Audio Decorrelator Using Subband Delays for Low-Latency Transients
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing decorrelators for audio signals, such as those used in parametric spatial audio coding, suffer from high computational complexity and processing delay, especially when handling transients.
Innovation Solution
A decorrelator system that divides the frequency representation into parts, processes each part with a separate delay unit, and uses phase shifters and Allpass filters to achieve low processing delay and complexity, allowing for parallel processing and high audio quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If time domain reverberation filters with long impulse response are used, then decorrelation quality is improved, but processing delay and computational complexity increase
Solution Approach 1:
The patent replaces time domain reverberation filters with frequency domain processing using delay units and envelope shapers. Instead of using long impulse response filters that cause high processing delay, the invention uses frequency domain operations with much shorter processing times, achieving comparable decorrelation quality with significantly reduced delay.
Solution Approach 2:
The patent changes the processing domain from time domain to frequency domain, and uses subband decomposition to process different frequency components separately. This parameter change allows for reduced processing delay while maintaining decorrelation effectiveness, as frequency domain operations can be performed with shorter delays than time domain convolution.
2Reliability
If Quadrature Mirror Filterbank with Lattice filters is used, then decorrelation is achieved, but processing delay and computational complexity increase
Solution Approach 1:
The patent divides the frequency representation into multiple subbands and processes each subband independently with separate delay units. This segmentation allows for simpler, less computationally intensive processing compared to using complex QMF banks and Lattice filters, while achieving comparable decorrelation effectiveness through parallel processing of subbands.
Solution Approach 2:
The patent substitutes complex filterbank structures with simpler delay units operating in the frequency domain. Instead of using computationally expensive QMF banks and Lattice filters, the invention employs straightforward delay operations combined with envelope shaping, significantly reducing computational complexity while maintaining decorrelation performance.
3Loss of time
If frequency domain operations with parallel processing are used, then processing delay is reduced, but implementation complexity increases
Solution Approach 1:
The patent segments the frequency spectrum into subbands and assigns separate delay units to each subband. This segmentation enables parallel processing that reduces overall processing delay, while the modular structure actually simplifies implementation compared to monolithic filterbank approaches, as each subband can be processed independently with standard delay elements.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A decorrelator comprises a plurality of delay units, wherein each delay unit is configured for receiving a part of a frequency representation being based on an audio signal, wherein each delay unit is configured for delaying the received part to provide a delayed part. The decorrelator comprises an envelope shaper configured for receiving and combining signals being based on the delayed parts of the frequency representation. The envelope shaper receives the frequency representation of the audio signal and is configured for adjusting an energy of the delayed parts in respect of the frequency representation of the audio signal. The envelope shaper is configured for providing a combined shaped frequency representation. Transient signal portions are handled by an adapted operation of the decorrelator.