CVSD Noise Feedback Coding for Lower Speech-Band Noise
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing Continuously Variable Slope Delta Modulation (CVSD) codec used in Bluetooth applications suffers from audible distortion due to insufficient stopband attenuation, leading to suboptimal speech quality, and modifying encoding rules or using adaptive post-filters to improve quality would compromise compatibility or introduce distortion.
Innovation Solution
Implementing a noise feedback coding technique with a second-order finite impulse response (FIR) filter in the CVSD codec to shape coding noise, reducing it in the 0-4 kHz passband while increasing it in the 4-32 kHz stopband, which can be filtered out during sampling rate conversion, thereby enhancing speech quality without affecting bit-stream compatibility.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional CVSD encoding is used, then bit rate is low (64 kbit/s), but speech quality is poor due to audible coding noise
Solution Approach 1:
The patent converts the harmful coding noise into a beneficial shaped noise spectrum by using noise feedback filtering. The filter shapes the quantization noise to have reduced power in the speech passband (0-4 kHz) and increased power in the stopband (4-32 kHz), transforming the harmful noise into a less audible form that even provides some masking benefit in the speech band
Solution Approach 2:
The patent implements a noise feedback loop where the quantization noise is filtered and fed back to the quantizer input. This feedback mechanism allows the system to actively control and shape the noise spectrum, reducing noise in critical speech frequencies while maintaining the low bit rate operation
2Manufacturing precision
If adaptive post-filters are used to reduce coding noise, then speech quality improves, but distortion is introduced
Solution Approach 1:
Instead of filtering the speech signal directly (which would cause distortion), the patent shapes the noise spectrum so that the noise itself becomes less harmful. The shaped noise has reduced energy in the speech passband and increased energy in the stopband, converting the harmful noise into a beneficial spectral distribution
3Manufacturing precision
If encoding rules are modified to improve quality, then speech quality improves, but compatibility is compromised
Solution Approach 1:
The patent separates the encoding process into two independent parts: the core CVSD encoding (which maintains compatibility) and the noise feedback filtering (which improves quality). The bit stream format remains unchanged, allowing compatibility with existing decoders, while the encoder includes an additional filtering stage that enhances speech quality
Solution Approach 2:
The noise feedback filter acts as an intermediary between the quantizer and the output, shaping the noise without altering the fundamental CVSD encoding process. This intermediary component improves quality while leaving the bit stream compatible with standard decoders
Data Source
AI summary
Systems and methods are described that apply a noise feedback coding (NFC) technique at the encoder of a delta modulation codec, such as a Continuously Variable Slope Delta Modulation (CVSD) codec, so as to shape the spectrum of the coding noise produced thereby in such a way that the speech quality of the delta modulation decoder output is enhanced. The techniques described herein are not limited to delta modulation codecs and may also be applied to any sample-by-sample codec, including a G.711 μ-law codec, a linear pulse code modulation (LPCM), or any other of a wide variety of sample-by-sample codecs, to improve the audio quality of the decoder output thereof.