Improving Pyramid Vector Quantizer with power projection

2017/05/03 by Jarek Duda, Duda, Jarek · 2 citations
Computer Science · #Advanced Data Compression Techniques #Digital Filter Design and Implementation #FOS: Computer and information sciences #FOS: Mathematics #Image and Signal Denoising Methods #Information Theory (cs.IT) #Optimization and Control (math.OC)

paper · pdf · doi:10.48550/arxiv.1705.05285

openalex publication_date 2017/05/03 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Pyramid Vector Quantizer (PVQ) is a promising technique especially for multimedia data compression, already used in Opus audio codec and considered for AV1 video codec. It quantizes vectors from Euclidean unit sphere by first projecting them to L¹ norm unit sphere, then quantizing and encoding there. This paper shows that the used standard radial projection is suboptimal and proposes to tune its deformations by using parameterized power projection: x→ x^p/‖x^p‖ instead, where the optimized power p is applied coordinate-wise, getting usually ≥ 0.5 dB improvement comparing to radial projection.

Cited by

• Improving distribution and flexible quantization for DCT coefficients
• Encoding of probability distributions for Asymmetric Numeral Systems

Related

• Vector Quantization for Machine Vision
• Pyramid Vector Quantization for LLMs
• Pyramid Vector Quantization for Deep Learning
• Pyramid Vector Quantization and Bit Level Sparsity in Weights for Efficient Neural Networks Inference
• PCDVQ: Enhancing Vector Quantization for Large Language Models via Polar Coordinate Decoupling