Hierarchical Audio Session Classification Model

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Electronic devices often erroneously upmix audio content, losing artistic intent and introducing spatial and timbre artifacts, particularly when converting stereo music to surround sound or vice versa, and struggle to accurately classify media as movie, music, or voice, which affects audio reproduction settings and multimedia indexing.

Innovation Solution

A hierarchical classification model that uses predetermined criteria to classify audio sessions based on source and employs machine learning analysis of metadata to accurately determine the type of content, reducing processing time and resource usage while increasing accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Area of stationary object

If audio content is upmixed from stereo to surround sound, then spatial coverage is improved, but artistic intent is lost and spatial/timbre artifacts are introduced

Engineering Contradiction:

Improvespatial coverageVSAvoidartistic intent preservation

Core Design Contradiction:

Area of stationary objectVSManufacturing precision

Solution Approach 1:

The system performs preliminary classification of the audio session type (movie, music, voice, gaming) before applying any audio processing. This preliminary identification allows the system to select appropriate processing paths that preserve artistic intent for music while still providing spatial enhancement for movie content, thus resolving the contradiction between spatial coverage and artistic intent preservation

Inventive Principle:

Principle #10Preliminary action

2Measurement precision

If machine learning analysis is used to classify audio sessions, then classification accuracy is improved, but processing time and resource usage increase

Engineering Contradiction:

Improveclassification accuracyVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The classification system is segmented into multiple stages: first using lightweight predetermined criteria (source identification, metadata analysis) for quick initial classification, then applying machine learning analysis only to sessions that require more sophisticated classification. This segmentation reduces overall processing time while maintaining high accuracy for complex cases

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system applies partial machine learning analysis by using it only when predetermined criteria are insufficient, rather than applying full machine learning to all audio sessions. This selective approach maintains high classification accuracy while significantly reducing processing time and resource consumption for straightforward cases

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS20220191636A1Audio session classification

Publication Date: 2022.06.16 HEWLETT PACKARD DEVELOPMENT COMPANY LP

AI summary

Examples of methods for audio session classification are described herein. In some examples, a method may include determining, at a first classification stage, whether an audio session is classifiable with predetermined criteria. In some examples, the method may include classifying, at a second classification stage, the audio session based on a machine learning analysis of metadata in a case that the audio session is not classifiable at the first classification stage.