Perceptual Delay Evaluation in Teleconferencing via Conversation Pattern Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional delay evaluation methods in audio communication scenarios, such as teleconferencing, fail to accurately reflect the perceived quality of conversation due to their inability to differentiate the impact of delay based on conversation patterns, leading to inadequate quality of experience (QoE) monitoring.
Innovation Solution
A method and system that detect conversational patterns to evaluate perceptual delay impact by determining conversational parameters, such as waiting and response intervals, and using a model to quantify the quality of conversation, providing a more precise measurement of the annoying degree of delay perceived by users.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional delay evaluation methods are used, then the measurement process is simple, but the measurement precision is insufficient to reflect actual user perception
Solution Approach 1:
The patent changes the evaluation parameter from simple absolute delay values to conversational parameters (waiting interval, response interval, interruption ratio) that reflect actual user perception. This allows the system to evaluate delay impact more accurately by considering the conversational context rather than just raw delay measurements.
Solution Approach 2:
The patent introduces a model as an intermediary that maps conversational parameters to perceived delay quality. This model acts as a bridge between the measured conversational parameters and the actual user perception, enabling more accurate evaluation without requiring direct access to user subjective experience.
2Reliability
If absolute delay values are measured, then the monitoring process is straightforward, but the ability to reflect actual conversation quality is insufficient
Solution Approach 1:
The patent segments the conversation into discrete units (speech segments) and identifies specific conversational events (turn-taking, interruptions, overlaps). By analyzing these segmented elements, the system can reliably measure QoE impact of delay on different conversation types without needing to analyze the entire conversation as a single unit.
Solution Approach 2:
The patent transforms the measurement approach from absolute delay values to relative conversational parameters (waiting interval, response interval, interruption ratio) that are more indicative of actual conversation quality. This parameter transformation enables more reliable QoE monitoring by capturing the nuanced impact of delay on different interaction patterns.
3Adaptability or versatility
If simple delay measurement is used, then the implementation is easy, but the adaptability to different conversation scenarios is poor
Solution Approach 1:
The patent implements a dynamic evaluation system that adapts to different conversation scenarios by detecting conversational patterns in real-time. The model adjusts the weight and interpretation of different parameters based on the detected conversation type (e.g., debate vs. presentation), enabling the system to adapt to various scenarios without requiring separate evaluation methods for each case.
Solution Approach 2:
The patent uses multiple conversational parameters (waiting interval, response interval, interruption ratio) that can be adjusted and weighted differently based on the conversation scenario. This parameter-based approach provides versatility across different conversation types while maintaining a unified evaluation framework.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
Example embodiments disclosed herein relate to an evaluation of perceptual delay impact on a conversation. A method of evaluating perceptual delay impact on a conversation is disclosed. The method includes detecting a conversational event between a first participant and a second participant, the conversational event including a plurality of speech segments; in response to the conversational event being detected, obtaining at least one conversational parameter for the first and second participants, the at least one conversational parameter indicating a relationship between the plurality of speech segments in the conversational event; and determining a quality of the conversation according to a model based on the at least one conversational parameter, the model characterizing the quality with the at least one conversational parameter. A method of obtaining a model for evaluating perceptual delay impact on a conversation is also disclosed. Corresponding systems and computer program products are also disclosed.