Pitch-Independent Timbre Extraction for Audio Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio processing technologies struggle to extract timbre attributes independently of pitch, requiring large databases of pitch-dependent timbres and consuming significant resources for classification and identification.
Innovation Solution
The method involves extracting a pitch-independent timbre log-spectrum from audio signals using a constant Q transform (CQT) to determine a log-spectrum, which is then processed to isolate timbre-independent attributes, allowing for efficient classification and identification of audio sources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If pitch-dependent timbre databases are used for audio classification, then classification accuracy is maintained, but resource requirements and system complexity increase significantly
Solution Approach 1:
The patent extracts pitch as a separate variable from timbre representation. By removing pitch dependency from timbre databases, the system stores only pitch-independent timbre characteristics, dramatically reducing database size while maintaining classification accuracy through separate pitch tracking
Solution Approach 2:
The audio analysis is segmented into two independent components: pitch extraction and timbre extraction. This segmentation allows the system to handle pitch and timbre separately, enabling the use of a compact pitch-independent timbre database while still achieving accurate instrument classification
2Productivity
If pitch-independent timbre extraction is implemented, then resource requirements are reduced, but the difficulty of detecting and measuring timbre attributes increases
Solution Approach 1:
The patent applies preliminary pitch detection and removal before timbre analysis. By pre-processing the audio signal to eliminate pitch components, the subsequent timbre extraction becomes simpler and more efficient, as it only needs to analyze the remaining pitch-independent characteristics
Data Source
AI summary
Methods and apparatus to extract a pitch-independent timbre attribute from a media signal are disclosed. An example apparatus includes an audio characteristic extractor to determine a logarithmic spectrum of an audio signal; transform the logarithmic spectrum of the audio signal into a frequency domain to generate a transform output; determine a magnitude of the transform output; and determine a timbre attribute of the audio signal based on an inverse transform of the magnitude.