Parametric Audio Stream Mixing Without Full Signal Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio coding techniques lack the ability to efficiently combine multiple parametrically coded audio streams, leading to high computational complexity and delay in scenarios like teleconferencing, where real-time processing is required.
Innovation Solution
An audio signal generator that combines down-mix channels and associated parameters directly within the parameter domain, using simple arithmetic operations such as summation and weight calculations, without reconstructing the individual audio signals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multiple parametrically coded audio streams are combined using conventional decoding and re-encoding techniques, then the combination of audio sources is achieved, but the computational complexity and processing delay increase significantly
Solution Approach 1:
The patent extracts only the essential parametric information (spatial parameters, energy parameters, and downmix signals) from the encoded audio streams, rather than fully decoding the audio signals. This extraction approach allows combining multiple audio sources by manipulating only the parameter data, significantly reducing computational complexity while maintaining the ability to reproduce the combined audio scene.
Solution Approach 2:
The patent performs preliminary combination of parametrically coded audio streams at the encoder side by deriving combined spatial and energy parameters from individual stream parameters before transmission. This preliminary action eliminates the need for complex real-time decoding and re-encoding at the decoder side, enabling real-time processing with minimal computational effort.
2Productivity
If multiple parametrically coded audio streams are combined using conventional techniques, then the audio sources can be integrated, but the processing delay increases
Solution Approach 1:
By extracting and manipulating only parametric data (spatial parameters, energy parameters, and downmix signals) rather than full audio signals, the patent dramatically reduces processing time. The parameter manipulation operations are computationally lightweight compared to full decoding and re-encoding, enabling real-time combination with minimal delay.
Solution Approach 2:
The patent changes the domain of operation from time-domain audio signal processing to parameter-domain processing. By working with spatial parameters, energy parameters, and downmix signals instead of full audio waveforms, the processing operations become much faster and more efficient, reducing processing delay while maintaining audio quality.
3Device complexity
If downmix channels are combined directly in the parameter domain using arithmetic operations, then computational effort is minimized, but the ability to preserve individual channel properties must be maintained
Solution Approach 1:
The patent applies local quality by preserving individual channel properties through separate parameter processing. Each audio source's spatial parameters and energy parameters are processed independently and then combined, ensuring that individual channel characteristics are maintained in the final output. This allows the system to use simple arithmetic operations while still preserving the unique properties of each source.
Solution Approach 2:
The patent uses parameter changes to maintain audio quality by carefully managing the transformation of spatial and energy parameters during combination. The downmix signals are combined using simple arithmetic operations, while the associated parameters are transformed and merged in a way that preserves the perceptual qualities of individual channels, ensuring reliable audio reproduction.
Data Source
AI summary
According to the present invention, multiple parametrically encoded audio signals can be efficiently combined using an audio signal generator, which generates an audio output signal by combining the down-mix channels and the associated parameters of the audio signals directly within the parameter domain, i.e. without reconstructing or decoding the individual input audio signals prior to the generation of the audio output signal. This is achieved by direct mixing of the associated down-mix channels of the individual input signals. It is one key feature of the present invention that the combination of the down-mix channels is achieved by simple, computationally inexpensive arithmetic operations.