Video Area Home >> Video Codecs Comparisons >> MSU Video Codecs Comparison 2023-2024

MSU Video Codecs Comparison 2023-2024
Part 3: Hardware encoders

Eighteenth Annual Video-Codecs Comparison by MSU

Video group head:	Dr. Dmitriy Vatolin
Project head:	Dr. Dmitriy Kulikov
Measurements, analysis:	Nickolay Safonov, Aleksandr Kostychev, Anna Bigler

compression.ru

Lomonosov
Moscow State University (MSU)
Graphics and Media Lab

Dubna International
State University

MSU Institute of Advanced
Studies of Artificial Intelligence
and Intelligent Systems

News

28.08.2024 Report release

Description

In MSU Hardware Video Codecs Comparison 2023-2024, we analyzed 15 hardware codecs. There were 4 different speed usecases: 30 fps, 60 fps, 120 fps, 240 fps. 53 FullHD video sequences were used for testing.

Results

The places below are given by quality scores
Encoders with scores closer than ~1% share one place

Fast (30 fps)

Very-fast (60 fps)

Super-fast (120 fps)

Ultra-fast (240 fps)

Best quality
(YUV-SSIM 6:1:1)

	Tencent canghaiV2 (FPGA)
	Streamlake-200 (ASIC), BVE1.2 (ASIC)
	Reference x265 (CPU)

	Tencent canghaiV2 (FPGA)
	Streamlake-200 (ASIC), BVE1.2 (ASIC)
	Reference x265 (CPU)

	Tencent canghaiV2 (FPGA)
	BVE1.2 (ASIC), Streamlake-200 (ASIC)
	Netint Quadra T1A H.265 (ASIC)

	BVE1.2 (ASIC)
	Streamlake-200 (ASIC)
	NVENC AV1 (GPU)

Best quality
(YUV-PSNR avg.MSE 6:1:1)

	Tencent canghaiV2 (FPGA)
	BVE1.2 (ASIC)
	Streamlake-200 (ASIC)

	Tencent canghaiV2 (FPGA)
	BVE1.2 (ASIC)
	Streamlake-200 (ASIC)

	Tencent canghaiV2 (FPGA)
	BVE1.2 (ASIC)
	Streamlake-200 (ASIC)

	BVE1.2 (ASIC)
	Streamlake-200 (ASIC)
	NVENC AV1 (GPU)

Best quality
(Y-VMAF 0.6.1)

	Tencent canghaiV2 (FPGA)
	BVE1.2 (ASIC)
	Streamlake-200 (ASIC)

	Tencent canghaiV2 (FPGA)
	BVE1.2 (ASIC)
	Streamlake-200 (ASIC)

	Tencent canghaiV2 (FPGA)
	BVE1.2 (ASIC)
	Streamlake-200 (ASIC)

	BVE1.2 (ASIC)
	Streamlake-200 (ASIC)
	NVENC AV1 (GPU)

Best quality
(Y-VMAF-NEG 0.6.1)

	Tencent canghaiV2 (FPGA)
	BVE1.2 (ASIC)
	Streamlake-200 (ASIC)

	Tencent canghaiV2 (FPGA)
	BVE1.2 (ASIC)
	Streamlake-200 (ASIC)

	Tencent canghaiV2 (FPGA)
	BVE1.2 (ASIC)
	Streamlake-200 (ASIC)

	BVE1.2 (ASIC)
	Streamlake-200 (ASIC)
	NVENC AV1 (GPU)

Best* GPU encoder

Intel ARC H.265 (GPU)

Intel ARC H.265 (GPU)

Intel ARC H.265 (GPU)

NVENC AV1 (GPU)

Best* FPGA encoder

Tencent canghaiV2 (FPGA)

Tencent canghaiV2 (FPGA)

Tencent canghaiV2 (FPGA)

Best* ASIC encoder

BVE1.2 (ASIC)

BVE1.2 (ASIC)

BVE1.2 (ASIC)

BVE1.2 (ASIC)

* The encoder that is the best in the largest number of main metrics

The winners vary for different objective quality metrics. The participants were rated using BSQ-rate (enhanced BD-rate) scores [1].

[1] A. Zvezdakova, D. Kulikov, S. Zvezdakov, D. Vatolin, "BSQ-rate: a new approach for video-codec performance comparison and drawbacks of current solutions," 2020.

Download and buy report

Free

Enterprise

15 hardware-accelerated encoders
FPGA: Tencent canghaiV2
ASIC: Streamlake-200, BVE1.2, Netint Quadra T1A AV1, Netint Quadra T1A H.265, Netint Quadra T1A H.264
CPU: Reference x265 , Reference x264, Intel QSV H265, Intel QSV H264
GPU: NVENC AV1, NVENC H.265, Intel QSV AV1, Intel QSV H265, Intel QSV H264, AMD H.265, AMD H.264

Number of test sequences

5 FullHD videos
(120 fps usecase)

53 FullHD videos
(30, 60, 120 and 240 fps usecases)

Test video descriptions

Basic codec info

Objective metrics

Only 4 metrics

28+ objective metrics
(VMAF, SSIM, MS-SSIM, PSNR of different variants)

Test videos download

Encoders presets description

HTML report

80 interactive charts

5900+ interacive charts

Price

Free

950 USD

Download/Buy

Objective comparison of ultra-fast and hardware-accelerated video encoders

HTML report (ZIP)

You will receive enterprise versions of all 2023-2024 reports (FullHD, Subjective, 4K)

Participants

	Codec name	Platform	Standard	Usecases	Presets provided
1	Reference x264 MulticoreWare, Inc.	CPU Intel Core i7 12700K (Alder Lake)	H.264/AVC	Ultra-fast (240 fps)	by MSU team
2	Reference x265 MulticoreWare, Inc.	CPU Intel Core i7 12700K (Alder Lake)	H.265/HEVC	Fast (30 fps), Very-fast (60 fps), Super-fast (120 fps)	by MSU team
3	Tencent canghaiV2 Tencent	FPGA Xilinx Alveo U250	H.266/VVC	Fast (30 fps), Very-fast (60 fps), Super-fast (120 fps) Ultra-fast (240 fps)	by developer
4	Streamlake-200 Streamlake	ASIC	H.265/HEVC	Fast (30 fps), Very-fast (60 fps), Super-fast (120 fps) Ultra-fast (240 fps)	by developer
5	BVE1.2 Bytedance	ASIC	H.265/HEVC	Fast (30 fps), Very-fast (60 fps), Super-fast (120 fps) Ultra-fast (240 fps)	by developer
6	Netint Quadra T1A AV1 Netint	ASIC Netint Quadra	AV1	Fast (30 fps), Very-fast (60 fps), Super-fast (120 fps) Ultra-fast (240 fps)	by MSU team (MSU team comment) by MSU team: We have selected presets based on provided with installation guide report Quadra_Video_Quality_Report_V4.7.*.pdf
7	Netint Quadra T1A H.265 Netint	ASIC Netint Quadra	H.265/HEVC	Fast (30 fps), Very-fast (60 fps), Super-fast (120 fps) Ultra-fast (240 fps)	by MSU team (MSU team comment) by MSU team: We have selected presets based on provided with installation guide report Quadra_Video_Quality_Report_V4.7.*.pdf
8	Netint Quadra T1A H.264 Netint	ASIC Netint Quadra	H.264/AVC	Fast (30 fps), Very-fast (60 fps), Super-fast (120 fps) Ultra-fast (240 fps)	by MSU team (MSU team comment) by MSU team: We have selected presets based on provided with installation guide report Quadra_Video_Quality_Report_V4.7.*.pdf
9	NVENC AV1 Nvidia Corp.	GPU Nvidia RTX 4070TI	AV1	Fast (30 fps), Very-fast (60 fps), Super-fast (120 fps) Ultra-fast (240 fps)	by MSU team
10	NVENC H.265 Nvidia Corp.	GPU Nvidia RTX 4070TI	H.265/HEVC	Fast (30 fps), Very-fast (60 fps), Super-fast (120 fps) Ultra-fast (240 fps)	by MSU team
11	Intel QSV AV1 Intel Corp.	GPU Intel Arc A380	AV1	Fast (30 fps), Very-fast (60 fps), Super-fast (120 fps) Ultra-fast (240 fps)	by MSU team
12	Intel QSV H.265 Intel Corp.	GPU Intel Arc A380	H.265/HEVC	Fast (30 fps), Very-fast (60 fps), Super-fast (120 fps) Ultra-fast (240 fps)	by MSU team
13	Intel QSV H.264 Intel Corp.	GPU Intel Arc A380	H.264/AVC	Fast (30 fps), Very-fast (60 fps), Super-fast (120 fps) Ultra-fast (240 fps)	by MSU team
14	Intel QSV H.265 Intel Corp.	CPU (built-in hardware acceleration)	H.265/HEVC	Fast (30 fps), Very-fast (60 fps), Super-fast (120 fps) Ultra-fast (240 fps)	by MSU team
15	Intel QSV H.264 Intel Corp.	CPU (built-in hardware acceleration)	H.264/AVC	Fast (30 fps), Very-fast (60 fps), Super-fast (120 fps) Ultra-fast (240 fps)	by MSU team
16	AMD H.265 AMD	GPU AMD Radeon RX 6800 XT	H.265/HEVC	Fast (30 fps), Very-fast (60 fps), Super-fast (120 fps) Ultra-fast (240 fps)	by MSU team
17	AMD H.264 AMD	GPU AMD Radeon RX 6800 XT	H.264/AVC	Fast (30 fps), Very-fast (60 fps), Super-fast (120 fps) Ultra-fast (240 fps)	by MSU team

Comparison Rules

Hardware-accelerated and ultra-fast codecs testing objectives

The main goal of this report is the presentation of a comparative evaluation of the quality of new and existing codecs using objective measures of assessment. The comparison was done using settings provided by the developers of each codec. Nevertheless, we required all presets to satisfy minimum speed requirement on the particular use case. The main task of the comparison is to analyze different encoders for the task of transcoding video – e.g., compressing video for personal use.

Test Hardware Characteristics for GPU/ASIC encoders

CPU: Intel Core i7 12700K (Alder Lake)
SSD: 1Tb
RAM: 4x16GB (64GB)
OS: Windows 11 x64, Ubuntu 22.04 LTS

Test Hardware Characteristics for FPGA encoders

CPU: Intel Core i7 12700K (Alder Lake)
SSD: 1Tb
FPGA: Xilinx Alveo U250 Accelerator
RAM: 4x16GB (64GB)
OS: Windows 11 x64, Ubuntu 22.04 LTS

See more on Call For Codecs 2023-2024 page

Videos

Videos for testing set were chosen from MSU video collection via a voting among comparison participants, organizers and an independend expert.

Number of videos in MSU video collection

Year	# FullHD videos	# FullHD samples	# 4K videos	# 4K samples	Total # of videos	Total # of samples
2016	3	7	882	2902	885	2909
2017	1996	4638	1544	4561	3540	9299
2018	4342	10330	1946	5503	6288	15833
2020	4945	12402	2091	6016	7036	18418
2021	4945	12402	2091	6016	7036	18418
2022	7379	19546	2091	6016	9470	25562

Bitrate distribution of videos in MSU video collection Videos bitrate distribution

Final video set consists of 53 sequences including new videos from Vimeo and media.xiph.org derf's collection.

Descriptions of all test videos are presented in a separate PDF provided with the report.

Codec Analysis and Tuning for Codec Developers and Codec Users

Computer Graphics and Multimedia Laboratory of Moscow State University:

20+ years working in the area of video codec analysis and tuning using objective quality metrics and subjective comparisons.
30+ reports of video codec comparisons and analysis (H.265, H.264, AV1, VP9, MPEG-4, MPEG-2, decoders' error recovery).
Methods and algorithms for codec comparison and analysis development, separate codec's features and codec's options analysis.

Strong and Weak Points of Your Codec

Deep encoder parts analysis (ME, RC on GOP, mode decision, etc).
Weak and strong points for your encoder and complete information about encoding quality on different content types.
Encoding Quality improvement by the pre and post filtering (including technologies licensing).

Independent Codec Estimation Comparing to Other Codecs for Different Use-cases

Comparative analysis of your encoder and other encoders.
We have direct contact with many codec developers.
You will know place of your encoder between other newest well-known encoders (compare encoding quality, speed, bitrate handling, etc.).

Encoder Features Implementation Optimality Analysis

We perform encoder features effectiveness (speed/quality trade-off) analysis that could lead up to 30% increase in the speed/quality characteristics of your codec. We can help you to tune your codec and find best encoding parameters.

Discussion

This study presents a comprehensive comparison of hardware video codecs with a focus on video quality assessed through objective metrics such as PSNR, SSIM, VMAF etc. The research could be helpful in selecting the encoder with the best quality for the community applications. In addition, benchmarking of the encoders can help developers to independently compare the solutions.

However, we acknowledge that the research has some limitations as we only focus on stream quality analysis, while for hardware encoders, business applications latency, density and power consumption may be critical. We are working on adding more parameters to the benchmark.