Adaptive ITD Estimation During DTX Speech Pauses
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional DTX systems in communications networks face challenges in accurately adapting the inter-channel time difference (ITD) during speech pauses, leading to sudden changes in spatial characteristics of comfort noise, which can be perceived as annoying by listeners.
Innovation Solution
The method involves speeding up the low-pass filtering of cross-correlation to adapt ITD estimation by modifying the low-pass filter coefficient during speech pauses, using adaptive filtering techniques such as adjusting filter coefficients based on spectral flatness measures and averaging cross-spectra during hangover and pause periods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If conventional low-pass filtering is used for ITD estimation during speech pauses, then the ITD estimation is stable, but the adaptation speed is slow causing sudden changes in spatial characteristics
Solution Approach 1:
The patent applies dynamics by making the low-pass filter coefficient adaptive rather than fixed. The filter coefficient changes dynamically based on the speech activity state: during active speech encoding, one coefficient is used, while during pause periods with CNG, a different coefficient is applied to speed up adaptation. This dynamic adjustment resolves the contradiction between stability and adaptation speed.
Solution Approach 2:
The patent changes the parameter of the low-pass filter coefficient based on the encoding state. By modifying this parameter according to whether the system is in active speech encoding or CNG pause encoding, the patent achieves both stability during speech and faster adaptation during pauses, directly resolving the technical contradiction.
2Loss of energy
If DTX encoding switches from active content to background noise during pause periods, then resource usage is reduced, but spatial characteristics change suddenly causing annoyance
Solution Approach 1:
The patent applies preliminary action by preparing the ITD estimation during the hangover period before the actual pause begins. The low-pass filter coefficient is adjusted in advance during the hangover period, so that when the pause period starts and CNG encoding begins, the ITD estimation is already adapted and ready, preventing sudden spatial characteristic changes.
Solution Approach 2:
The patent uses beforehand cushioning by implementing a hangover period where the ITD estimation is gradually adapted before the full pause effect takes over. This gradual transition during the hangover period cushions the transition, preventing the sudden change that would otherwise occur when switching from active to background noise encoding.
3Speed
If ITD estimation is updated frequently during active encoding, then adaptation to speech changes is fast, but computational complexity increases
Solution Approach 1:
The patent applies local quality by using different ITD estimation update strategies for different time periods. During active speech encoding, the patent uses one approach (faster updates), while during pause periods with CNG, it uses a different approach (slower updates with adjusted filter coefficients). This localized differentiation optimizes the balance between update speed and computational complexity for each specific operating condition.
Data Source
AI summary
A method to estimate an inter-channel time difference (ITD) in an encoder using a discontinuous transmission (DTX) is disclosed. The method includes receiving a time domain audio input including audio input signals and processing the audio input signals in frames to produce a mono mixdown signal and one or more stereo parameters. The method further includes encoding the mono mixdown signal on a frame-by-frame basis by: encoding of active content of the mono mixdown signal at a first bit rate until a pause period is detected; estimating ITD parameters during the encoding of active content; switching the encoding from the active encoding content to inactive encoding to encode background noise at a second bit rate during the pause period; and estimating ITD parameters during the pause period. The method further includes encoding the ITD estimated parameters and other stereo parameters periodically during the pause period.