. In the closing months of 2013, while the opportunity to voice loudness complaints remained, the number of complaints dropped to no more than 656 in any month. Clearly, there has been a drop in complaints since the CALM Act went into effect.
Worldwide, the conversion from analog to digital television delivery is extensive. The reaction of viewers around the world appears to be quite consistent when it comes to loudness issues. It is now becoming difficult to find someone involved in the delivery of digital television over terrestrial, cable, satellite, or IP, who is not aware of audio loudness recommendations and regulations. Broadcasters that make the transition to digital soon adopt industry recommendations or are required to comply with national regulations regarding commercial loudness. North America, some Central American countries, Korea and a few others follow the ATSC A/85 loudness recommendations. All members of the EBU are expected to follow the EBU recommended practice expressed in EBU R-128, Loudness Normalization and permitted Maximum Levels of Audio Signals. Some countries have introduced their own changes to an existing standard. Australia adopted a standard based on ATSC A/85 and Brazil adopted a standard based on EBU R-128. Japan and China developed their own loudness standards. It must be noted that the ITU-R BS. 1770 measurement standard, “Algorithms to measure audio programme loudness and true-peak audio level”, is THE standard for digital television delivery audio loudness measurements all over the world.
It should not have been a surprise that commercial loudness, as well as program-to-program level variations, would be a problem with digital broadcasting. Digital television broadcasts can deliver multichannel sound, 20 Hz to 20 KHz frequency response, typical THD + noise below .01%, a noise floor below the capability of almost all home audio playback systems, and a very large dynamic range. These capabilities far exceed the limitations of analog television audio delivery. The large dynamic range, in particular, can provide engaging and exciting audio in viewers’ homes. It is this dynamic range that plays a part in the problem of loud commercials. It has also become a casualty in the effort to control commercial loudness.
Analog television has built-in loudness limits. The maximum allowable modulation of the FM audio part of the TV signal has to be limited to protect the video signal and adjacent RF channels, from interference. This sets an upper limit on the received audio level. Low audio FM modulation levels result in an increase in noise and distortion. The use of audio compression and limiting is a requirement for analog delivery of audio programs. It is needed to both meet technical requirements and to present acceptably low noise and low distortion audio to viewers.
Dolby Laboratories did research on some of the audio preferences of home viewers. It is interesting to note how closely analog modulation limits and viewer’s loudness preferences align. Figure 1 shows FM audio modulation levels compared to viewer loudness preferences.
Figure 1- Dolby Laboratories Viewer Preferences and FM audio Modulation Comparison
Dolby Laboratory’s chart of viewer preferences is shown with audio FM carrier modulation beside it. 100% modulation is placed to coincide with the level where viewers will reach for the volume control. For analog TV, audio processing is used to keep average modulation between 25% and 50% which is a very close match to what the chart shows as the Comfort Zone. It is also used to limit both peak and minimum levels. 100% modulation is a hard limit, so program peaks are limited to this value. The peak headroom depends on the average modulation level. Many broadcasters set nominal levels of -18 dBFS allowing 18 dB for peaks. However, this puts the average level at only 12 % modulation. Theoretically, this level is 42 dB above the noise floor of an analog transmission system. Practically speaking, it is closer to 32 dB above the noise and only about 8 dB above the viewer’s desire to increase the volume. Again, audio processing can be used to both reduce high levels and increase low levels in order to maintain average program levels where the broadcaster chooses to place them and then to limit peaks.
Compare these analog audio levels to digital broadcasting where the target loudness level is -23 or -24 dBFS, and the useable noise floor ranges from -60 dBFS (limited by the TV receiver and audio system) to – 90 dBFS, or better, for a good home theater system. The theoretical noise floor in the digital broadcast system can reach from -96 dBFS for 16 bit resolution files to -144 dBFS for 24 bit resolution audio files. Now, peaks can be 23 or 24 dB above average program level. True Peak (TP) levels have recommended limits of -1 or -2 dBFS, providing 22 dB of useable headroom. A home theater system that takes a digital audio output from the transmission can have a noise floor that is 50 dB, or more, below average program level. There is no noise, distortion or frequency response penalty for low levels in a digital delivery system, unlike FM modulation. So, how does this relate to commercial loudness issues?
Creative use of compression and limiting has long been used to increase the average volume and maximize the impact of commercials. In analog broadcasting the commercials could sound louder than program audio, and some commercials did. However, the modulation limits inherent in analog transmission, combined with audio processing, prevented the level differences from being very large, let alone extreme. In digital broadcasting the modulation limits are gone. Audio put in the broadcast end can emerge at the viewer’s end untouched. Using digital audio production tools commercials and other content can be produced with average levels of -12 dBFS with peaks at -2 dBFS. Now the difference between the minus -23/-24 dBFS target for program levels and the level of commercial content can be very large and very annoying. Even content that is not radically compressed, limited, and normalized to very high peak and/or average levels, can be noticeably louder than program content. This is where the commercial loudness problems begin in digital delivery systems.
Where compliance with regulations has been mandated, broadcasters responded immediately by controlling loudness in the air-chain. A single device in the air-chain can ensure that loud content, from any source, is controlled. It is a set and forget solution that does not require any additional personnel, expertise, or time to achieve compliance with loudness targets. To a great extent this is the same approach used by broadcasters when compliance with loudness recommendations is voluntary. The good news is that with proper audio loudness control there are fewer complaints. This is proving to be true all over the world. The bad news is that, in many cases, the audio that is delivered is needlessly and radically reduced in dynamic range. Not only are long term level changes nearly eliminated but the short term level changes and peaks may be reduced or eliminated entirely. Audio controlled to extremes is described as compressed, dull, lifeless, and can be quite different than the program creators intended. Why does it have to be this way?
Real time processing must be set to control the most out-of-compliance content that is played through it. When loudness limits are initially imposed there is typically a mix of content that is much higher, and also much lower, than the target loudness levels. When a 10 second advertisement is delivered at -12 dBFS the real time loudness control must be set to act and reduce the audio to the proper average loudness level within the 10 second duration. To do this in real time requires very fast gain reduction. Unfortunately, the settings required to reduce a very loud, short duration, advertisement will also reduce levels in a program segment that are intended to be loud. There is nothing wrong with loud audio! It is not disallowed by loudness regulations. A program segment can be in compliance and still contain very loud, and very soft, music and effects and even dialogue. Programs and commercials should have the same average levels, not the same instantaneous level throughout a program! The ATSC A/85 document says that normally spoken dialog, of all content types, must average -24 LKFS +/- 2 dB. Audio levels can vary as creativity and intent allow. Audio level changes are the source of much of the drama, and fun, of television programs. Even loudness controllers that are capable of fine control, allowing high peak levels while still maintaining long term loudness levels, will change compliant content when configured to control content of wildly varying levels.
Digital broadcasting is capable of delivering engaging and exciting program audio. The easiest methods of complying with loudness regulations are reducing the truly amazing capabilities of digital delivery to analog proportions. There are a number of steps that can be taken to improve both compliance and audio quality. Any one of these steps will have a positive impact on program dynamic range. Used in combination air-chain loudness controllers can not only help to control loudness but can also improve audio for the viewer.
Some broadcasters are using file based audio measurement and normalization to make ingested content comply with loudness regulations. Content is measured when it is ingested. If it does not meet broadcast loudness requirements it can be: 1) returned to the provider for them to make the appropriate changes or; 2) normalized by the broadcaster to meet broadcast requirements. If 100% of a channel’s broadcast content can be corrected on ingest, or before broadcast, then the need for large amounts of real time processing in the air-chain is greatly reduced. If there is no live content, or if live content can be accurately and reliably controlled by the audio mixer, real time loudness control might even be eliminated. This solution has proven to be very effective in allowing some broadcasters to reduce the control that real time processors have to impose. When all content in the air chain is close to the correct loudness then the real time processor can be set to correct minor issues rather than maintain compliance.
Part of reducing the need for the broadcaster to adjust content is getting program suppliers to deliver content that is already compliant. Having delivery requirements that inform program providers about what audio standards they need to meet puts the responsibility for proper loudness levels on the program creators. Broadcasters should be able to tell providers how content will be measured; what the target measurements are; what are acceptable variances to the measurements; what will happen if submitted content does not comply with the standards and; what is the basis of the requirements? If program providers understand what is expected from them, they have an opportunity to deliver compliant content while retaining full creative control.
When content in the playout system is at, or near, the correct loudness levels, it is time to reduce real time compression and limiting in the air chain. Real time loudness controllers should only use the minimum control required to correct the content that is being played! Minimum real time loudness control also requires that all live content be well controlled. This means that all audio operators must understand proper loudness levels, have proper loudness metering and be willing to deliver audio at the proper levels. For live 5.1 surround mixing the monitoring environment must be properly calibrated. Too often the center channel level is not reduced to compensate for the mixer’s position being closest to the center speaker. This invariably results in dialogue that is particularly low in the downmix. Anecdotal evidence indicates that if long term dialogue level is more than 6 dB LKFS below long term average program level, then the dialogue is too low for home viewers.
Many broadcasters feel that a real time loudness controller is required in the air chain to ensure control of any and all content that may be broadcast. If the above steps can be taken, and content that is close to the correct level is played out, then fine tuning an air chain processor can help deliver audio that sounds great. Some real time loudness controllers provide a very fine control over a large number of audio parameters. Stations with specialized broadcasts such as old TV programs, movies, sports, or music, to name a few, can benefit from tuning their loudness controllers to enhance this content. Older programs tend to have higher noise floors and narrower frequency response than modern productions. An in-line loudness controller with advanced features can help make this content sound its best. The very wide dynamic range of Hollywood style productions is beyond many thin panel TVs, and broadcasters may elect to judicially reduce the dynamic range of these films to fit the viewer’s device. Live sports can often benefit from frequency response tailoring and increase in commentator level. Music programs can be tailored to provide a radio approach to broadcast processing. When freed from simply reducing loud commercials, in-line loudness controllers can help make audio more consistent, reduce unwanted noise, and even enhance audio according the broadcasters mission.
There is no single solution to getting audio loudness levels right. A watchful ear and an understanding of what different kinds of audio correction do will help to configure your audio tools. Even file-based measurement and scaling is not perfect unless great care is taken in how it is done. As an example, Figure 2 shows a measurement of a movie segment sent to me by a client. The piece was described by the client as being difficult to correct. The ATSC A/85 recommended target level of -24 dBFS is shown in green. The red line is the integrated LKFS level of the entire program. The dialog level, shown in blue, is a Dolby Media Meter Infinite Dialog measurement. The yellow line indicates the maximum TP level(s) in the segment. At -2 dBFS it appears above the scale for the graph. The waveform was created by graphing ITU BS. 1770-2 measurements of rolling 3 second integrations.
Figure 2: LKFS and TP Measurements of a Movie Soundtrack Segment
Correcting this program segment by simply scaling the gain up 3dB will have several consequences. The overall level of the segment will be increased 3 dB to bring the measured integrated loudness of -27 LKFS up to -24 LKFS. The segment will now comply with this broadcaster’s target of -24 dBFS. However, this gain shift will put the maximum TP levels above 0 dBFS. So, it is imperative that a normalization system be able to limit True Peaks to the allowed maximum level. Hopefully this will be done in a way that does not simply clip the peaks.
After gain correction the segment will have the correct average level of -24 LKFS. However, the A/85 recommendation is that, if possible, the dialogue level should be at the -24 LKFS target. While this should be the responsibility of the program creator the dialogue level can be measured by the broadcaster. There are several methods for doing this even though dialogue is within the program. The Dolby Media Meter showed an Infinite Dialog measurement at -31 LKFS. Should the program be scaled up 7 dB? This will place the dialog at the target but the average program level is now at -20 dB and the peak correction required will be very large. Using integrated program level or integrated dialog level is allowed in A/85. Using dialog level is preferred when possible. Which is correct? The answer is that neither is ideal for the viewer though both may be considered compliant with the CALM Act.
This is an example of how difficult it is to do a good job of preparing wide dynamic range content for home viewers. The dialogue, music, and effects levels in this movie segment may have been fine in a movie theater. However, a viewer at home does not listen at theater levels. Most home theater systems cannot produce theater levels. The speakers and amplifiers in the typical thin panel TV cannot approach the output level and dynamic range required to reproduce this audio track like it was heard in the theater.
This illustrates a situation where dynamic range control, or better yet, a combination of gain scaling and dynamic range control, can really work to improve the viewer’s audio experience. It also shows that the more program producers learn about broadcast audio loudness requirements and the home viewing experience, the better they can provide content that is already compliant with broadcast needs. Knowledge of the requirements will give content providers more control over the audio that viewers hear. The judicious use of audio processing can improve the presentation of wide dynamic range content in the home when necessary. It would be better if the content creators did this themselves, retaining creative control.
There is another model for broadcast loudness control. This model can allow content with proper levels to pass through the air chain with no changes. New standardized metadata structures like Dolby EMDF can carry program and loudness statistics enabling an instant snapshot of the entire program, if pre-recorded, or a snapshot of dynamic statistics for live programming. This metadata is intended to travel alongside of, or as part of, the content itself. This information, or the lack of this information, can be used to impose loudness control when necessary, and conceivably, apply only as much control as the broadcaster deems necessary for compliance. This means that content that is marked as compliant could pass untouched.
Further, since such a system would operate from production through distribution and finally to delivery to the viewer, it could help better match the audio content to the delivery platform. From home theatre systems to television sets, handheld and mobile devices, tablets and PCs - each platform would get the most appropriate version of a single audio asset.
There are already some production/playout systems that can do this for a very limited portion of content’s life cycle and only in proprietary systems. Recent standardization of several parts, including Dolby EMDF, will speed the completion of the flow of this useful information through professional environments and to consumers, thus enabling the full performance of digital television audio to be appreciated by the entire viewing audience. Until then, the only sure way to get the best quality to viewers is for everyone involved in audio, from content creation to delivery, to continue to learn about, and care about, television audio.
Addendum
California’s SB 576
The passing of California’s SB 576 ("No More Loud Commercials" law), signed in October 2025, mandates that streaming (OTT) service advertisements cannot be louder than the surrounding program content. Effective July 1, 2026, this regulation extends the Federal CALM Act (ATSC A/85) standards to OTT platforms like Netflix and Hulu, requiring ad audio levels to match program audio. While this regulation was only passed in California, it will have wide-reaching effects on audio production. Program creators and distributors are not going to create separate versions of content solely for distribution in California. Instead, distribution will remain unchanged, and ALL programs OTT will be loudness controlled to meet California’s new standard. This presents a new opportunity for Telos Alliance ATS file based-audio processing.
Audio Level Correction vs Quality
There is a very large difference in opinion about what audio processing for TV audio should do. The U.S. broadcaster’s view, shared with Canadian and Mexican broadcasters, and in fact with many broadcasters around the world, is that if all the audio delivered to viewers can be improved, it should be. There is no reason that poorly produced audio should be delivered to the home if it can be improved. The need to make audio complaint with local loudness regulations can be combined with the processing necessary to make every program sound good. Or, if not good, at least sound better than it would sound without any processing. In contrast to this view is the opinion that a broadcaster should deliver program content to the viewer exactly as the broadcaster receives it. Compliance with applicable audio loudness regulations is a legislated necessity. But, if the audio sounds bad that is the way viewers should hear it. Correct the loudness but do this with as little change in timbre, and noise level, as possible.
These two, very different, goals have given rise to two very different approaches to TV audio processing. One is exemplified by the Linear Acoustic AERO processors. The other by LA5300 with APTO and Junger’s Level Magic. AERO has the tools to improve audio and achieve a high degree of consistency between good audio sources and not so good, or even terrible, audio sources. APTO and Level Magic attempt to bring every audio source to the proper loudness level, leaving little or no evidence that the audio was processed at all. A critical element of selling Linear Acoustic and Junger audio processing globally is to understand a client’s predisposition for one over the other.
These two processing personalities are leading Junger to add processing blocks to their products that will provide qualitative audio control to their level control only products. The same is true of Linear Acoustic NGA products with APTO processing.