Layered CELP Encoder Pitch Lag Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing layered CELP speech encoders perform poorly at higher bit rates due to optimization limitations and bit-rate penalties, as they cannot jointly quantize parameters like non-layered encoders, leading to suboptimal performance across various bit rates.
Innovation Solution
A layered CELP encoder architecture with a core layer and enhancement layers, each equipped with first and second adaptive codebooks, allowing for closed-loop pitch lag estimation and separate gains for adaptive and fixed codebook contributions, enabling efficient encoding and decoding across different bit rates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If layered CELP encoder architecture is used with separate adaptive codebooks in core and enhancement layers, then bit rate scalability is improved, but encoding efficiency deteriorates due to inability to jointly quantize parameters
Solution Approach 1:
The enhancement layer adaptive codebook is nested within the core layer adaptive codebook structure, allowing the enhancement layer to refine pitch lag estimates based on the core layer's initial estimation. This nested architecture enables bit rate scalability while maintaining encoding efficiency through hierarchical parameter optimization.
Solution Approach 2:
The enhancement layer uses feedback from the core layer's pitch lag estimation to perform closed-loop optimization. The pitch lag estimate from the core layer is fed back to the enhancement layer adaptive codebook, allowing iterative refinement of pitch parameters and improving overall encoding efficiency despite the layered structure.
2Reliability
If layered CELP encoder with separate gains for adaptive and fixed codebook contributions is used, then performance across different bit rates is improved, but device complexity increases
Solution Approach 1:
The encoder architecture is segmented into core layer and enhancement layer subencoders, each with separate adaptive and fixed codebook contributions. This segmentation allows independent optimization of parameters for different bit rates while maintaining manageable complexity through modular design.
Solution Approach 2:
The system dynamically adjusts which layers and codebook contributions are active based on the target bit rate. The separate gains for adaptive and fixed codebook contributions allow flexible parameter adjustment, enabling the encoder to adapt to different bit rate requirements without requiring a completely different encoder architecture for each rate.
3Productivity
If non-layered encoder jointly quantizes parameters is used, then encoding efficiency is maximized at specific bit rate, but adaptability to multiple bit rates deteriorates
Solution Approach 1:
The encoder is segmented into functional components (core layer subencoder, enhancement layer subencoders) that can be independently configured. Each subencoder handles specific bit rate ranges with optimized parameters, allowing the system to achieve high encoding efficiency at each bit rate while maintaining adaptability across multiple rates through selective activation of layers.
Data Source
AI summary
A layered code-excited linear prediction (CELP) encoder, an Adaptive Multirate Wideband (AMR-WB) encoder and methods of CELP encoding and decoding. In one embodiment, the encoder includes: (1) a core layer subencoder and (2) at least one enhancement layer subencoder, at least one of the core layer subencoder and the enhancement layer subencoder having first and second adaptive codebooks and configured to retrieve a pitch lag estimate from the second adaptive codebook and perform a closed-loop search of the first adaptive codebook based on the pitch lag estimate.