Audio Codec Loudness Metadata Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional audio leveler devices cause loudness errors and artifacts, such as audio pumping, due to divergence between measured and target loudness levels, especially at the beginning of audio streams, leading to undesirable audio effects and inability to reproduce original audio levels.
Innovation Solution
An audio codec system that encodes audio content with loudness metadata, allowing a decoder to adjust loudness levels based on source and target loudness, reducing errors by utilizing lookahead and providing continuous loudness measurement updates, thus eliminating pumping artifacts and additional delay.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional audio leveler devices adjust loudness in real-time, then loudness normalization is achieved, but loudness errors and artifacts (such as audio pumping) occur due to divergence between measured and target loudness levels
Solution Approach 1:
The encoder performs loudness measurement and metadata generation in advance during the encoding process, rather than adjusting loudness in real-time during playback. This preliminary measurement allows the system to capture accurate source loudness values without causing audio pumping artifacts, as the measurement is performed on the original audio signal before any processing that might cause instability.
Solution Approach 2:
The patent introduces loudness metadata as an intermediary element that carries measurement information from the encoder to the decoder. This metadata acts as a mediator that enables the decoder to perform accurate loudness normalization without requiring the encoder to perform real-time adjustments that would cause audio quality degradation.
2Measurement precision
If loudness measurement is performed on the entire audio signal, then accurate overall loudness is obtained, but the measurement cannot be completed for live or real-time streaming where the entire audio is not yet received
Solution Approach 1:
For live streaming, the encoder performs preliminary loudness measurements on the audio portion that has been received so far, generating intermediate loudness metadata. This allows real-time processing to proceed without waiting for the complete audio signal, while still providing useful loudness information for current playback.
Solution Approach 2:
The system implements continuous loudness measurement updates where the encoder periodically measures loudness on received audio portions and updates the metadata stream. This feedback mechanism allows the loudness measurement to converge toward the overall loudness as more audio is received, providing progressively more accurate normalization while maintaining real-time capability.
3Manufacturing precision
If loudness adjustment is performed at the encoder side, then loudness normalization is achieved, but the original audio signal quality cannot be reproduced and additional processing delay is introduced
Solution Approach 1:
The patent extracts only the loudness measurement information from the audio signal and transmits it as separate metadata, rather than modifying the audio signal itself. This extraction approach allows the original audio signal to be preserved in its entirety during encoding, with only the essential loudness characteristics being processed and transmitted separately for later normalization at the decoder.
Solution Approach 2:
The encoder performs preliminary loudness measurement and stores the results in metadata without applying any gain adjustments to the audio signal. This preliminary measurement phase preserves the original audio fidelity, while the actual normalization is deferred to the decoder side where the audio signal can be adjusted based on the pre-computed metadata without introducing encoder-side processing delays.
Data Source
AI summary
A method that includes receiving an audio component associated with an audio scene, the audio component including an audio signal, determining a loudness level of the audio component based on the audio signal, receiving a target loudness level for the audio component, producing a bitstream with the audio component by encoding the audio signal and including metadata that has the loudness level and the target loudness level, and transmitting the bitstream to an electronic device.