Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

An Empirical Study of Reducing AV1 Decoder Complexity and Energy Consumption via Encoder Parameter Tuning

Vibhoothi Vibhoothi, Julien Zouein, Shanker Shreejith, Jean-Baptiste Kempf, Anil Kokaram

TL;DR

This study addresses the high decoding energy of AV1 on battery-powered devices by evaluating encoder-side parameter tuning in libaom-av1 and SVT-AV1. It employs a tool-off analysis and system-level energy measurements using RAPL and Intel SoC Watch to quantify trade-offs between decoding complexity, energy consumption, and compression quality. Key findings show that disabling tools like CDEF in libaom-av1 or using SVT-AV1's fast-decode preset can reduce decoding cycles by about 10–24% with minimal perceptual quality loss, while system-level measurements reveal nuanced energy effects not captured by CPU metrics alone. The work provides actionable guidelines for content providers to generate energy-efficient AV1 streams and highlights the importance of holistic energy profiling, with future work extending to ARM/mobile platforms and further multi-parameter optimization including tiling.

Abstract

The widespread adoption of advanced video codecs such as AV1 is often hindered by their high decoding complexity, posing a challenge for battery-constrained devices. While encoders can be configured to produce bitstreams that are decoder-friendly, estimating the decoding complexity and energy overhead for a given video is non-trivial. In this study, we systematically analyse the impact of disabling various coding tools and adjusting coding parameters in two AV1 encoders, libaom-av1 and SVT-AV1. Using system-level energy measurement tools like RAPL (Running Average Power Limit), Intel SoC Watch (integrated with VTune profiler), we quantify the resulting trade-offs between decoding complexity, energy consumption, and compression efficiency for decoding a bitstream. Our results demonstrate that specific encoder configurations can substantially reduce decoding complexity with minimal perceptual quality degradation. For libaom-av1, disabling CDEF, an in-loop filter gives us a mean reduction in decoding cycles by 10%. For SVT-AV1, using the in-built, fast-decode=2 preset achieves a more substantial 24% reduction in decoding cycles. These findings provide strategies for content providers to lower the energy footprint of AV1 video streaming.

An Empirical Study of Reducing AV1 Decoder Complexity and Energy Consumption via Encoder Parameter Tuning

TL;DR

This study addresses the high decoding energy of AV1 on battery-powered devices by evaluating encoder-side parameter tuning in libaom-av1 and SVT-AV1. It employs a tool-off analysis and system-level energy measurements using RAPL and Intel SoC Watch to quantify trade-offs between decoding complexity, energy consumption, and compression quality. Key findings show that disabling tools like CDEF in libaom-av1 or using SVT-AV1's fast-decode preset can reduce decoding cycles by about 10–24% with minimal perceptual quality loss, while system-level measurements reveal nuanced energy effects not captured by CPU metrics alone. The work provides actionable guidelines for content providers to generate energy-efficient AV1 streams and highlights the importance of holistic energy profiling, with future work extending to ARM/mobile platforms and further multi-parameter optimization including tiling.

Abstract

The widespread adoption of advanced video codecs such as AV1 is often hindered by their high decoding complexity, posing a challenge for battery-constrained devices. While encoders can be configured to produce bitstreams that are decoder-friendly, estimating the decoding complexity and energy overhead for a given video is non-trivial. In this study, we systematically analyse the impact of disabling various coding tools and adjusting coding parameters in two AV1 encoders, libaom-av1 and SVT-AV1. Using system-level energy measurement tools like RAPL (Running Average Power Limit), Intel SoC Watch (integrated with VTune profiler), we quantify the resulting trade-offs between decoding complexity, energy consumption, and compression efficiency for decoding a bitstream. Our results demonstrate that specific encoder configurations can substantially reduce decoding complexity with minimal perceptual quality degradation. For libaom-av1, disabling CDEF, an in-loop filter gives us a mean reduction in decoding cycles by 10%. For SVT-AV1, using the in-built, fast-decode=2 preset achieves a more substantial 24% reduction in decoding cycles. These findings provide strategies for content providers to lower the energy footprint of AV1 video streaming.
Paper Structure (13 sections, 2 figures, 1 table)

This paper contains 13 sections, 2 figures, 1 table.

Table of Contents

  1. Introduction
  2. Background
  3. Energy Estimation Methodology
  4. Running Average Power Limit (RAPL)
  5. Intel SoC Watch
  6. Experimental Setup
  7. Dataset
  8. Encoder and Decoder Setup
  9. Experimental Results
  10. Analysing Overall Decoding Efficiency Gains
  11. Impact of Decoding Duration on Active Power Savings
  12. Practical Encoder Configurations Guidelines for Energy Efficient Decoding
  13. Conclusion

Figures (2)

  • Figure 1: UGC dataset used in the study. The dataset contains a mix of Landscape and Portrait 1080p videos.
  • Figure 2: Heatmaps of Active Power Gain (%) as a Function of Decode Duration for (a) libaom-av1 and (b) SVT-AV1. The colour of each cell represents the power saving (negative values, shown in purple, indicate greater savings) for a specific encoder flag, measured with either RAPL or SoC Watch over an increasing number of decode loops (from 1 to 200). The data illustrates that power gains are often most significant during the initial decode loops and stabilise over time.