DTX Control for Music Signal Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speech encoders in mobile communication systems incorrectly classify music signals as background noise, leading to unpleasant silences during music transmission and inefficient radio resource usage, as the voice activity detector (VAD) fails to differentiate music from speech effectively.
Innovation Solution
A system that determines an activity factor based on the ratio of sound and silence signal periods in a call input signal to differentiate music from speech, enabling or disabling the discontinuous transmission (DTX) mode only when the activity factor exceeds a specified threshold, typically above 80%, thereby distinguishing music from speech and optimizing radio capacity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If VAD/DTX is enabled to save radio capacity, then radio interference is reduced and radio planning is optimized, but music signals are falsely classified as background noise causing unpleasant silences
Solution Approach 1:
The patent changes the operational parameters of the VAD by introducing a music detection mechanism that analyzes the activity factor and signal characteristics. When music is detected (activity factor > threshold), the VAD is disabled or its threshold is adjusted, preventing false classification. This dynamic parameter adjustment resolves the contradiction by adapting the VAD behavior to the signal type.
Solution Approach 2:
The patent introduces an intermediary music detection mechanism that sits between the signal input and the VAD. This intermediary analyzes the signal characteristics (activity factor, silence frame percentage) and controls whether the VAD should process the signal. This intermediary layer prevents music signals from being incorrectly processed by the VAD while allowing normal speech processing to continue.
2Reliability
If VAD/DTX is disabled for all calls to correctly recognize music signals, then music signals are no longer falsely classified, but radio interference increases and radio capacity is reduced
Solution Approach 1:
The patent applies local quality by making the VAD behavior signal-dependent rather than uniformly applied to all calls. The system locally adjusts VAD operation based on the detected signal type: VAD remains active for speech signals but is disabled or adjusted for music signals. This localized approach ensures music recognition accuracy without sacrificing overall radio capacity.
Solution Approach 2:
The patent introduces dynamics by making the VAD configuration adaptive and changeable during call operation. The system continuously monitors the activity factor and signal characteristics, dynamically switching between VAD-enabled and VAD-disabled states based on whether speech or music is detected. This dynamic adaptation allows the system to optimize both music recognition and radio capacity utilization.
3Reliability
If activity factor threshold is set low to disable DTX for music, then music signals are protected from false classification, but speech signals may also have DTX disabled reducing radio efficiency
Solution Approach 1:
The patent applies partial action by using a combination of multiple criteria (activity factor threshold, silence frame percentage, signal continuity) rather than relying on a single threshold. This partial approach ensures that DTX is disabled for music signals through the cumulative effect of multiple conditions being met, while speech signals that don't fully satisfy all conditions maintain DTX operation, preserving radio efficiency.
Data Source
AI summary
The invention relates to a method for disabling a discontinuous transmission node DTX of a speech encoder if a music signal is detected in a call input signal. The music signal is detected by determining an activity factor corresponding to the relation of sound signal periods relative to scheme signal periods. If the activity factor is higher than a specified activity factor, the DTX is disabled.