Fixed Codebook Pulse Search for Rate-Distortion Audio Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional fixed codebook (FCB) methods for audio codecs, such as Codebook Excited Linear Prediction (CELP), face challenges in optimizing error values and computational resources, as the computational requirements increase exponentially with the number of pulses, leading to unfeasible processing and lower quality signal output.
Innovation Solution
A method for FCB searching in audio codecs that determines the number of FCB pulses based on rate distortion criteria, selects pulses across the time domain, and generates a FCB signal, optimizing error values and computational resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of FCB pulses is increased to reduce signal error, then the signal quality is improved, but the computational resources required increase exponentially
Solution Approach 1:
The patent changes the parameter of pulse selection by introducing a rate-distortion optimization criterion that dynamically determines the optimal number of pulses based on the trade-off between signal quality (distortion reduction) and computational complexity (rate). This allows the system to adaptively select the number of pulses rather than using a fixed number, resolving the contradiction between signal quality and computational resources.
Solution Approach 2:
The patent makes the FCB pulse selection dynamic by allowing the number of pulses to vary based on the signal characteristics and rate-distortion criteria. Instead of a static fixed number of pulses, the system dynamically adjusts the pulse count and positions to optimize the balance between signal quality and computational complexity for each subframe.
2Device complexity
If the number of FCB pulses is decreased to reduce computational resources, then the computational complexity is reduced, but additional error is introduced into the signal
Solution Approach 1:
The patent applies parameter changes by using rate-distortion optimization to dynamically determine the optimal number of pulses. This criterion allows the system to find the minimum number of pulses required to achieve acceptable signal quality, thereby reducing computational resources while minimizing the introduction of error.
Solution Approach 2:
The rate-distortion optimization criterion provides feedback mechanisms that evaluate the impact of pulse selection on signal quality. This feedback allows the system to adjust the number and positions of pulses to maintain acceptable signal quality while minimizing computational complexity.
Data Source
AI summary
Methods and systems are described for encoding voice speech. A method may include receiving, by an audio encoder, an audio signal comprising a plurality of subframes; determining, for a first subframe of the plurality of subframes, a number of fixed codebook (FCB) pulses according to a rate distortion criteria; selecting, in the subframe, a first set of one or more FCB pulses across a time domain and according to the determined number of FCB pulses; and generating a FCB signal based on the selected first set of FCB pulses.