Energy-Rate-Quality Tradeoffs of State-of-the-Art Video Codecs Angeliki Katsenou

2025-04-29 0 0 296.67KB 5 页 10玖币

侵权投诉

Energy-Rate-Quality Tradeoffs of State-of-the-Art

Video Codecs

Angeliki Katsenou

Trinity College Dublin, Dublin, IE

angeliki.katsenou@tcd.ie

Jingwei Mao

University of Bristol, Bristol, UK

jingwei.mao.2020@bristol.ac.uk

Ioannis Mavromatis

BRIL, Toshiba Europe Ltd., Bristol, UK

ioannis.mavromatis@toshiba-bril.com

Abstract—The adoption of video conferencing and video com-

munication services, accelerated by COVID-19, has driven a

rapid increase in video data trafﬁc. The demand for higher

resolutions and quality, the need for immersive video formats,

and the newest, more complex video codecs increase the energy

consumption in data centers and display devices. In this paper, we

explore and compare the energy consumption across optimized

state-of-the-art video codecs, SVT-AV1, VVenC/VVdeC, VP9, and

x.265. Furthermore, we align the energy usage with various

objective quality metrics and the compression performance for

a set of video sequences across different resolutions. The results

indicate that from the tested codecs and conﬁgurations, SVT-

AV1 provides the best tradeoff between energy consumption and

quality. The reported results aim to serve as a guide towards

sustainable video streaming while not compromising the quality

of experience of the end user.

Index Terms—Video Codecs, Energy Consumption,

VVenC/VVdeC, SVT-AV1, VP9, x.265.

I. INTRODUCTION

Over the past years, video network trafﬁc is rapidly increas-

ing and currently accounts for the highest Internet-exchanged

trafﬁc [1]. In addition, the recent COVID-19 pandemic con-

tributed to the rapid adoption of digital online services. As

a result, live and on-demand video exchange becomes the

norm for daily work and leisure activities [2]. Popular exam-

ples are on-demand streaming platforms (Netﬂix, Apple TV,

HBO, Amazon Prime, etc.) and live video conferencing and

collaborative online workspaces (Zoom, Webex, MS Teams,

etc.). Furthermore, the accessibility to powerful and affordable

devices, and the advances in cloud-computing technologies,

enable users to create and share live or on-demand short user-

generated content clips over social media/sharing platforms

(Instagram, TikTok, YouTube, etc.).

Associated with the demands and drivers described above,

the content creation and video communications pipelines

contribute signiﬁcantly to global energy consumption. Cloud

computing services, data centers, display devices, and video

delivery are the main contributors to this increased energy

expenditure. While most climate change organizations focus

on the transport and energy sectors’ emissions, it is essential to

recognize that ICT technologies also generate a considerable

carbon emissions footprint [3]. Hence, efﬁciency must improve

as technology usage increases if sustainability targets are to

be met. According to a recent study by Huawei [4], data

centers currently consume about 3% of global electricity. This,

This work has been supported by Bristol+Bath Creative Industry Cluster.

however, is expected to rise to over 8% by 2030, a ﬁgure

larger than the energy consumption of some nations. While

estimates vary, there is a consensus of an impending major

global issue. Another critical issue is the energy required from

the users to capture, transmit, and display the video data.

Recent research has shown that the energy consumed on the

user side is much higher than on the provider side [5] given

that a single encoding is delivered to thousands of viewers.

While video streaming companies are highly engaged in

optimizing their algorithms to offer the highest quality of

experience, the energy consumption is not part of this process

yet. Each new generation of video codec reduces the amount

of data transmitted over the network at the cost of increased

computational complexity. A ∼50% efﬁciency gain of each

new codec usually comes with a vast increase in computational

complexity [6], [7] yielding to signiﬁcantly increased encod-

ing times. However, decoding has been kept relatively low

complying with the requirement for smooth play-outs without

rebuffering. With this growth in computational load, video

providers, like Netﬂix, BBC, and others, are working towards

assessing the environmental cost and committing towards net

zero emissions [8], [9].

Various research activities have focused on modeling and

predicting the energy consumption at the decoder side [10],

[11]. As a result, tools that analyze the encoding statis-

tics and estimate the decoding energy consumption for

H.264/Advanced Video Coding (AVC) [12], High Efﬁciency

Video Coding (HEVC) [13], and Versatile Video Coding

(VVC) [14] are currently being developed. Similarly, re-

searchers in [7], [15] explore both the encoder and the decoder

on the latest VVC standard and compare it against HEVC.

Challenged by the above, in this work, we investigate the

energy, quality, and bitrate tradeoff across different state-

of-the-art codecs, particularly, x.265, VVenC/VVdeC, VP9,

and Scalable Video Technology AV1 (SVT-AV1). The energy

is measured both at the encoder and the decoder side. We

selected these production-optimized versions of codecs instead

of the reference software implementations as these are usually

deployed by the industry. After collecting the quality, rate,

and energy statistics, we compare their tradeoffs. Although

previous studies have performed codec comparisons in terms

of delivered quality and compression effectiveness [16], [17],

to the best of our knowledge, this is the ﬁrst work comparing

these codecs with regard to their energy-rate-quality tradeoffs.

For the evaluation of the results, a new metric to reﬂect the

arXiv:2210.00618v1 [eess.IV] 2 Oct 2022

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

Energy-Rate-QualityTradeoffsofState-of-the-ArtVideoCodecsAngelikiKatsenouTrinityCollegeDublin,Dublin,IEangeliki.katsenou@tcd.ieJingweiMaoUniversityofBristol,Bristol,UKjingwei.mao.2020@bristol.ac.ukIoannisMavromatisBRIL,ToshibaEuropeLtd.,Bristol,UKioannis.mavromatis@toshiba-bril.comAbstractTheadopti...

展开>> 收起<<

Energy-Rate-Quality Tradeoffs of State-of-the-Art Video Codecs Angeliki Katsenou.pdf

共5页,预览1页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

Energy-Rate-Quality Tradeoffs of State-of-the-Art Video Codecs Angeliki Katsenou

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: