Decoder-Side Loudness Normalization for Accurate DRC Gain Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Dynamic range compression in audio signals during live streaming or recording is challenging when the program loudness is unknown, leading to potential biases in DRC gains and undesirable loudness shifts due to incorrect predictions.
Innovation Solution
Deferring loudness normalization from the encoder side to the decoder side, where integrated loudness is calculated and used to adjust the DRC gains dynamically, allowing for real-time adaptation and improved audio quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If loudness normalization is performed at the encoder side using predicted program loudness, then real-time dynamic range control is achieved, but loudness shifts and pumping artifacts occur when actual program loudness deviates from predictions
Solution Approach 1:
The encoder performs preliminary DRC gain calculation using predicted program loudness before the actual loudness is known. This allows real-time processing to proceed with preliminary gains, while the deferred loudness normalization mechanism later adjusts these gains once actual loudness measurements are available, resolving the contradiction between real-time capability and accuracy.
Solution Approach 2:
The system implements feedback by measuring actual program loudness after encoding and using this information to adjust the DRC gains at the decoder side. This feedback loop corrects the preliminary gains that were calculated using predictions, eliminating loudness shifts and pumping artifacts while maintaining real-time processing benefits.
2Speed
If DRC gains are calculated using predicted program loudness at the encoder, then processing can proceed in real-time, but incorrect predictions lead to biases in DRC gains
Solution Approach 1:
The encoder calculates DRC gains in advance using predicted program loudness to enable real-time processing. The predicted loudness allows the system to proceed with gain application immediately, while the actual loudness measurement is deferred to after encoding completes, thus maintaining processing speed without sacrificing measurement accuracy.
Solution Approach 2:
Instead of waiting for actual loudness measurement before calculating DRC gains (which would slow down processing), the system inverts the approach by first calculating gains using predictions and then adjusting them with actual measurements. This inversion maintains processing speed while ultimately achieving accurate loudness control.
3Reliability
If loudness normalization is deferred to the decoder side, then accurate program loudness can be used for DRC gain adjustment, but additional processing steps are required
Solution Approach 1:
The loudness normalization function is extracted from the encoder side and moved to the decoder side. This allows the encoder to focus on real-time DRC gain calculation using predictions, while the decoder handles the adjustment using actual loudness measurements. The extraction reduces encoder complexity while maintaining or improving overall system accuracy.
Solution Approach 2:
The patent introduces an intermediary mechanism where the encoder outputs preliminary DRC gains based on predictions, and the decoder acts as an intermediary that adjusts these gains using actual loudness measurements. This intermediary approach distributes complexity between encoder and decoder, achieving accurate loudness control without overburdening either component.
4Productivity
If predicted program loudness is used for encoder-side normalization, then real-time streaming is enabled, but loudness shifts occur when actual loudness differs significantly from predictions
Solution Approach 1:
The encoder performs preliminary DRC processing in real-time using predicted loudness values, enabling live streaming without delay. The actual loudness normalization is then applied as a corrective preliminary action at the decoder side, eliminating loudness shifts and pumping effects that would otherwise be generated during real-time processing.
Solution Approach 2:
The system applies preliminary anti-action by first applying DRC gains based on predicted loudness (which may cause harmful effects), then applying a corrective anti-action at the decoder using actual loudness measurements to counteract and eliminate the loudness shifts and pumping artifacts that were preliminarily introduced.
Data Source
AI summary
A bitstream is obtained by a decoder side, that contains an encoded version of an audio signal and an instantaneous loudness sequence of the audio signal. The instantaneous loudness sequence has not been loudness normalized. A dynamic range control, DRC, gain sequence is produced by applying the instantaneous loudness sequence to a DRC characteristic, with loudness normalization. The DRC gain sequence is applied to the decoded audio signal. Other aspects are also described and claimed.