Reference-Band Spatial Parameter Signalling for Low-Bitrate Audio

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing spatial audio compression methods face challenges in efficiently managing bitrate allocation between audio channels and spatial metadata, particularly at low bitrates, leading to suboptimal quality and increased metadata overhead.

Innovation Solution

A method that selects a single frequency band with the highest directional energy weight factor to represent all frequency bands, reducing metadata bitrate by encoding only the direction and energy ratios for that band, while maintaining perceptual quality through spatial audio synthesis.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If spatial parameters for all frequency bands are encoded separately, then spatial accuracy is improved, but metadata bitrate increases significantly

Engineering Contradiction:
Improvespatial accuracyVSAvoidmetadata bitrate
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent merges the spatial parameters of multiple frequency bands by selecting a reference frequency band and using its spatial parameters (direction, energy ratios) to represent multiple bands. This combining approach reduces the total number of parameters that need to be encoded and transmitted, directly addressing the contradiction between spatial accuracy and metadata bitrate by maintaining accuracy through reference band selection while reducing bitrate through parameter sharing.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent makes the spatial parameters of the reference frequency band universal by using them to represent multiple frequency bands simultaneously. The selected reference band's parameters serve multiple purposes: they directly represent the reference band and also represent other bands through the bandwidth extension mechanism, thereby reducing overall metadata requirements while maintaining acceptable spatial accuracy across bands.

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Productivity

If metadata bitrate is reduced by encoding only one frequency band, then bitrate efficiency is improved, but spatial information loss increases

Engineering Contradiction:
Improvebitrate efficiencyVSAvoidspatial information loss
Core Design Contradiction:
ProductivityVSLoss of information

Solution Approach 1:

The patent changes the parameter representation by selecting specific parameters (direction, direct-to-total energy ratio, diffuse-to-total energy ratio) from a reference frequency band and using them to represent multiple bands. This parameter selection and reuse strategy improves bitrate efficiency by reducing the number of encoded parameters while minimizing spatial information loss through intelligent reference band selection and parameter extrapolation.

Inventive Principle:
Principle #35Parameter changes

3Productivity

If clustering and quantization are applied to spatial parameters, then compression ratio is improved, but computational complexity increases

Engineering Contradiction:
Improvecompression ratioVSAvoidcomputational complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent segments the frequency spectrum into multiple bands and then selects a reference band for parameter encoding. This segmentation approach allows clustering and quantization to be applied more efficiently by focusing computational resources on one reference band rather than all bands, thereby improving compression ratio while managing computational complexity through selective processing.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3844748B1Spatial parameter signalling
Publication Date: 2025.07.23 NOKIA TECHNOLOGIES OY
  • EP3844748B1 patent drawingFigure 1
  • EP3844748B1 patent drawingFigure 2
  • EP3844748B1 patent drawingFigure 3

AI summary

An apparatus comprising means for:obtaining at least one audio signal;obtaining at least one parameter respectively for each of at least two frequency bands associated with the at least one audio signal; and selecting a frequency band of the at least two frequency bands based on comparing at least one further respective parameter for each of the at least two frequency bands wherein the at least one further respective parameter is determined from each of the at least two frequency bands; generating an output comprising a selection of the at least one parameter associated with the selected frequency band of the at least two frequency bands, such that the selection of the at least one parameter associated with the selected frequency band is configured to reduce a bitrate or size of the output and wherein the at least one parameter of the selected frequency band is configured to represent respective parameters of the at least two frequency bands.