Speech Packet Identifier Encoding for Bandwidth Adaptation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current speech processing technologies face challenges in efficiently encoding and decoding speech signals across different bandwidths, particularly in combining narrowband and highband signals without significant computational complexity and maintaining low sampling rates, while also supporting compatibility with narrowband systems.

Innovation Solution

The method involves a system that includes a wideband speech encoder and decoder capable of partitioning signals into frames, determining whether to encode as wideband or narrowband packets, and embedding an identifier to select appropriate decoding modes, allowing for efficient transmission and decoding of speech signals by combining encoded narrowband and highband signals into a single bitstream.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If speech signals are encoded as wideband packets, then speech quality is improved, but device complexity increases

Engineering Contradiction:
Improvespeech qualityVSAvoiddevice complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system dynamically selects between wideband and narrowband encoding modes based on network conditions and capabilities. The encoder can switch between encoding the full wideband signal and encoding only the highband portion, allowing adaptation to different operational requirements and reducing complexity when full wideband encoding is not necessary.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The speech signal is divided into lowband and highband portions. Instead of always encoding the entire wideband signal, the system can segment the encoding process to only process the highband portion when operating in half-rate mode, reducing computational complexity while maintaining speech quality improvements.

Inventive Principle:
Principle #1Segmentation

2Measurement precision

If identifier bits are added to packets, then decoding accuracy is improved, but loss of information increases

Engineering Contradiction:
Improvedecoding accuracyVSAvoidloss of information
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The identifier serves multiple functions: it indicates the encoding mode (wideband or narrowband), guides the decoder in selecting the appropriate decoding algorithm, and enables proper reconstruction of the speech signal. This multi-functionality justifies the additional bits by providing comprehensive control over the decoding process.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The identifier is embedded in the packet during the encoding phase, before transmission. This preliminary action ensures that the decoder has all necessary information available at the start of the decoding process, enabling accurate mode selection and preventing information loss that would occur if mode information had to be determined later.

Inventive Principle:
Principle #10Preliminary action

3Use of energy by moving object

If sampling rate is reduced, then use of energy is improved, but measurement precision deteriorates

Engineering Contradiction:
Improveuse of energyVSAvoidmeasurement precision
Core Design Contradiction:
Use of energy by moving objectVSMeasurement precision

Solution Approach 1:

The speech signal is segmented into lowband and highband frequency ranges. By reducing the sampling rate and only encoding the highband portion in half-rate mode, the system reduces energy consumption while preserving the perceptually important high-frequency components that distinguish wideband from narrowband speech.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different portions of the speech signal are treated differently. The lowband portion is fully encoded to maintain basic speech intelligibility, while the highband portion is selectively encoded based on energy availability, optimizing the balance between energy consumption and speech quality.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP2047461B1Systems and methods for including an identifier with a packet associated with a speech signal
Publication Date: 2013.04.17 QUALCOMM INC
  • EP2047461B1 patent drawingFigure 1a~1b
  • EP2047461B1 patent drawingFigure 2a~2b
  • EP2047461B1 patent drawingFigure 3a~3b

AI summary

A method for including an identifier with a packet associated with a speech signal is described. A signal is received. The signal is partitioned into a plurality of frames. A frame of the signal is encoded into a packet. A determination is made if the packet is encoded as a wideband packet or a narrowband packet. An identifier is packed in the packet based on the determination. The packet is transmitted. At least two illegal values are provided from an N-bit parameter, wherein at least one bit from the N-bit parameter is used to carry information. A number of bits from the N-bit parameter that are used to carry information is equal to log2(X), wherein X is the number of illegal values provided from the N-bit parameter.