Non-Speech Audio Processing Using Mel Frequency Cepstrum Coefficients
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current audio processing techniques are inadequate in handling non-speech sounds, as they primarily focus on identifying words and phrases, making it difficult to process and differentiate other audio file aspects when speech is not present.
Innovation Solution
The development of non-speech audio processing systems and methods that utilize techniques such as Mel Frequency Cepstrum Coefficients, dynamic time warping, and machine learning algorithms to compare and identify non-speech audio files, enabling interactive games and assessments with social bots, such as noise imitation games, where users can guess or imitate sounds like alarm clocks or cellphone ringtones.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If speech processing techniques are used to process audio files, then speech identification accuracy is improved, but non-speech sound processing capability deteriorates
Solution Approach 1:
The audio processing system segments the audio signal into different components and applies specialized processing techniques to each. Speech segments are processed using speech-specific algorithms while non-speech segments are processed using audio analysis techniques, allowing both speech and non-speech sounds to be handled effectively without interference
Solution Approach 2:
The system implements a universal audio processing framework that can handle both speech and non-speech sounds. By integrating multiple processing capabilities into a single system that automatically adapts to the input type, the system achieves versatility without sacrificing the specialized accuracy needed for speech recognition
2Productivity
If traditional audio processing techniques are used, then processing speed is maintained, but non-speech sound identification accuracy deteriorates
Solution Approach 1:
The system performs preliminary classification of audio segments to identify whether they contain speech or non-speech sounds before applying detailed analysis. This preliminary action allows the system to quickly route audio segments to appropriate processing pipelines, maintaining high processing speed while enabling accurate non-speech sound identification through specialized analysis techniques
Solution Approach 2:
The system introduces intermediary processing stages that bridge traditional audio processing and advanced non-speech sound analysis. These intermediary components include feature extraction modules and classification layers that enhance non-speech sound identification accuracy while maintaining compatibility with existing fast processing pipelines
Data Source
AI summary
Systems and methods may be used to process and output information related to a non-speech vocalization, for example from a user attempting to mimic a non-speech sound. A method may include determine a mimic quality value associated with an audio file by comparing a non-speech vocalization to a prerecorded audio file. For example, the method may include determining an edit distance between the non-speech vocalization and the prerecorded audio file. The method may include assigning a mimic quality value to the audio file based on the edit distance. The method may include outputting the mimic quality value.