Speech Interval Detection Using Dynamic Pitch Gain Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speech production interval detection methods fail to accurately detect the end of a speech interval in audio signals, especially in environments where the pitch parameter is affected by sounds other than the speaker's voice, leading to incorrect detection and prolonged response times.
Innovation Solution
A speech production interval detection apparatus that calculates a pitch gain for each frame of an audio signal, sets a first threshold for detecting the start of speech and a lower second threshold for detecting the end, adjusting the threshold difference based on the difference between representative pitch gain values in the speech and preceding intervals, to improve detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a fixed threshold is used to detect speech production interval end, then the detection process is simple, but the detection accuracy deteriorates in noisy environments
Solution Approach 1:
The patent applies dynamics by making the second threshold variable rather than fixed. The second threshold is dynamically adjusted based on the pitch gain value at the detected start time of the speech production interval. This allows the threshold to adapt to different speech characteristics and environmental conditions, resolving the contradiction between simple detection process and accurate detection in varying environments.
2Measurement precision
If the second threshold is set lower to detect speech end accurately, then detection accuracy improves, but false detection increases due to environmental noise
Solution Approach 1:
The patent applies parameter changes by adjusting the second threshold based on the pitch gain parameter. Instead of using a fixed threshold, the system calculates the pitch gain at the detected start time and sets the second threshold relative to this value. This dynamic parameter adjustment allows the system to maintain high detection accuracy while adapting to different noise levels and speech characteristics, reducing false detections.
3Measurement precision
If manual operation is required to instruct recording start and end, then recording control is precise, but ease of operation deteriorates when hands are occupied
Solution Approach 1:
The patent applies self-service by enabling the system to automatically detect and determine speech production intervals without requiring manual user input. The apparatus uses pitch gain calculation and threshold comparison to autonomously identify when speech starts and ends, allowing hands-free operation while maintaining precise recording control through automatic speech interval detection.
Data Source
AI summary
An apparatus for speech processing: calculates a pitch gain indicating a magnitude of periodicity of an audio signal for each frame, the audio signal representing speaker's voice to be divided into the frames each having a predetermined length; determines that a speech production interval has started, when the pitch gain becomes equal or greater than a first threshold after a non-speech production interval; sets a second threshold that is lower than the first threshold by a reduction amount corresponding to a value acquired by subtracting a second representative value of the pitch gain in an interval preceding the start of the speech production interval from a first representative value of the pitch gain in the speech production interval; and determines that the speech production interval has terminated, when the pitch gain becomes smaller than the second threshold after the speech production interval has started.