Shared Speakerphone Switching for Active Speaker Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In videoconferencing scenarios where multiple devices are used in a physical location, such as a conference room, issues arise from cross-talk and interference among different participant devices, and identifying the active speaker can be challenging.
Innovation Solution
A speakerphone system is shared among multiple devices, using talker identification (talker ID) and active speaker recognition to identify the active speaker, automatically switching to the device with the best microphone, and displaying the active speaker's video feed to all participants.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple participant devices are used in a physical location for videoconferencing, then participants can join the meeting through their individual devices with microphones and speakers, but cross-talk and interference occur between different devices' audio equipment
Solution Approach 1:
The system segments the audio functions by designating one device as the primary audio source while other devices focus on video transmission. The conference chair device is responsible for capturing audio through its microphone, while participant devices primarily transmit video feeds, thereby separating audio capture functions from video capture functions across multiple devices.
Solution Approach 2:
The conference chair device acts as an intermediary between the environment and other participant devices. It receives audio input from the physical environment through its microphone, processes the audio feed, and transmits it to the conferencing service, which then distributes it to all participants. This intermediary role prevents direct audio interference between multiple device microphones and speakers.
2Quantity of substance
If multiple devices capture audio simultaneously, then each device can independently pick up participant speech, but it becomes difficult to identify which participant is actually speaking
Solution Approach 1:
The system performs preliminary identification of the active speaker by analyzing video feeds from multiple participant devices before audio capture becomes critical. The conference chair device or conferencing service examines video data to determine which participant is actively speaking, then designates that participant's device as the primary audio source, ensuring accurate speaker identification before audio routing decisions are made.
Solution Approach 2:
The system implements feedback mechanisms where the conferencing service continuously monitors video feeds and audio inputs to identify the active speaker. Based on this feedback, the system dynamically adjusts which device's microphone is used for audio capture and which device's video feed is displayed to other participants, ensuring the correct speaker is always identified and highlighted.
3Ease of operation
If each participant device uses its own speakers for audio output, then audio can be provided locally to each participant, but cross-talk and interference occur between multiple speakers
Solution Approach 1:
The system extracts the audio output function from individual participant devices and consolidates it at the conferencing service level. The conference chair device receives the audio feed from the conferencing service through its microphone input, processes it locally, and outputs it through its speaker. This extraction of the audio processing function from multiple devices and concentration at a single device eliminates speaker interference while maintaining local audio output capability.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A speakerphone system is shared with multiple participant devices of participants in a physical meeting that are using a web conferencing service. An active speaker is identified from the participants. The participant device of the active speaker is switched, such that the speakerphone system receives and renders audio of the active speaker. Video of the participant device of the active speaker is enabled, such that the web conferencing service displays the video to the participant devices.