Adaptive Comfort Noise Parameters Through Inactive-Segment Averaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing DTX systems face issues with undesired fluctuations in comfort noise (CN) generation during speech pauses, particularly when the first CN parameter set is based on a single frame or a short period, leading to an annoying listening experience due to rapid changes in background noise characteristics.
Innovation Solution
A method for determining CN parameters by calculating a weighted sum of signal characteristics from the current and previous inactive segments, using weighting functions based on the length of the active segment to stabilize the CN generation, especially in cases where background noise conditions change over time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If CN parameters are sent less frequently to save bandwidth, then transmission bandwidth is reduced, but CN stability deteriorates leading to undesired fluctuations
Solution Approach 1:
The system performs preliminary averaging of CN parameters over multiple frames before transmission. Instead of transmitting raw single-frame parameters that cause fluctuations, the encoder averages parameters across several frames (e.g., 8 frames) in advance, so that the transmitted parameters represent a stable baseline that reduces fluctuations when used by the decoder during speech pauses.
2Device complexity
If CN parameters are based on a single frame or short period, then encoding complexity is reduced, but CN quality deteriorates due to rapid changes in background noise
Solution Approach 1:
The system merges multiple frames of CN parameters into a single averaged parameter set for transmission. By combining information from multiple frames (e.g., averaging spectral shape and energy parameters across 8 frames), the system creates a more reliable and stable CN representation that better represents the background noise characteristics, while still maintaining efficient single-frame transmission.
3Adaptability or versatility
If CN parameters are updated frequently to track background noise changes, then adaptability improves, but transmission bandwidth increases
Solution Approach 1:
The system implements periodic transmission of CN parameters at specific intervals (e.g., every 8 speech encoder frames) rather than continuous transmission. This periodic update strategy allows the system to track background noise changes over time while maintaining bandwidth efficiency. The averaged parameters transmitted at these periodic intervals provide stable CN during speech pauses without requiring frequent updates.
Data Source
AI summary
A method for generating a comfort noise (CN) parameter is provided. The method includes receiving an audio input; detecting, with a Voice Activity Detector (VAD), a current inactive segment in the audio input; as a result of detecting, with the VAD, the current inactive segment in the audio input, calculating a CN parameter CNused; and providing the CN parameter CNused to a decoder. The CN parameter CNused is calculated based at least in part on the current inactive segment and a previous inactive segment.


