Voice Processing Apparatus Speaker Association Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current voice processing systems lack the ability to accurately determine the combination of voice data corresponding to multiple speakers in conversation from recorded voice data, making it difficult for evaluators to identify speaker associations, especially when the data size is large.
Innovation Solution
A voice processing apparatus that includes an acquisition unit to collect voice signals, a detecting unit to calculate signal intensities, and a determining unit to calculate correlation coefficients between signal intensities, determining whether voices are in a conversation state based on specified thresholds, thereby identifying speaker associations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If voice data is continuously recorded to learn communication patterns, then the quantity of voice data increases, but the difficulty of identifying speaker associations increases
Solution Approach 1:
The patent applies preliminary action by calculating correlation coefficients between signal intensities of different voice data in advance. The determining unit computes these correlations proactively before evaluation is needed, storing the results for later use. This allows evaluators to quickly identify speaker associations without manually analyzing large volumes of voice data, resolving the contradiction between data quantity and identification difficulty.
2Measurement precision
If manual evaluation is used to identify speaker associations, then accuracy can be maintained, but the workload and time consumption increase
Solution Approach 1:
The patent implements self-service by enabling the system to automatically determine speaker associations through correlation coefficient calculations. The determining unit autonomously analyzes signal intensity patterns and identifies which voice data corresponds to speakers in conversation, eliminating the need for manual evaluation. This maintains accuracy while dramatically reducing time consumption and workload.
3Measurement precision
If correlation coefficient calculation is performed on all voice data combinations, then identification accuracy improves, but computational complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the voice data into individual data units, each with its own signal intensity time sequence. The determining unit then calculates correlation coefficients between pairs of these segmented units, comparing signal intensity patterns of each voice data against others. This segmented approach maintains identification accuracy while making the computational process more manageable compared to analyzing all data as a single complex unit.
Data Source
AI summary
A voice processing apparatus including a memory, and a processor coupled to the memory and the processor configured to acquire a first input signal containing a first voice, and a second input signal containing a second voice, obtain a first signal intensity of the first input signal, and a second signal intensity of the second input signal, specify a correlation coefficient between a time sequence of the first signal intensity and a time sequence of the second signal intensity, determine whether the first voice and the second voice are in the conversation state or not based on the specified correlation coefficient, and output information indicating an association between the first voice and the second voice when it is determined that the first voice and the second voice are in the conversation state.