Dynamic Audio Equalization Using Instantaneous Loudness Metadata
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional loudness equalization methods neglect the time-varying levels of spectral bands in audio content, leading to spectral distortions due to nonlinear loudness perception, and often require complex real-time processing with high latency.
Innovation Solution
A metadata-based loudness equalization scheme that uses instantaneous loudness values to control a time-varying spectral shaping filter, compensating for spectral distortions by adjusting the playback level, which can be seamlessly integrated into existing MPEG-D DRC standards, reducing complexity and latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional adaptive filter loudness equalization is used, then loudness perception is improved, but processing complexity and latency increase
Solution Approach 1:
The patent pre-calculates and stores equalization filter coefficients at the encoding stage based on program loudness metadata, eliminating the need for complex real-time adaptive filtering at playback. The pre-computed coefficients are retrieved and applied directly, significantly reducing decoder complexity while maintaining accurate loudness equalization.
Solution Approach 2:
The patent uses program loudness metadata (a copy of loudness information captured at encoding) to control the equalization process at playback without requiring real-time analysis of the audio signal. This metadata copy enables the system to reproduce the intended loudness characteristics without complex real-time processing.
2Reliability
If real-time adaptive filtering is implemented, then loudness equalization quality improves, but processing delay increases
Solution Approach 1:
Equalization coefficients are pre-computed during encoding based on program loudness metadata and stored for later use. At playback, these pre-computed coefficients are retrieved and applied immediately without requiring real-time signal analysis, eliminating processing delay while maintaining equalization accuracy.
3Manufacturing precision
If multi-band dynamic range compression is used to achieve EQ, then spectral control improves, but device complexity increases
Solution Approach 1:
The patent combines dynamic range control and loudness equalization into a single unified process. The same filter coefficients applied for DRC are also used for EQ, eliminating the need for separate multi-band processing while achieving both spectral control and loudness compensation through metadata-driven parameter adjustment.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Dynamic loudness equalization of received audio content in a playback system, using metadata that includes instantaneous loudness values for the audio content. A playback level is derived from a user volume setting of the playback system, and is compared with a mixing level that is assigned to the audio content. Parameters are computed, that define an equalization filter that is filtering the audio content before driving a speaker with the filtered audio content, based on the instantaneous loudness values and the comparing of the playback level with the assigned mixing level. Other embodiments are also described and claimed.