Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Energy-efficient Adaptive Video Streaming with Latency-Aware Dynamic Resolution Encoding

Vignesh V Menon, Amritha Premkumar, Prajit T Rajendran, Adam Wieckowski, Benjamin Bross, Christian Timmerer, Detlev Marpe

TL;DR

The paper addresses energy-efficient adaptive streaming under encoding latency constraints by introducing LADRE, a latency-aware dynamic per-title encoding framework. LADRE uses spatiotemporal features and random-forest models to predict per-segment resolution and rate factors while enforcing a maximum encoding time $\tau_L$, enabling online optimization without pre-encoding steps. Empirical results on the Inter-4K dataset show LADRE improves XPSNR/PSNR with bitrate savings and dramatically reduces encoding energy (around $84.17\%$) relative to a fixed HLS ladder, with negligible additional streaming latency. The work demonstrates practical potential for greener streaming and establishes a foundation for future latency-constrained, ML-driven encoding optimization.

Abstract

Traditional per-title encoding schemes aim to optimize encoding resolutions to deliver the highest perceptual quality for each representation. However, keeping the encoding time within an acceptable threshold for a smooth user experience is important to reduce the carbon footprint and energy consumption on encoding servers in video streaming applications. Toward this realization, we introduce an encoding latency-a ware dynamic resolution encoding scheme (LADRE) for adaptive video streaming applications. LADRE determines the encoding resolution for each target bitrate by utilizing a random forest-based prediction model for every video segment based on spatiotemporal features and the acceptable target latency. Experimental results show that LADRE achieves an overall average quality improvement of 0.58 dB PSNR and 0.43 dB XPSNR while maintaining the same bitrate, compared to the HTTP Live Streaming (HLS) bitrate ladder encoding of 200 s segments using the VVenC encoder, when the encoding latency for each representation is set to remain below the 200 s threshold. This is accompanied by an 84.17 % reduction in overall encoding energy consumption.

Energy-efficient Adaptive Video Streaming with Latency-Aware Dynamic Resolution Encoding

TL;DR

The paper addresses energy-efficient adaptive streaming under encoding latency constraints by introducing LADRE, a latency-aware dynamic per-title encoding framework. LADRE uses spatiotemporal features and random-forest models to predict per-segment resolution and rate factors while enforcing a maximum encoding time , enabling online optimization without pre-encoding steps. Empirical results on the Inter-4K dataset show LADRE improves XPSNR/PSNR with bitrate savings and dramatically reduces encoding energy (around ) relative to a fixed HLS ladder, with negligible additional streaming latency. The work demonstrates practical potential for greener streaming and establishes a foundation for future latency-constrained, ML-driven encoding optimization.

Abstract

Traditional per-title encoding schemes aim to optimize encoding resolutions to deliver the highest perceptual quality for each representation. However, keeping the encoding time within an acceptable threshold for a smooth user experience is important to reduce the carbon footprint and energy consumption on encoding servers in video streaming applications. Toward this realization, we introduce an encoding latency-a ware dynamic resolution encoding scheme (LADRE) for adaptive video streaming applications. LADRE determines the encoding resolution for each target bitrate by utilizing a random forest-based prediction model for every video segment based on spatiotemporal features and the acceptable target latency. Experimental results show that LADRE achieves an overall average quality improvement of 0.58 dB PSNR and 0.43 dB XPSNR while maintaining the same bitrate, compared to the HTTP Live Streaming (HLS) bitrate ladder encoding of 200 s segments using the VVenC encoder, when the encoding latency for each representation is set to remain below the 200 s threshold. This is accompanied by an 84.17 % reduction in overall encoding energy consumption.
Paper Structure (13 sections, 4 equations, 3 figures, 4 tables)

This paper contains 13 sections, 4 equations, 3 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related work
  3. Latency-aware dynamic resolution encoding
  4. Spatiotemporal complexity feature extraction
  5. Optimized resolution estimation
  6. Optimized rate factor estimation
  7. Constrained variable bitrate encoding
  8. Experimental design
  9. Evaluation setup
  10. Performance metrics
  11. Reference schemes
  12. Evaluation results
  13. Conclusions and future directions

Figures (3)

  • Figure 1: Rate-distortion (RD) and rate-encoding time curves of representative sequences (segments) of Inter-4K dataset inter4k_ref encoded at 540p, 1080p and 2160p resolutions using VVenC at faster preset. Here, XPSNR is used as the quality metric.
  • Figure 2: Encoding using LADRE envisioned in this paper for online streaming applications.
  • Figure 3: RD curves and encoding times of representative video sequences (segments) using default encoding (blue line), OPTE (purple line), LADRE (red line).