Dolby has invested many years researching volume level problems in order to develop effective solutions. Dolby's first major solution to resolve volume level issues is a metadata concept featured in Dolby Digital and Dolby Digital Plus. While Dolby audio metadata is well suited to resolving volume level issues (it provides content providers, broadcasters, and operators with a tool called "dialog normalization"), global volume level challenges remain on broadcast systems where Dolby audio formats are not being used. Such systems have no built-in mechanisms such as metadata to manage volume levels across channels, or content. In addition, significant level differences also occur due to level differences between analog and digital transmission or between different input sources connected to a TV set.

Dolby's Innovative Solution
Recognizing the widespread and intrusive issues related to volume variations, Dolby has leveraged its extensive understanding of audio and years of research to create Dolby Volume, the next generation of volume control technology for televisions. Dolby Volume features a new, groundbreaking volume leveling solution that, when integrated into televisions, provides a unique solution that finally overcomes the volume level differences in all content experienced through television sets. By continually monitoring and adjusting the perceived volume level of the audio, Dolby Volume allows any audio source played through a TV's speakers to match a desired target playback level set by the user. The new volume control technology offered in Dolby Volume works regardless of the audio format (such as analog, digital, or low-bit-rate coded). While Dolby Volume's volume leveling function resolves loudness differences in a single-ended way (that is, without pre-processing or metadata), it can also take advantage of and benefit from Dolby's powerful audio metadata when available and provide an even higher quality volume control solution.

How does it Work?
Dolby Volume is comprised of a suite of complementary volume level processing technologies designed to work together seamlessly and provides a complete high-quality audio solution. Unlike previous attempts which have addressed the problem by using standard audio compressors or automatic gain controls (AGCs), Dolby Volume's uniqueness lies in the fact that for the first time the measurement, analysis, and control of the volume level of audio is performed using a perceptual processing engine built on an advanced psychoacoustic model of human hearing. This engine simulates, in real-time, the complex, non-linear mechanisms through which humans perceive loudness and organize sound. With the guidance of this perceptual engine, Dolby Volume knows not only how and when to process the audio but, equally important, when not to process the audio. As a result, Dolby Volume is a new approach uniquely capable of reproducing any audio signal, at any desired playback level set by the user using a standard volume controller, free of any of the audio artifacts (like pumping and breathing) introduced by the previous standard compressor or AGC solutions. Dolby Volume achieves this high-quality and artifact-free performance by combining two patented features derived from its powerful analysis and processing engine:

Volume Leveler
Volume Leveler
Dolby Volume's main feature is the Volume Leveler, a powerful and unique solution that continually measures the perceptual loudness of audio content and dynamically applies multi-band gain modifications to the audio using its perceptual processing engine so that the perceived loudness of the audio content remains consistent. This allows audio from multiple sources (for example different television channels or audio coming from different external inputs on a television set, such as DVD-Players or gaming consoles) to have the same perceived loudness level. When the Volume Leveler is used in conjunction with the volume controller of an audio system, the volume controller is transformed and no longer emulates an electrical resistor that controls the audio signal level being sent to an amplification section. Instead, the volume controller now provides input to the Dolby Volume system indicating the users desired perceived loudness reproduction level. This information is then used to modify the perceived loudness of the reproduced audio to match that of the users desired loudness level.

A comparison between an existing, commercially available leveling solution and Dolby Volume's powerful Volume Leveler is outlined below in Figure 1.

The top of Figure 1 shows the time domain waveform of the original, unprocessed audio signal as a user switches between two different television channels. As shown in the figure, the audio in Television Channel 2 is at a significantly lower level than the audio in Channel 1 and would require the user to modify the volume on the television to adjust for the differences and maintain a consistent experience. The middle of Figure 1 shows the time domain waveform of the audio processed using an existing leveling solution found in a commercially available television. As can be seen in the figure, the leveler does not provide a sufficient increase in gain to match the level between the two channels. The lower part of Figure 2 shows the same audio processed by Dolby Volume's Volume Leveler, which has properly matched the level between the two channels. While an audio signal's time domain waveform does not specifically indicate the perceived loudness of an audio signal, it should be noted that Dolby Volume's Volume Leveler has adjusted the audio in Channel 2 so that it matches the perceived loudness of the audio in Channel 1. This is possible because of Dolby's powerful new perceptual audio processing engine that understands how we perceive loudness and organize sound.

A comparison
A comparison between a second, commercially available leveling solution and Dolby Volume's powerful Volume Leveler solution is outlined in Figure 2. In Figure 2, the audio being processed is from a film soundtrack where the audio switches between dialog and quiet background wind sounds. In this case the wind sounds should remain quieter than the dialog to maintain the intelligibility of the dialogue and the overall balance of the soundtrack. By comparing the processing results of the two existing, non-Dolby leveling solutions shown in Figures 1 and 2, a common and serious problem found with existing leveling solutions is highlighted. In previous, non-Dolby Volume solutions, either the leveler does not provide sufficient gain to correct the large variations in loudness found between different television channels (as exhibited by Existing Leveler 1) or the leveler introduces an inappropriate and time-varying increase in the amount of gain applied to the audio (too much, inappropriate gain applied to the wind sound by Existing Leveler 2, commonly referred to as a compression "breathing" artifact). Only Dolby Volume's Volume Leveler solution, which a contains powerful loudness and audio content analysis processing capability, provides the appropriate amount of gain for all types of audio content.

Because of Dolby Volume's new approach to controlling audio in the perceptual loudness domain, the Volume Leveler allows high-quality and artifact-free control of the audio, heretofore unavailable from previous solutions. Dolby Volume truly provides a paradigm shift in next generation audio control by very effectively eliminating loudness level differences while at the same time maintaining the integrity of the audio signal without imposing any quality degradations.

Volume Modeler
Volume Modeler
The Volume Modeler technology featured in Dolby Volume was originally developed based on research for Dolby's Professional Digital Cinema systems. In Digital Cinema, audio content is mastered and expected to be reproduced at a specific reference level (85 dB C-weighted measurement of pink noise in one channel, or a setting of 7 on a Dolby Cinema Processor) in Dolby equipped theatres. However, many times theatres play back soundtracks at reduced levels to address complaints from movie audiences who often perceive the sound in theatres to be too loud when played back at reference level (setting 7). Due to the variation in the perception of audio at any given reproduction level (which is related to psychoacoustics and the equal loudness contours and the threshold of hearing in quiet), reduced playback levels result in dramatic differences in timbre and spatial reproduction of the audio when compared to when the same content is played back at reference level.

Dolby Volume's Volume Modeler technology continually analyzes and modifies the audio depending on both the content and the playback level, restoring it to the way it would be perceived at the reference playback level, resulting in improved imaging, intelligibility and audibility of the audio content, for example improved audibility of the surround channels in multi-channel content when downmixed and mapped to stereo for a 2-channel reproduction through the speakers of a television set and played back at lower than reference levels).

An example of the content and playback level dependent processing provided by Dolby Volume's Volume Modeler is outlined below in Figure 3.

As shown in the top left of the figure, when a louder section of audio is reproduced at the reference level, nearly all of the audio is above the threshold of hearing and therefore audible. If this same audio is reproduced 40dB quieter using a normal volume control which applies the same gain to all frequencies and channels (top middle of the figure), some of the audio's lower frequencies and all the audio's frequencies above 10 kHz are below the threshold of hearing and therefore inaudible. The result is a dramatically different listening experience when compared to the audio being played back at the reference level. However, when the same audio is played back at the lower level using Dolby Volume's Volume Modeler (top right of Figure 3), the audio is processed such that all frequencies of the audio remain audible, even at the reduced reproduction level, thereby maintaining the reference level experience for the listener. The relationships between frequencies are maintained such that the original frequency relationships or timbre are preserved.

The lower half of Figure 3 shows the results for a second, softer section of audio, for example low level ambience in a soundtrack. In this case, when the softer audio is played back 40dB quieter with a normal volume control (lower middle of the figure), all of the audio falls below the threshold of hearing and the channel is completely inaudible to the listener. However, when the same audio is played back using Dolby Volume's Volume Modeler (lower right in Figure 3), the audio is processed such that all frequencies that were audible at the reference playback level of the audio remain audible, even at the reduced reproduction level, thereby maintaining the reference level experience even for softer sections of audio.

Dolby's Concept of Audio Metadata
How does this affect Dolby's Concept of Audio Metadata?
Dolby Volume has been carefully designed to work seamlessly in conjunction with the metadata included in Dolby Digital and Dolby Digital Plus. In the many cases when metadata is present and has been set correctly, Dolby Volume complements this concept by staying subtly in the background until the next level change occurs, from switching the channel on the TV or if a program element is broadcast that may not have Dolby audio metadata. In such examples, Dolby metadata is more important than ever and broadcasters and operators should insure that they set dialnorm and Dynamic Range Control profiles correctly for each program element. This can be most easily achieved using the Dolby DP600 program optimizer. In such an environment Dolby Volume will focus on eliminating the level differences between content and input sources delivered in different audio formats (for example MPEG1LII used for legacy stereo in DVB applications).

Every television set using Dolby Volume will be calibrated by the manufacturers in the factory in order to allow for perfect interoperability between Dolby Audio metadata, the Volume Leveler and the Volume Modeler.

About the author
Page Shaper Haun is theWorldwide Director of the Broadcast Business Segment of the Consumer Division of Dolby Laboratories. . As the worldwide director, Page Haun is responsible for worldwide marketing, strategy and deployment of Dolby's audio technologies to the cable, satellite and telecommunications industries as well as to television manufacturers. Ms Haun joined Dolby in February 2006 and has held strategic marketing and management positions for some of the world's leading cable service providers. As the Corporate Director of Marketing for New Products at Charter Communications, she oversaw the company's digital video recording, HDTV service and interactive television initiatives. Previously, she worked at AT&T Broadband, where she led partner marketing and retail efforts. Ms. Haun has a bachelor's degree from Vanderbilt University and an MBA from Harvard Business School. She was named among the "Women to Watch" in the Women in Cable Telecommunications Association in 2004. Ms. Haun can be reached at pshaun@dolby.com.