Glossary
Album Normalization: A method of Loudness Normalization applied to entire mastered albums. It preserves the relative loudness balance between tracks, respecting the artists’ intent. Album normalization consists of measuring the Integrated Loudness of each track, then calculating the gain or attenuation necessary to normalize the loudest track to a specified Distribution Loudness and applying that gain or attenuation equally to all tracks.
Anchor / Anchor Element: The perceptual loudness reference point or element (typically dialogue) around which other elements are balanced in producing the final mix of the content, or that a reasonable viewer would focus on when setting the volume control. Example: “We normalized this file to –24 LKFS using a dialogue anchor.”
Audiogram: A graph that shows a given listener’s audible threshold for standardized frequencies as measured by an audiometer. The Y axis represents intensity measured in decibels (dB) and the X axis represents frequency measured in hertz (Hz). The threshold of hearing is plotted relative to a standardized curve that represents ‘normal’ hearing, in dB (HL) (hearing level).
Automatic Dialogue Replacement (ADR): A post-production process where actors re-record their dialogue in a studio to replace the original audio captured during filming. This is done to improve sound quality, fix mistakes, or change dialogue after the fact. The new recordings are then synchronized with the video footage, matching the original lip movements and emotional intent.
AVR (Audio Video Receiver): A combination audio amplifier and audio/video switching device for a home theater. It contains inputs for all the audio and video sources and outputs to one or more sets of speakers and one or more monitors (without a tuner) or TVs.
Bottom-Up Processing: Speech understanding that occurs subconsciously and effortlessly. Typically, listeners will subconsciously augment it with Top-Down Processing.
BS.1770 Loudness / ITU-R BS.1770: An electrical measurement using the ITU-R BS.1770 algorithm. This estimates the perceived loudness of content, referenced to 0 dBFS, in a unit of “LKFS.” LUFS and LKFS are interchangeable and specify identical values. The algorithm consists of a K-weighting filter feeding a level-gated RMS detector followed by a logarithmic converter scaled such that a loudness change of 1 LU corresponds to a level change of 1 dB. The first stage is a high frequency 4 dB shelving boost accounting for the acoustical effects of the head; the second stage is a low-frequency roll off based on “B-weighting,” roughly approximating the 70 phon equal-loudness contour.
CODEC (short for Coder-Decoder): An algorithm or device which converts a signal into another format and converts the coded signal back to the original format. For example, MP3, AAC, AC-3.
Content: Material or essence used for distribution by an operator.
Content Dynamic Range (see Dynamic Range).
dBFS; dB FS (decibels full-scale): RMS amplitude of digital samples expressed as a level in decibels relative to a sine wave whose peak level is 100% full-scale and whose frequency is asynchronous with the sample frequency. [per AES17]
dBTP; dB TP (decibels True Peak): Decibels, True Peak relative to 100% full-scale (per ITU-R BS.1770 Annex 2).
Dialog (Dialogue) Integrated Loudness: See Integrated Dialogue Loudness.
Dialogue: Any speech elements within content.
Dialogue Intelligibility: How well speech within content can be understood. It is a continuum between effortless understanding and total unintelligibility regardless of listening effort. In technical literature, the term intelligibility is reserved for the proportion of an unknown, meaningful message that a listener can understand.
Digital Media Adapter (DMA): (also known as a Digital Media Receiver or Digital Media Player): A device that allows digital media sources to connect to home media systems like TVs, soundbars and audio-video receivers (AVRs). It enables the transfer and display of digital content, such as photos, music, and videos, over a network, either wired or wireless. The most common DMAs (for example: Roku, Apple TV, and Fire TV) run multiple applications that each support streaming from one content provider.
Distribution: A Distributor’s act of sending content to end users.
Distributor: The entity in the delivery chain that may be responsible for the normalization of content loudness before release to end users. This entity is typically a network, branded streaming service or on-demand content provider. (See also Content Distributor)
Downmixing: The process of reducing the number of audio channels by mixing some of the source content’s channels together. For example, 5.1 content can be downmixed to stereo. Downmixing can impact dialogue intelligibility.
Downward Normalization: Applying a uniform amount of attenuation to a piece of content to bring its Integrated Loudness down to the desired Distribution Loudness.
DRC Profile: A collection of parameters that describe how dynamic range control Metadata is calculated. For example, Film Standard with AC-3/E-AC-3 or General for MPEG-D DRC.
Dynamic Range (also referred to as Content Dynamic Range): The difference in Loudness between the loudest and softest passages of content, excluding silence. A large difference between the Loudness of the dialog and the Loudness of the Full Program Mix Measurement may be an indication of wide Content Dynamic Range. Similarly, large values of LRA may be an indication of wide Content Dynamic Range.
Dynamic Range Control (DRC): The process of continually adjusting audio signal level to control the loudness difference between loud and soft (quiet) passages. This can help overcome noise in the listening environment, meet the dynamic range capability of the playback equipment, or both. The term typically implies that the codec’s encoder generates the gain control signal as metadata and that the player device’s decoder acts on this metadata, as distinct from “dynamic range compression” performed internally by a single device without DRC metadata.
EDID (Extended Display Identification Data): A data structure provided over HDMI by a display or other device to describe its capabilities to a video source device (e.g., graphics card or set-top box).
Full Program Mix Measurement: The Integrated Loudness measurement using the BS.1770 algorithm, in units of LKFS or LUFS, based on all channels and all elements of the audio Content, over the duration of the Program.
High Resolution Content: Content that is linearly or losslessly encoded. As used generally, losslessly encoded content that exceeds CD sample rate and/or bit depth.
ID3 Tagging (ID3 Metadata): A type of metadata container used to store information about the audio; this information is embedded within the audio file. Allows a file to include relevant information such as Now Playing information (artist name, track title, album, track number, genre) and can be used to access other databases.
Infinite Look-Ahead: A variation of look-ahead technology used in digital audio file processing, where the entire file is first read and a processing algorithm then uses this global knowledge to compute local gain reduction throughout the file with minimal side-effects.
Integrated Dialogue Loudness (aka Speech Loudness, Dialogue(Dialog) Loudness): The BS.1770 integrated loudness of dialogue content that is not mixed with other elements such as a music bed. Measurement requires isolating speech segments of content either manually, or by measuring the dialogue stem, or with an automatic speech detection or gating tool.
Integrated Loudness: The average loudness between two points in time, measured electrically using the BS.1770 loudness algorithm with gating. To estimate the Integrated Loudness of a continuous stream, it is necessary to choose a sliding rectangular integration time window long enough to prevent the measurement from being disturbed by short-term dynamics in the content.
Interstitial Content; Interstitial: (see Short-form content)
Isolated Dialogue (aka Clean Speech): Describes the contents of an audio channel or object containing only dialogue without music or effects.
K-weighting: The frequency-weighting defined in Recommendation ITU-R BS.1770 and applied to incoming audio prior to RMS block measurement in the specified loudness algorithm; the “K” in LKFS refers to this weighting curve.
LKFS: Loudness K-weighted, relative to digital Full Scale. LKFS is synonymous with LUFS.
Long-form Content: “Show or program-related material or essence. The typical duration is greater than approximately two to three minutes.” [per ATSC A/85] Examples include radio shows, concerts, dramas and musical albums.
Look-Ahead: In a dynamics processor such as a compressor or peak limiter, technology that presents the gain-control sidechain with audio that is earlier than the audio applied to the processor’s gain control element. This allows the sidechain to apply strategies, such as ramping the gain down before a peak is encountered, that minimize audible artifacts introduced by limiting.
Loudness: A subjective metric of the perceived audio level. No simple physical property like amplitude fully describes it; it is determined in a complex way by the content’s amplitude and spectrum, and how these vary over time. For the loudness measurement of media content, the algorithm specified in ITU-R BS.1770 is typically used.
Loudness Management: A process that adjusts the Integrated Loudness of content or loudness metadata to provide loudness consistency. Can be implemented via Loudness Normalization and/or Dynamic Range Control (DRC).
Loudness Normalization: A process that applies a fixed gain or attenuation to content so that its loudness matches a specified Target Loudness. This can also be performed non-destructively when loudness metadata accompanies the content and is applied.”
Loudness Range (LRA): A metric (formally defined in EBU Tech 3342) that can provide information about the loudness variation of the full mix or dialogue specifically (when applied to the dialogue track/stem/element).
LU (Loudness Units): A unit of measure defined in ITU-R BS.1771 as the difference between two ITU-R BS.1770 measurements and/or specified target or reference loudness values. A loudness change of 1 LU corresponds to a level change of 1 dB.
LUFS: (see LKFS)
Masking: (aka auditory masking) The hearing phenomenon where the response to one stimulus is reduced in the presence of another, often louder, sound. For example, with respect to dialogue intelligibility, it is phenomenon where competing sounds such as music and effects “drown out” dialogue.
Maximum True Peak (Level): The largest True Peak in the content under consideration. This indicates whether the signal reaches amplitudes that could overload the following stages of an audio system.
Measured Loudness: The value reported, in units of LKFS or LUFS, when an audio signal is measured with equipment that implements the algorithm specified by BS.1770. It is an approximation of perceived Loudness.
Metadata; Audio Metadata: Data (other than essence) included within content, streaming, or broadcasting used to identify, label and present content — including content loudness, codec in use, bitrate, artist, genre, label, and titles.
Microdynamics: The subtle and intricate level variations and dynamic contrasts within a small portion of a sound or signal, often within a single measure of music or a specific instrument’s performance.
MVPD (Multichannel Video Programming Distributor): Includes DBS service operators, local cable system operators, and cable multiple system operators.
Nearfield Mixing: A monitoring technique where the mix engineer sits close to the monitor speakers so that the direct sound from the speakers is substantially more intense than the mixing room’s reverberation. This minimizes the effect of room acoustics on what the mixer hears and can yield an overly optimistic estimate of dialogue intelligibility compared to what the consumer typically hears.
Next-Generation Audio (NGA) codec: A codec, such as MPEG-H 3D Audio, Dolby AC-4, and DTS-UHD, that, when compared to legacy codecs like Dolby AC-3 or AAC-LC, provides additional features such as support for immersive audio, audio objects (including separated dialogue), and more versatile metadata-driven Dynamic Range Control.
Normalization: See Loudness Normalization.
On-Demand File Playback: Playout or streaming of a File through interconnected audio facilities, usually between the point of distribution and the consumer (listener).
OTT (Over-the-Top-Television): The means to deliver video content via Internet streaming, VOD, pay TV, IPTV and download via IP mechanisms.
OVD (Online Video Distributor): Any entity that offers video Content by means of the Internet or other Internet Protocol (IP)-based transmission path provided by a person or entity other than the OVD.
Parallel Compression: A dynamic range compression technique used in sound recording and mixing to implement upward compression, accomplished by mixing an unprocessed audio signal with a compressed version of the same signal. Particularly useful for classical music.
Peak Limiting (Limiting): A process allowing signals below a specified input power or level to pass unaffected, while attenuating the peaks of stronger signals that exceed this threshold. Limiting is a type of dynamic range compression. Clipping is an extreme version of limiting.
Peak Normalization: The practice of adjusting the peak level of audio content to (usually) full scale. This produces widely-varying loudness levels and has been cited as the original cause of loudness wars in broadcast and music production.
PLR; Peak to Loudness Ratio: The ratio of Maximum True Peak level of an audio content segment to its integrated BS.1770 Loudness.
PPM (Peak Program Meter) An audio meter which measures and displays or records the maximum excursion of an audio program. Developed in Europe to measure peak audio signals, complementary to the VU meter, per IEC 60268-10/-18.)
Program: An individual, self-contained audio-visual or audio-only item to be presented in Radio, Television or other electronic media. An advertisement (commercial), trailer, promotional item (‘promo’), interstitial or similar item shall be a Program in this context.
Program Loudness: (aka Full Program Mix Loudness) The average loudness of all elements of the content, between two points in time, measured using the BS.1770 loudness algorithm with gating.
Program to Dialogue Loudness Ratio (PDLR): The difference between Program Loudness and Integrated Dialogue Loudness.
Set-top Box: A type of Digital Media Adapter; a device that accepts an RF or IP signal from a broadcast antenna or cable TV provider, decodes it, and presents it to a TV or audio-video receiver via HDMI.
Short-Form Content: Advertising, commercial, promotional or public service related material or essence. The typical duration is less than approximately two to three minutes. Synonymous with Interstitial Content.
Speech Loudness: See Integrated Dialogue Loudness.
SPL (Sound Pressure Level): Sound pressure level in decibels referenced to 20 µN/m²(µPa).
Streaming: A continuous transmission over a network (typically the Internet) that consists of pieces of content presented sequentially.
Target Loudness: The intended Integrated Loudness or Speech Loudness of a given piece of content or an entire stream.
Top-Down Processing: Listeners using context along with their experience and expectations to help understand speech that cannot be understood by Bottom-Up Processing alone.
Track Normalization: A method of Loudness Normalization applied independently to each music track, so each track plays at the same BS.1770 Loudness. This may conflict with the creative intent of the artists. (See Album Normalization.)
True Peak (Level): Amplitude of peaks (positive or negative) of a digitized signal in the continuous time domain after the signal has been applied to a mathematically ideal digital-to-analog converter and reconstruction filter. True Peaks (aka “intersample peaks”) can be substantially higher than the level of digital samples adjacent to them and can exceed 0 dBTP.
Upward Compression: Upward compression amplifies low-loudness audio while mostly preserving the dynamics of high-loudness audio. Can also be achieved by mixing the output of a downward compressor with its input (parallel compression).
Upward Normalization: Applying a uniform amount of positive gain to a piece of content to bring its Integrated Loudness up to the desired Distribution Loudness. May require peak limiting to prevent peak levels from exceeding 0 dBTP.
