Releases: intel/media-delivery
Media Delivery Software Stack v3.1.2
That's release update of v3.1 release. Important changes:
- Cherry-picked fix for 1N transcoding use cases which error out handling EOS (trac-10261, 10262)
- Updated to use
-framerateinstead of-rforrawvideo
Full Changelog: v3.1.1...v3.1.2
Media Delivery Software Stack v3.1.1
That's release update of v3.1 release. Important changes:
- FFmpeg updated to n6.0
- IGT updated to v1.27.1
- Updated ffmpeg cmdlines to avoid usage of options deprecated by ffmpeg
- Better document docker setup
- Wrap up documented ffmpeg features (as available in n6.0)
Full Changelog: v3.1...v3.1.1
Media Delivery Software Stack v3.1
What's new in v3.1?
- Simplified and updated project documentation (see #124 for details)
- Updated media stack to later version pulling in new features and fixes (see #126 and #135 for details). Major changes:
- Added support of AV1 enctools
- Added support of 8K encoding with Intel® Deep Link Hyper Encode
- Added 10-bit encoding into quality benchmarking scripts (see #139 for details)
What's new in v3?
- Provided quality and performance benchmark data for Intel® Data Center GPU Flex Series
- Updated ffmpeg version to the one supporting Intel® oneAPI Video Processing Library
What was new in v2.1?
- Enabled Intel® Data Center GPU Flex Series (products formerly Arctic Sound) support
- Further quality improvements via usage of new software bitrate controller (dubbed EncTools)
- Added quality benchmarking path to assess Low Latency encoding
- Added AV1 decoding/encoding support to quality and performance benchmarking scripts
- Switched to use Intel® oneAPI Video Processing Library
- Switched to use M4 templates to generate dockerfiles
- Updated container ingredients (ffmpeg, vmaf, media-driver, etc.)
Summary
Intel Media Delivery Software Stack samples demonstrate how to configure publicly available media stack for Intel GPU and use best practices to achieve optimal video quality and performance on Intel GPUs for content delivery networks.
Featured documentation:
- README
- Quality assessment methodology and infrastructure:
- Performance assessment methodology and infrastructure
- GPU Virtualization Guide for Passthrough and SR-IOV
Hardware requirements:
- System with Gen8+ integrated Intel GPU
Software requirements:
- Linux* Operating System with up to date kernel supporting underlying Intel GPU:
| Intel GPU | Supported Setup |
|---|---|
| Intel® Data Center GPU Flex Series | Intel GPU DKMS Setup Guide |
| Other Intel GPUs | Vanilla kernels |
- Docker*
Verified configurations:
- Intel® Data Center GPU Flex Series (products formerly Arctic Sound)
- Gen9 Intel GPUs: SKL (Skylake) and KBL (Kaby Lake)
© 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.
Media Delivery Software Stack v3.0.1
That's release update of v3 release. Important changes:
- #119: added quick example scripts
- #123: clarified ATS-M setup instruction that
pci=realloc=offmight be needed on Linux kernel cmdline - #137: fixed issue with reference for AV1 encoding data in measure-quality script
- #137: fixed measure-perf to support multiple devices
- #117,, #137, etc.: various documentation fixes and updates
Full Changelog: v3...v3.0.1
Media Delivery Software Stack v3
What's new in v3?
- Provided quality and performance benchmark data for Intel® Data Center GPU Flex Series
- Updated ffmpeg version to the one supporting Intel® oneAPI Video Processing Library
What was new in v2.1?
- Enabled Intel® Data Center GPU Flex Series (products formerly Arctic Sound) support
- Further quality improvements via usage of new software bitrate controller (dubbed EncTools)
- Added quality benchmarking path to assess Low Latency encoding
- Added AV1 decoding/encoding support to quality and performance benchmarking scripts
- Switched to use Intel® oneAPI Video Processing Library
- Switched to use M4 templates to generate dockerfiles
- Updated container ingredients (ffmpeg, vmaf, media-driver, etc.)
Summary
Intel Media Delivery Software Stack samples demonstrate how to configure publicly available media stack for Intel GPU and use best practices to achieve optimal video quality and performance on Intel GPUs for content delivery networks.
Featured documentation:
- README
- Quality assessment methodology and infrastructure:
- Performance assessment methodology and infrastructure
- GPU Virtualization Guide for Passthrough and SR-IOV
Hardware requirements:
- System with Gen8+ integrated Intel GPU
Software requirements:
- Linux* Operating System with up to date kernel supporting underlying Intel GPU:
| Intel GPU | Supported Setup |
|---|---|
| Intel® Data Center GPU Flex Series | Intel GPU DKMS Setup Guide |
| Other Intel GPUs | Vanilla kernels |
- Docker*
Verified configurations:
- Intel® Data Center GPU Flex Series (products formerly Arctic Sound)
- Gen9 Intel GPUs: SKL (Skylake) and KBL (Kaby Lake)
© 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.
Media Delivery Software Stack v2.1
What's new in v2.1?
- Enabled Intel® Data Center GPU Flex Series (products formerly Arctic Sound) support
- Further quality improvements via usage of new software bitrate controller (dubbed EncTools)
- Added quality benchmarking path to assess Low Latency encoding
- Added AV1 decoding/encoding support to quality and performance benchmarking scripts
- Switched to use Intel® oneAPI Video Processing Library
- Switched to use M4 templates to generate dockerfiles
- Updated container ingredients (ffmpeg, vmaf, media-driver, etc.)
Summary
Intel Media Delivery Software Stack samples demonstrate how to configure publicly available media stack for Intel GPU and use best practices to achieve optimal video quality and performance on Intel GPUs for content delivery networks.
Featured documentation:
- README
- Quality assessment methodology and infrastructure:
- Performance assessment methodology and infrastructure
- GPU Virtualization Guide for Passthrough and SR-IOV
Hardware requirements:
- System with Gen8+ integrated Intel GPU
Software requirements:
- Linux* Operating System with up to date kernel supporting underlying Intel GPU:
| Intel GPU | Supported Setup |
|---|---|
| Intel® Data Center GPU Flex Series | Intel GPU DKMS Setup Guide |
| Other Intel GPUs | Vanilla kernels |
- Docker*
Verified configurations:
- Intel® Data Center GPU Flex Series (products formerly Arctic Sound)
- Gen9 Intel GPUs: SKL (Skylake) and KBL (Kaby Lake)
© 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.
Media Delivery Software Stack v2
This release adjusts samples to support Intel® Iris® Xe MAX Graphics (formerly DG1).
Quality and Performance Collaterals
- Quality and performance collateral for Intel® Iris® Xe MAX Graphics (formerly DG1)
Summary
Intel Media Delivery Software Stack samples demonstrates how to easy configure publicly available media stack for Intel GPU and use best practices to achieve optimal video quality and performance on Intel GPUs for content delivery networks.
Featured documentation:
- README
- Quality assessment methodology and infrastructure
- Performance assessment methodology and infrastructure
Hardware requirements:
- System with Gen8+ integrated Intel GPU
Software requirements:
- Linux* Operating System with up to date kernel supporting underlying Intel GPU
- See setup instructions for Intel® Iris® Xe MAX Graphics (formerly DG1)
- Docker*
Verified configurations:
- Gen9 Intel GPUs: SKL (Skylake) and KBL (Kaby Lake)
- Gen12 Intel GPUs: Intel® Iris® Xe MAX Graphics (formerly DG1)
What's new in v2?
- Enabled Intel® Iris® Xe MAX Graphics (formerly DG1) platform support thru updated docker image ingredients and scripts fixes (see details below)
- Fixed issues with multi-gpu systems support in ffmpeg command lines and demo scripts
- Switched to intel_gpu_top which supports multi-gpu systems
- Added VDENC support for measure-perf and minor fixes for VDENC path in measure-quality
- measure-perf fixed to properly support .mp4 streams with 2+ substreams (#33)
- measure-quality fixed to properly handle streams having out-of-bound bdrate (#29)
© 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.
Media Delivery Software Stack v1
This is initial release of Media Delivery Software Stack. This collection of samples demonstrates how to easy configure publicly available media stack for Intel GPU and use best practices to achieve optimal video quality and performance on Intel GPUs for content delivery networks.
Featured documentation:
- README
- Quality assessment methodology and infrastructure
- Performance assessment methodology and infrastructure
Hardware requirements:
- System with Gen8+ integrated Intel GPU
Software requirements:
- Linux* Operating System with up to date kernel supporting underlying Intel GPU
- Docker*
Verified configurations:
- Gen9 Intel GPUs: SKL (Skylake) and KBL (Kaby Lake)
© 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.