Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Rate-Quality or Energy-Quality Pareto Fronts for Adaptive Video Streaming?

Angeliki Katsenou, Xinyi Wang, Daniel Schien, David Bull

TL;DR

The paper challenges rate-only optimizations in adaptive video streaming by introducing energy-quality (EQ) curves and corresponding Pareto fronts, alongside conventional rate-quality (RQ) analysis. It builds three-resolution decoding-energy and VMAF-based quality spaces, derives rate-quality-energy (RQE) and energy-quality (EQ) Pareto fronts via Akima interpolation, and constructs both rate-driven and quality-driven ladders from these fronts. Empirical results on YouTube-UGC using x265 show decoding energy reductions of up to ~31% at comparable QoE, with quality-driven ladders achieving additional bitrate and energy savings. The work highlights the potential of energy-aware ladder design to improve streaming efficiency and motivates future work on including display power and carbon emissions and refining perceived-quality sufficiency.

Abstract

Adaptive video streaming is a key enabler for optimising the delivery of offline encoded video content. The research focus to date has been on optimisation, based solely on rate-quality curves. This paper adds an additional dimension, the energy expenditure, and explores construction of bitrate ladders based on decoding energy-quality curves rather than the conventional rate-quality curves. Pareto fronts are extracted from the rate-quality and energy-quality spaces to select optimal points. Bitrate ladders are constructed from these points using conventional rate-based rules together with a novel quality-based approach. Evaluation on a subset of YouTube-UGC videos encoded with x.265 shows that the energy-quality ladders reduce energy requirements by 28-31% on average at the cost of slightly higher bitrates. The results indicate that optimising based on energy-quality curves rather than rate-quality curves and using quality levels to create the rungs could potentially improve energy efficiency for a comparable quality of experience.

Rate-Quality or Energy-Quality Pareto Fronts for Adaptive Video Streaming?

TL;DR

The paper challenges rate-only optimizations in adaptive video streaming by introducing energy-quality (EQ) curves and corresponding Pareto fronts, alongside conventional rate-quality (RQ) analysis. It builds three-resolution decoding-energy and VMAF-based quality spaces, derives rate-quality-energy (RQE) and energy-quality (EQ) Pareto fronts via Akima interpolation, and constructs both rate-driven and quality-driven ladders from these fronts. Empirical results on YouTube-UGC using x265 show decoding energy reductions of up to ~31% at comparable QoE, with quality-driven ladders achieving additional bitrate and energy savings. The work highlights the potential of energy-aware ladder design to improve streaming efficiency and motivates future work on including display power and carbon emissions and refining perceived-quality sufficiency.

Abstract

Adaptive video streaming is a key enabler for optimising the delivery of offline encoded video content. The research focus to date has been on optimisation, based solely on rate-quality curves. This paper adds an additional dimension, the energy expenditure, and explores construction of bitrate ladders based on decoding energy-quality curves rather than the conventional rate-quality curves. Pareto fronts are extracted from the rate-quality and energy-quality spaces to select optimal points. Bitrate ladders are constructed from these points using conventional rate-based rules together with a novel quality-based approach. Evaluation on a subset of YouTube-UGC videos encoded with x.265 shows that the energy-quality ladders reduce energy requirements by 28-31% on average at the cost of slightly higher bitrates. The results indicate that optimising based on energy-quality curves rather than rate-quality curves and using quality levels to create the rungs could potentially improve energy efficiency for a comparable quality of experience.
Paper Structure (8 sections, 1 equation, 5 figures, 1 table)

This paper contains 8 sections, 1 equation, 5 figures, 1 table.

Table of Contents

  1. Introduction
  2. Test sequences, Video Codec, and Energy Measurements
  3. Exploring Adaptive Streaming based on Quality-Energy curves
  4. Rate-Quality-Energy Parameter Space
  5. Rate-Quality and Energy-Quality Pareto Fronts
  6. Building the RQ and EQ Ladders
  7. Results and Discussion
  8. Conclusions

Figures (5)

  • Figure 1: Overview of the proposed methodology. Black denotes the typical process of pre-processing for adaptive video streaming, green the energy computation as in KatsenouPCS2024, and blue the proposed.
  • Figure 2: The two figures illustrate the quality-rate-energy points for encodes with x.265 across three spatial resolutions: 2160p, 1080p, and 720p. The energy here refers to the energy consumed during the decoding process.
  • Figure 3: This figure summarises the findings in the energy-driven PF computation. The two histograms show the share of different spatial resolutions on the PFs. The bottom two plots illustrate the RQ-PF and EQ-PF for the test sequence "Gaming_2160P-67b0" in the (c) RQ domain and (d) EQ domain.
  • Figure 4: The two top plots illustrate the rate-driven RQ-PF and EQ-PF ladders for the test sequence "Gaming_2160P-67b0" in the RQ and EQ domain, while the two bottom the quality-driven RQ-PF and EQ-PF ladders in the RQ and EQ domain.
  • Figure 5: Mean Ladders with standard error over all videos.