Name
Performance Comparison of Emerging EVC and VVC Video Coding Standards with HEVC and AV1
Date & Time
Wednesday, November 11, 2020, 7:00 PM - 7:30 PM
Timezone
UTC
Session Type
Technical session
Dan Grois
Description
The recent dramatic increase in video content consumption requires efficient video coding standards, and this is especially true for UltraHD resolutions, such as 4K and 8K (i.e. 3840x2160 or 7680x4320 resolutions in terms of luma samples, respectively). The well-known HEVC (H.265/MPEG-H) video coding standard was approved in 2013, but despite providing approximately 50% coding gain compared to its predecessor AVC (H.264/MPEG-4), the HEVC adoption is still relatively slow. In addition, larger bit-rate savings than those provided by HEVC are currently desired. In turn, the work on the Versatile Video Coding (VVC) and Essential Video Coding (EVC) standards started in 2018, and after intensive development efforts that continued about two and a half years, these two emerging video coding standards have been recently finalized. VVC (H.266/MPEG-I) is developed jointly by the MPEG and ITU-T VCEG organizations, while EVC (MPEG-5) is an MPEG-only effort. In this paper, we compare the coding performance of EVC and VVC, in terms of both coding gains and computational complexity, to their predecessor – the HEVC video coding standard. In addition, given the growing popularity of the AV1 video codec, which was recently developed by the Alliance for Open Media (AOM), we also include AV1 as an alternative baseline and provide corresponding comparison results.

According to the experimental results, which have been carried out in a Constant Bit-Rate (CBR) mode, the EVC provides about 30% bit-rate savings compared to HEVC for encoding 4K/2160p entertainment content (such as VoD) in terms of BD-BR PSNRYUV, while introducing an encoding computational complexity increase of ~5 times. On the other hand, the VVC provides larger bit-rate savings of about 40% at a price of a significant encoding computational complexity increase of more than 9 times. On the other hand, when performance of the HEVC CBR encoding (i.e. with the rate-control disabled) is compared to performance of the AV1 VBR encoding (i.e. with the rate-control is enabled), it was found that AV1 provides bit-rate savings of about 20% compared to HEVC for encoding 4K/2160p video sequences, as a trade-off of the encoding computational complexity increase by a factor of ~4.

The authors find out both EVC and VVC to be very promising successors of HEVC in terms of coding gains and computational complexity, and the jury is still out on the speed of their adoption.

Keywords. video compression, coding efficiency, codec comparison, video coding gain, VVC, EVC, HEVC, AVC, AV1, MPEG, H.266, H.265, H.264, computational complexity.
Technical Depth of Presentation
This talk is appropriate for broad audience. We will start from an overview/introduction such that even those who are not familiar with video codecs still will significantly benefit by attending this talk. In the due course of this talk, we will provide more detailed observations and deep insights, such that professionals/experts in the field will be also able to use the presented information in their daily work practice.
What Attendees will Benefit Most from this Presentation
This talk is appropriate for broad audience, including decision-makers on all technical and non-technical levels, such as this talk will be very beneficial for all attendees: from those not familiar with video codecs to professionals/experts in the video coding field.
Take-Aways from this Presentation
1) There is a strong need for efficient video coding standards, especially for coding of UltraHD resolutions, such as 4K and 8K. 2) EVC and VVC video coding standards are very promising successors of the HEVC standard. 3) Based on the extensive experimentation on 4K UltraHD cinematic and broadcast content, EVC and VVC provide close to 30% and 40% bitrate savings, respectively, compared to HEVC for encoding entertainment content (such as VoD), while introducing a computational complexity increase, especially at the encoder end.