Shared Channel Speaker Profiling for User Count Capacity Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Shared-line telephony systems face challenges in accurately determining the number of distinct users accessing a shared phone line, which affects capacity planning and resource allocation, as endpoint addresses may not directly correspond to individual employees and can represent departments or functional units.
Innovation Solution
Analyzing audio data from shared communication channels to generate speaker profiles, comparing them with a library of existing profiles, and incrementing the count of distinct users when a new speaker is identified, without relying solely on endpoint addresses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If endpoint addresses are used to determine the number of distinct users, then the system can track user access, but the measurement precision is insufficient because endpoint addresses may represent departments or functional units rather than individual employees
Solution Approach 1:
The patent replaces the mechanical system of tracking endpoint addresses with an acoustic field-based speaker recognition system. Audio data captured during calls is analyzed to extract speaker characteristics, which are then compared against a library of known speaker profiles to identify distinct users with higher precision.
Solution Approach 2:
The system changes the identification parameter from endpoint addresses (network-level identifier) to speaker characteristics (acoustic identifier). This parameter change enables more accurate distinction between individual users, as speaker profiles can differentiate between people even when they share the same endpoint address.
2Measurement precision
If audio analysis is performed to identify speakers, then the precision of user identification improves, but the computing resources and processing time increase
Solution Approach 1:
The system performs preliminary action by pre-processing audio data during calls to extract and store speaker characteristics in a library of speaker profiles. This preliminary extraction and storage of acoustic features enables faster comparison and identification during subsequent calls, reducing the computational burden in real-time processing.
3Measurement precision
If speaker profiles are maintained in a library for comparison, then the accuracy of identifying distinct users improves, but the system complexity and storage requirements increase
Solution Approach 1:
The system extracts only the essential acoustic characteristics from audio data to create speaker profiles, rather than storing complete audio recordings or all possible audio features. This extraction of key identifying features reduces the storage requirements while maintaining the ability to accurately distinguish between different speakers.
Data Source
AI summary
This present disclosure provides techniques and solutions for identifying a number of discrete users of a shared communication channel, such as share telephone line associated with a telephone number. Information about the number of discrete users can be used for adjusting computing resource capacity associated with the shared communication channel. A target speaker profile is generated for audio sent over the shared communication channel and compared with speaker profiles in a library. If the target speaker profile does not match any speaker profile in the library, the system increments the number of distinct users associated with the shared communication channel. Disclosed techniques can be applied to various shared communication channels, including shared telephone lines, network addresses, radio frequencies, network links, or network channels.