Audio DRC Metadata for Consistent Loudness Without Clipping

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The varying dynamic range and loudness perception across different end-user devices and audio programs lead to an inconsistent user experience during digital audio playback, as existing technologies fail to effectively normalize loudness and control dynamic range across diverse devices and content types.

Innovation Solution

A systematic approach is implemented using metadata-enhanced digital audio coding and decoding, where dynamic range control (DRC) gain values are computed and embedded per frame, allowing for normalization and dynamic range adjustment during playback, utilizing a DRC processor in both encoding and decoding stages to ensure consistent loudness levels across different devices and content types.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Illumination intensity

If dynamic range control is applied to increase average loudness, then the average loudness is improved, but the peak levels must be reduced which causes clipping and audible distortion

Engineering Contradiction:
Improveaverage loudnessVSAvoidclipping distortion
Core Design Contradiction:
Illumination intensityVSObject-generated harmful factors

Solution Approach 1:

The patent applies preliminary dynamic range control during the encoding stage to pre-compress the audio signal's dynamic range before encoding. This preliminary action ensures that peak levels are controlled in advance, preventing clipping during playback while maintaining improved average loudness without requiring additional processing at decoding

Inventive Principle:
Principle #10Preliminary action

2Adaptability or versatility

If different audio programs are encoded with varying dynamic ranges, then the creative flexibility is improved, but the playback consistency across different devices deteriorates

Engineering Contradiction:
Improvecreative flexibilityVSAvoidplayback consistency
Core Design Contradiction:
Adaptability or versatilityVSStability of the object's composition

Solution Approach 1:

The patent embeds metadata parameters including DRC gain values and DRC characteristics indices that describe the dynamic range control applied to each audio program. These parameters are stored with the encoded audio and automatically applied during playback, allowing each program to maintain its creative dynamic range characteristics while ensuring consistent playback behavior across different devices through standardized parameter application

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If per-frame DRC gain values are embedded in metadata, then the loudness normalization precision is improved, but the data rate increases

Engineering Contradiction:
Improveloudness normalization precisionVSAvoiddata rate
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent embeds per-frame DRC gain values in metadata to provide precise loudness normalization control at the frame level. This partial action approach applies DRC processing selectively to individual frames rather than continuously, achieving high precision loudness normalization while minimizing the additional data rate overhead by only encoding necessary gain values for frames requiring adjustment

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP2956936B1Metadata for loudness and dynamic range control
Publication Date: 2019.04.24 APPLE INC
  • EP2956936B1 patent drawingFigure 1
  • EP2956936B1 patent drawingFigure 2
  • EP2956936B1 patent drawingFigure 3

AI summary

An audio normalization gain value is applied to an audio signal to produce a normalized signal. The normalized signal is processed to compute dynamic range control (DRC) gain values in accordance with a selected one of several pre-defined DRC characteristics. The audio signal is encoded, and the DRC gain values are provided as metadata associated with the encoded audio signal. Several other embodiments are also described and claimed.