QMF Subband BRIR Truncation for Real-Time Binaural Rendering

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Binaural rendering for multi-channel audio signals requires high computational complexity due to long filter lengths, especially when using binaural room impulse response filters, which is exacerbated by the need for multiple fast Fourier transforms and inefficient operations in the time domain.

Innovation Solution

The method involves converting binaural room impulse response filter coefficients into subband filter coefficients, determining truncation lengths based on average reverberation time information and curve-fitting, and performing binaural filtering in the QMF domain to reduce computational complexity.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If binaural rendering is performed using long BRIR filter coefficients in the time domain, then sound quality is improved, but computational complexity becomes enormous

Engineering Contradiction:
Improvesound qualityVSAvoidcomputational complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent transforms the BRIR filter coefficients from the time domain to the frequency domain using Fast Fourier Transform (FFT), changing the representation parameters to reduce computational complexity while maintaining sound quality. This allows the system to process long filters efficiently by operating in the frequency domain where convolution becomes multiplication.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent divides the BRIR filter into multiple subbands using Quadrant Mirror Filter (QMF) decomposition, segmenting the frequency spectrum into manageable portions. This segmentation allows parallel processing and reduces the computational burden by handling smaller filter sections independently across different frequency bands.

Inventive Principle:
Principle #1Segmentation

2Productivity

If fast convolution is performed multiple times for multi-channel audio, then processing capability is improved, but computational complexity increases enormously

Engineering Contradiction:
Improveprocessing capabilityVSAvoidcomputational complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent combines multiple channel processing operations into a unified frequency-domain approach. By transforming all input channels to the frequency domain simultaneously and then applying the BRIR filter once across all channels, the system merges what would otherwise be separate convolution operations into a single efficient processing stage, dramatically reducing total computational complexity.

Inventive Principle:
Principle #5Merging (Combining)

3Ease of manufacture

If binaural filtering is performed in the time domain, then implementation is straightforward, but additional QMF synthesis operations are required making it inefficient

Engineering Contradiction:
Improveimplementation simplicityVSAvoidprocessing efficiency
Core Design Contradiction:
Ease of manufactureVSProductivity

Solution Approach 1:

Instead of performing time-domain filtering followed by QMF synthesis, the patent inverts the approach by directly implementing the filter in the frequency domain. This reversal eliminates the need for separate QMF synthesis operations while maintaining the same filtering function, thereby improving processing efficiency without sacrificing implementation clarity.

Inventive Principle:
Principle #13The other way round (Inversion)

Data Source

PatentEP4712075A2Audio signal processing method and audio signal processing device
Publication Date: 2026.03.18 WILUS INSTITUTE OF STANDARDS & TECHNOLOGY INC
  • EP4712075A2 patent drawingFigure 1
  • EP4712075A2 patent drawingFigure 2
  • EP4712075A2 patent drawingFigure 3

AI summary

The present invention relates to an audio signal processing method, a parameterization device and an audio signal processing device for the same, and more particularly, to an audio signal processing method to implement filtering of an input audio signal with a low computational complexity, and a parameterization device and an audio signal processing device for the same. To this end, provided are a method for processing an audio signal, including: receiving an input audio signal; receiving at least one binaural room impulse response (BRIR) filter coefficients for binaural filtering of the input audio signal; converting the BRIR filter coefficients into a plurality of subband filter coefficients; obtaining flag information indicating whether the length of the BRIR filter coefficients in a time domain is more than a predetermined value; truncating each subband filter coefficients based on filter order information obtained by at least partially using characteristic information extracted from the corresponding subband filter coefficients, the truncated subband filter coefficients being filter coefficients of which energy compensation is performed based on the flag information and the length of at least one truncated subband filter coefficients being different from the length of the truncated subband filter coefficients of another subband; and filtering each subband signal of the input audio signal by using the truncated subband filter coefficients, and a parameterization device and an audio signal processing device for the same.