PVQ Indexing With Leading Sign Extraction for 32-Bit DSPs

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing pyramid vector quantization (PVQ) indexing schemes are limited by 32-bit integer ranges and cannot be efficiently implemented in digital signal processor (DSP) architectures that only support signed 32-bit integers, leading to increased complexity and overhead in indexing and de-indexing operations.

Innovation Solution

The proposed method extracts a leading sign from the input vector and uses a pyramid vector quantization enumeration scheme that neglects the sign of the terminal non-zero coefficient, allowing for the use of 33-bit indices in 32-bit DSP hardware and enabling efficient indexing and de-indexing operations by iteratively decomposing PVQ structures into leading sign sections, reducing dynamic range requirements.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If existing PVQ indexing schemes are used, then indexing can be performed, but the schemes are limited by 32-bit integer ranges and cannot be efficiently implemented in DSP architectures that only support signed 32-bit integers, leading to increased complexity and overhead

Engineering Contradiction:
Improvecompatibility with 32-bit DSP hardwareVSAvoidindexing and de-indexing operation complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent segments the PVQ indexing process by separating the sign extraction from the magnitude indexing. The leading sign bit is extracted and handled separately, while the remaining magnitude is indexed using a modified PVQ scheme that operates within 32-bit signed integer constraints. This segmentation allows the system to achieve 33-bit effective indexing capability using only 32-bit hardware operations.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent changes the parameter representation by modifying how the index is structured. Instead of using a straightforward 33-bit unsigned integer index, the system uses a signed 32-bit index where the most significant bit represents the leading sign and the remaining bits represent the magnitude. This parameter transformation enables efficient implementation on 32-bit DSP hardware while maintaining the ability to represent a larger index space.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If 33-bit PVQ indices are used in 32-bit DSP hardware, then indexing capability is improved, but dynamic range requirements and memory requirements increase

Engineering Contradiction:
Improveindexing precisionVSAvoidmemory requirements
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent introduces a new dimensional approach by utilizing the sign bit dimension separately from the magnitude dimension. Instead of requiring a full 33-bit unsigned integer representation, the system uses the most significant bit as a sign indicator and represents the magnitude using the remaining 31 bits. This dimensional reorganization allows efficient packing of 33-bit indexing capability into 32-bit hardware word structures.

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

3Device complexity

If the sign of the terminal non-zero coefficient is neglected in the enumeration scheme, then indexing complexity is reduced, but information loss may occur

Engineering Contradiction:
Improveenumeration scheme complexityVSAvoidsign information
Core Design Contradiction:
Device complexityVSLoss of information

Solution Approach 1:

The patent extracts the leading sign information from the terminal non-zero coefficient and handles it separately from the magnitude enumeration. By taking out the sign bit and processing it independently, the enumeration scheme can focus on magnitudes only, significantly simplifying the counting and indexing logic. The extracted sign information is then recombined with the enumerated magnitude index to form the complete index representation.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentUS10404984B2Method and apparatus for pyramid vector quantization indexing and de-indexing of audio/video sample vectors
Publication Date: 2019.09.03 TELEFONAKTIEBOLAGET LM ERICSSON (PUBL)
  • US10404984B2 patent drawing
  • US10404984B2 patent drawing
  • US10404984B2 patent drawing

AI summary

A method for pyramid vector quantization indexing of audio/video signals comprises obtaining of an integer input vector representing the audio/video signal samples. A leading sign is extracted from the integer input vector. The leading sign is a sign of a terminal non-zero coefficient in the integer input vector. The terminal non-zero coefficient is one of a first non-zero coefficient and a last non-zero coefficient in the integer input vector. The integer input vector is indexed with a pyramid vector quantization enumeration scheme into an output index representing the audio/video signal samples. The pyramid vector quantization enumeration scheme is designed for neglecting the sign of the terminal non-zero coefficient. The output index and the leading sign are outputted. A corresponding method for de-indexing, an encoder, a decoder, and computer programs therefore are also disclosed.