Scholarlink

Abstract <p>This paper extensively examines the error resilience performance of three prominent HEVC (high-efficiency video coding) software implementations – HM (HEVC Test Model), x265 with medium presets, and x265 with ultrafast presets—across diverse network conditions. Using a class C video coding dataset with varied content, our experiments aim to elucidate the impact of code optimizations on the resilience of HEVC bitstreams to network errors. Contrary to their observed influence on compression efficiency, code optimizations do not significantly affect the error resilience of HEVC bitstreams. Notably, x265-M “Medium” emerges as the most reliable software implementation for real-time video transmission, showcasing adequate performance in the face of network errors. These findings provide valuable insights for practitioners and researchers, informing the selection and design of video coding systems.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Comparative Evaluation of Various HEVC Implementations in Error-Prone Environments

  • 期刊论文

  • 发表日期：

    2025年6月1日

查看全文
  • Islem Mansri,
  • Nasreddine Kouadria,
  • Ahcen Aliouat,
  • Noureddine Doghmane

摘要

Abstract

This paper extensively examines the error resilience performance of three prominent HEVC (high-efficiency video coding) software implementations – HM (HEVC Test Model), x265 with medium presets, and x265 with ultrafast presets—across diverse network conditions. Using a class C video coding dataset with varied content, our experiments aim to elucidate the impact of code optimizations on the resilience of HEVC bitstreams to network errors. Contrary to their observed influence on compression efficiency, code optimizations do not significantly affect the error resilience of HEVC bitstreams. Notably, x265-M “Medium” emerges as the most reliable software implementation for real-time video transmission, showcasing adequate performance in the face of network errors. These findings provide valuable insights for practitioners and researchers, informing the selection and design of video coding systems.