CELP Decoder Bandwidth Extension via Up-sampling

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing audio signal processing technologies, such as ITU-T G.718 and G.729.1 compliant speech coders, face significant algorithmic delays when encoding and decoding audio signals with bandwidths beyond the core Code Excited Linear Prediction (CELP) codec's cut-off frequency, due to the need for sequential processing and analysis in the spectral domain using Modified Discrete Cosine Transform (MDCT).

Innovation Solution

A method for decoding audio signals that involves up-sampling the CELP excitation signal, applying non-linear operations, and using bandpass filtering to extend the audio bandwidth beyond the CELP-based decoder element, while synchronizing adaptive codebooks and minimizing alignment errors, and employing complementary all-pass filters to achieve equivalent filtering in both the encoder and decoder.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If MDCT analysis is used to code bandwidth beyond CELP core cut-off frequency, then coding capability for wideband/super-wideband signals is improved, but algorithmic delay increases significantly

Engineering Contradiction:
Improvecoding capabilityVSAvoidalgorithmic delay
Core Design Contradiction:
Adaptability or versatilityVSLoss of time

Solution Approach 1:

The audio signal is divided into two separate coding paths: CELP coding for the low-frequency band (0-6.4 kHz) and MDCT-based spectral coding for the high-frequency band (6.4-14 kHz). By segmenting the frequency spectrum and applying different coding methods to each segment, the patent achieves wideband/super-wideband coding capability while managing delay through parallel processing of the two bands

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transitions from time-domain sequential processing to frequency-domain parallel processing by applying MDCT to the high-frequency band separately. This dimensional change allows simultaneous encoding of low and high frequency components, reducing the cumulative algorithmic delay that would result from sequential processing

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Adaptability or versatility

If sequential CELP and MDCT processing is used, then bandwidth extension is achieved, but processing complexity and delay increase

Engineering Contradiction:
Improvebandwidth extensionVSAvoidprocessing complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The processing architecture is segmented into independent CELP and MDCT processing streams that operate in parallel rather than sequentially. The input signal is split into low-frequency (for CELP) and high-frequency (for MDCT) components, processed simultaneously through different algorithms, and then combined at the output, reducing overall processing complexity

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies preliminary frequency-domain decomposition using MDCT to the high-frequency band before encoding, allowing the high-frequency components to be processed independently and in parallel with the CELP low-frequency encoding, rather than requiring sequential time-domain processing

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP2628156B1Audio signal bandwidth extension in CELP-based speech coder
Publication Date: 2015.09.02 GOOGLE TECHNOLOGY HOLDINGS LLC
  • EP2628156B1 patent drawingFigure 1A
  • EP2628156B1 patent drawingFigure 1B
  • EP2628156B1 patent drawingFigure 2

AI summary

A method for decoding an audio signal in a decoder having a CELP-based decoder element including a fixed codebook component, at least one pitch period value, and a first decoder output, wherein a bandwidth of the audio signal extends beyond a bandwidth of the CELP-based decoder element. The method includes obtaining an up-sampled fixed codebook signal by up-sampling the fixed codebook component to a higher sample rate, obtaining an up-sampled excitation signal based on the up-sampled fixed codebook signal and an up-sampled pitch period value, and obtaining a composite output signal based on the up-sampled excitation signal and an output signal of the CELP-based decoder element, wherein the composite output signal includes a bandwidth portion that extends beyond a bandwidth of the CELP-based decoder element.