Accelerating Media Delivery
Intel’s media accelerators featuring Intel® Quick Sync Video transcode optimize both throughput and visual quality for video cloud distribution. Three classes of accelerator engines target high density video decode, encode, and video processing. These unlock real-time customers needing:
File based, Just-in-time, and Live consumption and creation
Bandwidth-efficient media delivery for Adaptive bitrate streaming, user generated content upload, and high visual quality broadcast
Intel’s discrete graphics accelerators are well integrated into open source media frameworks such as FFmpeg and the Intel® Media SDK. These popular frameworks allow both complex pipeline support as well as extreme customization of accelerator control. To make these tools even more accessible to Linux developers we’re now providing build scripts in Docker on the latest Linux kernels.
About Intel® Iris® Xe MAX graphics
Intel’s integrated graphics products have successfully enabled cloud
video in products such as Intel Visual Compute
Acccelerator.
Intel’s new discrete graphics and media accelerators are built upon the
same media technology used in 11th Gen Intel® Core™ i7-1185G7 (Product
formerly Tiger Lake).
This accelerator has three classes of video accelerator engine:
two or more video (codec) engines accelerate video decode, encode, and low-power encode
one or more video enhancement engine accelerating video resize, color space conversion, denoise, deinterlace and more
render engine - distributed execution units combined with media samplers
Caption: Picture provides brief overview of Intel® Iris® Xe MAX Graphics Adapter. Other Adapters might have different architecture and features. See more information about supported media features.
The combination of three accelerator types ensures typical transcode operations can pipeline execution to minimize latency. Multiple accelerator units allow concurrent execution of multiple frames to maximize throughput.
Performance
Each Intel® Iris® Xe MAX graphics die achieves the highest levels of
performance for the modern generation video standards like HEVC, while
still supporting ultra-high density and high quality AVC transcode.
When using high level API’s like FFMPEG, we provide three convenient
operating presets that offer different tradeoffs between speed and
quality (many additional controls are available for developers use). See
key platform capabilities highlight below:
Quality
Intel® Iris® Xe MAX graphics offers significant HEVC encode improvements
over the previous generation of hardware encoders. When compared to
typical presets on popular software video encoders (x264* and x265*)
Intel® Iris® Xe MAX graphics provides acceleration at similar quality.
The graphs (below) illustrate the video bitrate savings of Intel® Iris®
Xe MAX graphics compared with the most common presets.
Caption: These charts illustrate quality savings as a percent of bitrate saved for 8-bit 420 720p and 1080p compressed video streams. Bitrate savings are computed as BDRATE (using piecewise linear approach). Each point on the chart is the average BDRATE computed across 27 standard short sequences generated in both CBR and VBR. The objective visual quality metric used in the BDRATE calculation is Luma PSNR, averaged across frames. BDRATE is calucated using baselines of x264 for AVC or x265 medium for HEVC.
Caption: This chart uses the same approach but illustrating the consistent average bitrate savigns of the Intel HEVC encoders.
FFMPEG - Tips for performance
FFMPEG is a framework flexible in supporting many video transcode pipelines. To simplify the use our dockerfiles make it easy to build FFMPEG and its dependencies for your Linux platform. The FFMPEG command lines below illustrate good practices in using Intel® Quick Sync Video. The use of “extbrc” demonstrates the use of developer configurable bitrate control, in these examples the defaults generate streams using pyramid coding and other quality optimizations.
Example 1: AVC VBR Encode
ffmpeg -f rawvideo -pix_fmt yuv420p -s:v ${width}x${height} -r $framerate \
-i $input -vframes $nframes -y \
-c:v h264_qsv -preset medium -profile:v high \
-b:v $bitrate -maxrate $(bitrate*2) -bitrate_limit 0 \
-bufsize $(bitrate*4) -g 256 -extbrc 1 -b_strategy 1 -bf 7 -refs 5 \
-vsync 0 $output
Example 2: AVC CBR Encode
ffmpeg -f rawvideo -pix_fmt yuv420p -s:v ${width}x${height} -r $framerate \
-i $input -vframes $nframes -y \
-c:v h264_qsv -preset medium -profile:v high \
-b:v $bitrate -maxrate $bitrate -minrate $bitrate -bitrate_limit 0 \
-bufsize $(bitrate*2) -g 256 -extbrc 1 -b_strategy 1 -bf 7 -refs 5 \
-vsync 0 $output
Example 3: HEVC VBR Encode
ffmpeg -f rawvideo -pix_fmt yuv420p -s:v ${width}x${height} -r $framerate \
-i $input -vframes $nframes -y \
-c:v hevc_qsv -preset medium -profile:v main \
-b:v $bitrate -maxrate $(bitrate*2) -bitrate_limit 0 \
-bufsize $(bitrate*4) -g 256 -extbrc 1 -refs 5 -bf 7 \
-vsync 0 $output
Example 4: HEVC CBR Encode
ffmpeg -f rawvideo -pix_fmt yuv420p -s:v ${width}x${height} -r $framerate \
-i $input -vframes $nframes -y \
-c:v hevc_qsv -preset medium -profile:v main \
-b:v $bitrate -maxrate $bitrate -minrate $bitrate -bitrate_limit 0 \
-bufsize $(bitrate*2) -g 256 -extbrc 1 -refs 5 -bf 7 \
-vsync 0 $output
Try it out
See Intel® Iris® Xe MAX graphics running a simple content delivery network (CDN) leveraging high density and high performance transcoding. Explore our quality and performance measurement infrastructure to level set quality and performance expectations from our solution.
See more details and get source code at https://github.com/intel/media-delivery.
Caption: screen shot from the CDN demo showing [top left] client status, [top right] server video engine performance telemetry, [bottom left] server CPU utilization, and [bottom right] server status.
Detecting and selecting your Intel® Iris® Xe MAX graphics device:
Each Intel® Iris® Xe MAX SOC is exposed as a unique driver instance in a typical Linux guest OS configuration. Your platform may also contain other devices, for example integrated graphics. To launch one or more instances of our sample on different devices you will need to supply a different parameter to each application instance in the sample above, replace e.g. /dev/dri/renderD128 with the device you want ot target. More information on mapping the Linux Device instance to the underlying hardware type may be found at https://dgpu-docs.intel.com/releases/index.html
Prerequisites:
System with Gen9-11 or Xe-LP graphics (Intel® Iris® Xe MAX graphics recommended for best performance)
Ubuntu 20.04 (w/ the Linux kernel supporting underlying Intel GPU, see https://dgpu-docs.intel.com/releases/index.html)
Client system to receive streaming video (can run on same system as server)
We would like to thank Yash Raj Films for the courtesy of offering their trailer “WAR” (Copyright: Yash Raj Films Pvt. Ltd) for use in our demo.
git clone https://github.com/intel/media-delivery && cd media-delivery
docker build \
$(env | grep -E '(_proxy=|_PROXY)' | sed 's/^/--build-arg /') \
--file Dockerfile.ubuntu \
--tag intel-media-delivery \
.
DEVICE=${DEVICE:-/dev/dri/renderD128}
DEVICE_GRP=$(ls -g $DEVICE | awk '{print $3}' | \
xargs getent group | awk -F: '{print $3}')
docker run --rm -it \
-e DEVICE=$DEVICE --device $DEVICE --group-add $DEVICE_GRP \
--cap-add SYS_ADMIN \
-p 8080:8080 \
intel-media-delivery \
demo http://localhost:8080/vod/avc/WAR_TRAILER_HiQ_10_withAudio/index.m3u8
Try high quality tuned command lines for HEVC and AVC video encoding:
# Film: WAR – Courtesy & Copyright: Yash Raj Films Pvt. Ltd.
wget https://repositories.intel.com/media/WAR_TRAILER_HiQ_10_withAudio.mp4
ffmpeg -an -c:v h264_qsv -i WAR_TRAILER_HiQ_10_withAudio.mp4 \
-c:v hevc_qsv -preset medium -profile:v main -b:v 2000000 \
-extbrc 1 -bf 7 -refs 5 \
vsync 0 -y WAR_2Mbps_VBR_QSV.h265
Start Developing
Intel® Iris® Xe MAX graphics uses exactly the same API’s and software components as Intel’s integrated graphics adapters. Developers and users can easily access Linux drivers, Intel Media SDK, and FFMPEG. Our commitment to open source allows developers to easily customize these components for any video application.
Use FFmpeg command line tool to perform basic trancode operations. See command line examples to achieve optimal quality level for content delivery usage scenarious. Check out generic examples for Intel Media SDK Plugins for FFmpeg.
Start developing or enhance your own application reading Intel Media SDK Manual. See command line examples for Media SDK samples to achieve optimal quality level for content delivery usage scenarious.
Links
Intel® Media SDK Plugins for FFmpeg (also known as Intel® Quick Sync Video Plugins for FFmpeg)
Performance collection details
Configuration notes:
Test by Intel as of 11/20/2020, 11th Gen Intel® Core™ i7-1185G7 (Product formerly Tiger Lake) @ 3.00GHz, 1 Socket, 2 threads per core, 8 total CPUs, Intel Turbo Boost enabled, Total Memory 7714372kB, BIOS: TGLSFWI1.R00.3373.A00.2009091720 (ucode: 0x60), Ubuntu 20.04 LTS, gcc (Ubuntu 9.3.0-10ubuntu2) 9.3.0. Performance data is collected on production Intel® Iris® Xe MAX graphics adapter formerly DG1 (PCI ID 0x4905), EU total 96. Intel® Iris® Xe MAX graphics performance data is obtained with GPU operating at a fixed 1.0 GHz with 16 GB onboard memory. Commercial products may operate at higher or lower frequency.
Multi-stream performance data is collected using scripts noted above running file-to-file transcode. The scripts execute multiple concurrent 720p, 1080p, or 4K content streams, measuring the average frame rate of the transcoding process, at increasing numbers of streams to seek a target (typically 30 fps or 60 fps). The maximum stream density that meets or exceeds 98% of the target fps is reported.
The following is a table of the project versions used (listed Intel projects are provided on repositories.intel.com/graphics).
Project versions
Component |
Version |
|---|---|
Intel® Media driver for VAAPI |
20.4.1 |
Intel® Media SDK |
20.4.1 |
libva2 |
2.9.1 |
Intel® Graphics Memory Management Library |
20.3.2 |
ffmpeg |
n4.3 |
Notices & Disclaimers
Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors.
Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more complete information visit www.intel.com/benchmarks.
Performance results are based on testing as of dates shown in configurations and may not reflect all publicly available updates. See backup for configuration details. No product or component can be absolutely secure.
Your costs and results may vary.
Software and workloads used in performance tests may have been optimized for performance only on Intel graphics adapters. Performance tests used here are measured using specific computer systems, components, software, operations, and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases including the performance of this product when combined with other products. For more complete information visit https://www.intel.com/performance.
Results have been measured on pre-production and production systems, and provided to you for information purposes. Any differences in your system hardware, software, or configurations may affect your actual performance.
Intel technologies, features, and benefits depend on system configuration and may require enabled hardware, software, or service activation. Performance varies depending on system configuration. No computer system can be absoutely secure. Check with your system manufacturer retailer or learn more at ark.intel.com
Intel and Iris are trademarks of Intel Corporation or its subsidiaries.
© Intel Corporation. Intel, the Intel logo, and other Intel marks are trademarks of Intel Corporation or its subsidiaries. Other names and brands may be claimed as the property of others.