Audio Volume Levelling with Smooth Transition Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
State-of-the-art Volume-Levelling (VL) processing technologies introduce audible artefacts when switching between different audio sources due to sudden changes in volume levelling, which can result in shifts in timbre, and incur significant computational overhead from running multiple filters in parallel.
Innovation Solution
Implement a method that smoothly phases-in or phases-out volume levelling processing by gradually adjusting the weighting applied to the audio signal over a predetermined time period, allowing for a seamless transition between requiring and not requiring VL processing, using a combination of filters and a delay/gain module with a weight-and-sum module to adapt filter coefficients and apply frequency-independent gains.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If volume levelling processing is applied to reduce loudness inconsistencies, then loudness consistency is improved, but audible artefacts and timbre shifts are introduced during transitions
Solution Approach 1:
The system dynamically adjusts the transition profile based on audio characteristics and context. The cross-fade duration and weighting factors are not fixed but adapt according to the detected audio content, allowing optimal balance between loudness consistency and artefact avoidance for different programming material.
Solution Approach 2:
The patent modifies key parameters including the transition profile shape, cross-fade duration, and weighting factors applied to different audio components. By changing these parameters dynamically, the system optimizes the balance between maintaining loudness consistency and minimizing audible artefacts during volume levelling transitions.
2Stability of the object's composition
If multiple filters are run in parallel to maintain timbre consistency, then timbre consistency is improved, but computational overhead increases significantly
Solution Approach 1:
The patent combines multiple filter operations into a unified processing pipeline. Instead of running separate filters in parallel, the system integrates the filtering functions and shares common computational resources, reducing the overall computational overhead while maintaining timbre consistency through coordinated filter design.
Solution Approach 2:
The filter bank is designed with multi-functionality where a single set of filters can serve multiple purposes: maintaining timbre consistency, enabling smooth transitions, and adapting to different audio content types. This universal filter design reduces the need for multiple specialized filter chains, thereby reducing computational complexity.
Data Source
Figure 1~2
Figure 3
Figure 4~6
AI summary
A method in a computing device configured to perform volume-levelling processing on input audio data by at least applying one or more filters to the input audio data, the computing device being configured to obtain an estimated difference between a target output loudness level and a loudness level associated with the input audio data, and to adapt the filter coefficients of the one or more filters based on the estimated difference. The method involves starting, or stopping, the volume-levelling processing. The method comprises gradually increasing, or decreasing, a weighting applied to the estimated difference, in response to obtaining an indication to start, or stop, the volume-levelling processing.