ISO IEC 23003-3-2020 PDF

Full title and description

Information technology — MPEG audio technologies — Part 3: Unified speech and audio coding (ISO/IEC 23003-3:2020). This international standard specifies the Unified Speech and Audio Coding (USAC) codec: a single codec framework designed to efficiently code signals containing arbitrary mixes of speech and general audio content, supporting single- and multi-channel operation across a wide range of bitrates.

Abstract

The standard defines a unified speech and audio codec that combines perceptual audio compression techniques (perceptually shaped quantization noise, parametric coding of high-frequency content and stereo stage) with a source/model-based approach for speech production. USAC is intended to achieve perceptually transparent quality at high bitrates while providing very efficient low-bitrate coding with full audio bandwidth. The document includes bitstream syntax, decoder behaviour and tool descriptions; reference software and formal conformance procedures are provided via an amendment.

General information

• Status: Published
• Publication date: 24 June 2020
• Publisher: ISO / IEC (ISO/IEC JTC 1/SC 29 — Coding of audio, picture, multimedia and hypermedia information)
• ICS / categories: 35.040.40 (Coding of audio, video, multimedia and hypermedia information)
• Edition / version: Edition 2 (2020)
• Number of pages: 339

Scope

ISO/IEC 23003-3:2020 (USAC) specifies a unified codec framework for efficient coding of content containing mixtures of speech and general audio. The scope covers codec architecture, individual coding tools (time- and frequency-domain tools, parametric enhancement tools, speech-model tools), bitstream syntax, decoding processes and required decoder behaviour. It addresses single- and multi-channel configurations, a broad bitrate range from very low bitrates up to transparent quality, and interacts with related MPEG audio tools and systems. Reference software and conformance material are available through subsequent amendments.

Key topics and requirements

• Definition of the Unified Speech and Audio Coding (USAC) architecture and operation modes.
• Perceptual coding techniques including perceptually shaped quantization.
• Parametric coding tools for upper-spectrum and stereo/stage representation.
• Speech-production model integration for improved low-bitrate speech coding.
• Bitstream syntax, payload formats and decoder behaviour requirements.
• Support for single- and multi-channel audio configurations.
• Conformance testing and reference software provided via amendment (reference software and conformance procedures published as ISO/IEC 23003-3:2020/Amd 1:2021).
• Provision for future extensions and amendments (e.g., draft work on media authenticity features).

Typical use and users

USAC is used by codec implementers, silicon and consumer-device manufacturers, streaming and broadcast platform engineers, audio middleware developers, and researchers working on audio compression and perceptual coding. Typical applications include streaming audio services, digital broadcasting, multimedia file formats, low-bitrate voice+music transmission, and any product or service requiring a single codec to handle mixed speech and music content efficiently.

Related standards

ISO/IEC 23003 is the MPEG-D (MPEG audio technologies) series. Closely related documents include ISO/IEC 23003-1 (MPEG Surround / spatial audio), ISO/IEC 23003-2 (Spatial Audio Object Coding, SAOC), ISO/IEC 23003-4 (Dynamic Range Control) and ISO/IEC 23003-5 (Uncompressed audio in MPEG‑4 file format). USAC is also related to MPEG-4 Audio specifications (for example ISO/IEC 14496-3) where USAC tools are used or referenced. Conformance and reference software for ISO/IEC 23003-3 are published as an amendment to the 2020 edition.

Keywords

USAC; Unified Speech and Audio Coding; MPEG-D; MPEG audio technologies; audio codec; speech coding; parametric coding; perceptual coding; bitstream syntax; conformance; reference software; multi-channel audio.

FAQ

Q: What is this standard?

A: ISO/IEC 23003-3:2020 specifies the Unified Speech and Audio Coding (USAC) codec — a standardized codec for efficiently encoding signals containing mixtures of speech and general audio.

Q: What does it cover?

A: It covers the codec architecture, coding tools (perceptual and parametric techniques plus speech-model tools), bitstream and payload formats, decoder behaviour and the intended operating modes for single- and multi-channel content across a wide bitrate range. Conformance and reference software are available via amendment.

Q: Who typically uses it?

A: Codec developers, device and SoC manufacturers, streaming and broadcast engineers, multimedia application developers and audio researchers use the standard to implement interoperable encoders and decoders for mixed speech/music content.

Q: Is it current or superseded?

A: The 2020 edition (Edition 2) is the current published version of ISO/IEC 23003-3. It is maintained by ISO/IEC JTC 1/SC 29 and has at least one published amendment for reference software and conformance (2021). Users should check for any later amendments or revisions when implementing conformance requirements.

Q: Is it part of a series?

A: Yes — ISO/IEC 23003 is the MPEG-D (MPEG audio technologies) series. Part 3 (USAC) is one part of that series alongside Part 1 (MPEG Surround), Part 2 (SAOC), Part 4 (Dynamic Range Control) and Part 5 (Uncompressed audio in MPEG-4 file format).

Q: What are the key keywords?

A: USAC, unified speech and audio coding, MPEG-D, codec, parametric coding, perceptual coding, speech model, bitstream, conformance, reference software.