Speech Signal Processing for Teleconferencing Noise and Continuity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Teleconferencing systems face issues with background noise, transients, and abrupt speech cut-offs, which disrupt the listening experience by causing noise-dominant speech and misinterpreting pauses as complete speaker termination.
Innovation Solution
A method and apparatus that adjust and process speech signals by applying a threshold to remove noise, slowing rising and falling edges to attenuate transients, and holding signal levels during pauses to create a more continuous listening experience, using a control block with threshold, ramp-up, hold, and ramp-down functional blocks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If a simple threshold cutoff is used to process speech signals, then background noise can be minimized, but speech segments are incorrectly cut off during brief pauses
Solution Approach 1:
The system performs preliminary detection of falling edges and applies a holding period before actual cutoff occurs. This preliminary action prevents premature termination of speech segments during brief pauses by anticipating the continuation of speech based on the holding period mechanism.
Solution Approach 2:
The threshold mechanism dynamically adjusts its behavior based on signal characteristics. By detecting falling edges and applying conditional holding periods, the system transitions from a static threshold cutoff to a dynamic response that adapts to speech patterns, maintaining continuity during pauses while still eliminating background noise.
2Reliability
If transients are allowed to pass through the speech signal, then the signal remains authentic, but electrical noise and spikes are heard by listeners
Solution Approach 1:
The system applies different processing characteristics to different portions of the signal. Rising edges are slowed to attenuate transients, while the bulk of the speech signal maintains its authentic characteristics. This local differentiation allows transient reduction without compromising overall signal authenticity.
Solution Approach 2:
The system applies preliminary attenuation to rising edges before they can manifest as harmful transients in the output. By slowing the rate of change at rising edges in advance, the system prevents electrical noise and spikes from occurring, while maintaining the authenticity of the speech content.
3Object-affected harmful factors
If speech signals are abruptly cut off when a user finishes speaking, then background noise is eliminated, but the listening experience becomes unnatural
Solution Approach 1:
The system applies a holding period after detecting falling edges, which acts as a cushion to prevent abrupt cutoff. This preliminary cushioning maintains the speech signal for a brief period after the falling edge, creating a smoother transition that sounds more natural to listeners while still eliminating background noise after the holding period expires.
Data Source
AI summary
A method and apparatus for generating a control signal for processing a speech signal comprising the steps of: adjusting the signal relative to a threshold level; and responsive to detection of a falling edge of the signal, holding the signal level for a holding period. The technique further comprises ‘slowing’ each rising edge of the signal. The technique further comprises attenuating each falling edge of the signal. The steps are carried out on a signal representing the envelope of the speech signal.