Audio Signal Analysis via Dynamic Feature Vector Configuration

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing audio metadata systems are inflexible and challenging to configure for new applications, lacking a flexible framework for multi-level signal processing and integration of symbolic machine-learning operations, which hinders adaptive audio analysis and object recognition.

Innovation Solution

A multi-stage audio signal analysis method involving windowed signal analysis, statistical processing, and machine-learning techniques for sound object recognition and labeling, allowing for real-time audio feature extraction and mapping of metadata to multimedia applications.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If fixed feature vector formats are used in existing software implementations, then the system structure is simple and easy to implement, but the system lacks flexibility and is difficult to adapt for new applications

Engineering Contradiction:
Improveadaptability for new applicationsVSAvoidsystem complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent implements dynamic configurability of feature vectors through a parameter-based architecture. The feature extractor is designed to accept configuration parameters that define the structure and content of feature vectors, allowing the system to adapt to different applications by changing parameters rather than rewriting code. This enables the same core system to generate different feature vector formats dynamically based on application requirements.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system uses parameter changes to control feature vector characteristics. By modifying parameters such as window size, feature types, and extraction methods, the system can adapt to new applications without structural changes. The configurable parameters allow flexible adjustment of feature extraction behavior to match specific application needs while maintaining the same underlying system architecture.

Inventive Principle:
Principle #35Parameter changes

2Adaptability or versatility

If second-stage higher-level feature extraction is custom-coded for each application, then the feature extraction is precise for that specific application, but it becomes challenging to develop and configure for new applications

Engineering Contradiction:
Improveconfigurability for new applicationsVSAvoidease of development and configuration
Core Design Contradiction:
Adaptability or versatilityVSEase of manufacture

Solution Approach 1:

The patent creates a universal feature extraction framework that can perform multiple functions through configuration rather than custom coding. The second-stage feature extraction is designed as a configurable module that can be adapted to different applications by setting parameters rather than writing new code. This universal approach allows the same extraction logic to serve multiple applications with different requirements.

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

Solution Approach 2:

The system uses template-based configuration where common feature extraction patterns can be copied and reused across different applications. Instead of custom-coding each extraction process, pre-defined templates can be instantiated with different parameters for new applications, significantly reducing development effort while maintaining application-specific precision.

Inventive Principle:
Principle #26Copying

3Adaptability or versatility

If fixed frameworks supporting only one method of data-mining or application processing are used, then the framework is simple to implement, but it is neither run-time configurable nor easily integrated with various application run-time environments

Engineering Contradiction:
Improverun-time configurability and integration capabilityVSAvoidframework complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent segments the audio processing system into distinct modular components: feature extraction, data mining, and application processing. Each component can be independently configured and replaced, allowing run-time adaptability. The segmentation enables different combinations of components to be used for different applications without redesigning the entire framework.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces an intermediary layer between the processing components and application environments. This intermediary handles configuration management and integration logic, allowing the core processing framework to remain simple while providing run-time configurability and broad integration capability through standardized interfaces.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentUS9031243B2Automatic labeling and control of audio algorithms by audio recognition
Publication Date: 2015.05.12 NATIVE INSTR USA INC
  • US9031243B2 patent drawing
  • US9031243B2 patent drawing
  • US9031243B2 patent drawing

AI summary

Controlling a multimedia software application using high-level metadata features and symbolic object labels derived from an audio source, wherein a first-pass of low-level signal analysis is performed, followed by a stage of statistical and perceptual processing, followed by a symbolic machine-learning or data-mining processing component is disclosed. This multi-stage analysis system delivers high-level metadata features, sound object identifiers, stream labels or other symbolic metadata to the application scripts or programs, which use the data to configure processing chains, or map it to other media. Embodiments of the invention can be incorporated into multimedia content players, musical instruments, recording studio equipment, installed and live sound equipment, broadcast equipment, metadata-generation applications, software-as-a-service applications, search engines, and mobile devices.