Deliver Fast, Brilliant Video Experiences with New Intel® Media Server Studio 2017 R3

Developers: Optimize Video Performance & Quality

video codec developerVideo delivery and cloud graphics solution providers can create innovative, enterprise-grade media solutions that deliver fast, high quality video transcoding, speed transition to higher frame rates and resolutions, and reduce infrastructure and development costs with Intel® Media Server Studio.

Download FREE Community Edition Now

 

Quick Summary

The Linux* version delivers major HEVC encode video quality improvements, such as control over CU QP and support of P-frames.Significant enhancements were also made to AVC Flexible Encode Infrastructure (FEI) adding explicit weighted prediction for P-frames and bias adjustment. VDBOX balancing improves a performance for multisession and multiprocess decode, encode and encode for all codecs. New OpenCL™ extensions (VEBox, HEVC PAK and NV12) provide ultimate flexibility for programming Intel CPUs and graphics processors. Support of the latest CentOS v.7.3 deliver new operating system features and security updates.

The R3 release for the Windows* version brings substantial video encoding improvements for cloud gaming and remote desktop applications. These include support of weighted predictions for P- and B-frames and low delay bitrate control for AVC, and max frame size bitrate control for HEVC.

Get the Details: Key Linux Features

The key Visual Quality improvements that the Linux* version delivers include:

  • AVC FEI ENCODE Explicit Weighted P Prediction allows an encoder to specify the usage of scaling and to offset for P-frames when motion compensation is performed. It enables better compression efficiency in fade in and out, and when brightness is changed.
  • AVC FEI ENCODE Bias Adjustment provides four new low-level controls: 1) the ability to use raw references (useful to get better visual quality on high QP), 2) direct bias adjustment, 3) global motion bias adjustment, 4) motion vector cost scaling, which can be used for more optimal mode decision.
  • AVC FEI ENCODE Arbitrary Field Order in Reference List gives users full control over reference list for interlace encoding.
  • AVC FEI PAK productized: improves test coverage and finalized API as well, as it includes support of the memory control operation command to construct arbitrary references.
  • AVC Improved coding efficiency of B frames preceding forced IDR enables encoding optimization by changing frame type from B to P before inserting IDR.
  • HEVC Encode CU QP support allows users to assign QP for particular LCU to optimize quality within the frame and make smart bit distribution. It can be used to enhance flexibility for selecting ROI or building custom bitrate control.
  • HEVC Encode P Frame Support unlocks a wider range of applications to build by including solutions that limit support of B-frames.

Performance Enhancements include:

  • VDBOX Balancing is an optimization of the kernel mode driver (KMD) scheduler that improves a load balancing between two VDBOXs (video codec blocks) in Intel graphics processors. It refines a throughput and optimizes performance in multisession and multiprocess use cases for all codecs: AVC, HEVC, MPEG-2 and more.
  • AVC Encode Repartition Check allows to disable and enable a partition refinement to reduce the distortion by finding the most optimal partition. This feature is especially beneficial for improving encoding performance of low resolution streams.
  • HEVC Encode Multi Slice Support allows users to improve robustness and latency in end-to-end transmission.

The R3 release also brings new OpenCL™ features and extensions:

  • VEBox extension (preview) allows users to use an independent fixed-function block within the Intel graphics processor for a variety of image enhancement functions. VEBox is able to execute concurrently with regular OpenCL kernels running on GPGPU without impacting the EUs performance.
  • HEVC PAK extension (preview) unlocks an access to the final stages of the HEVC encoding pipeline, including residual computation, forward transformation, forward quantization, and entropy encoding (CABAC) with NV12 image formats.
  • Enhanced NV12 support in VA_API surface sharing allows to get an entire NV12 image instead of only separate luminance and chrominance channels.

Major Windows Release Enhancements

New features in the Intel Media Server Studio for Windows* version are primarily useful for cloud gaming and remote desktop application developers:

  • AVC Weighted P and B Predictions enable an encoder to specify the usage of scaling and offset for P- and B-frames when motion compensation is performed. It enables better compression efficiency in fade in and out, and when brightness is changed.
  • AVC Sliding Window Bitrate Control allows users to limit the average bitrate within a sliding window for better control over the video stream latency according to network conditions.
  • AVC Max Frame Size I/P provides an ability to set different restrictions for the maximum frame size of I and P-frames to improve the visual quality.
  • AVC Slice Based Intra Refresh enables users to insert refreshed areas aligned to slice size and structure. It improves stream robustness to transmission losses, allows to avoid overlapping between intra regions in consecutive frames, and as result the encoder uses fewer bits per stream and delivers better quality.
  • AVC Low Delay Bitrate Control provides strict control over maximum frame size and to set parameters not to exceed a defined threshold to secure low delay per frame transmission over network.
  • HEVC Max Frame Size Bitrate Control enables control over frame size to better align with network conditions.
Adaptive QVBR Low Delay in Intel® Media SDK 2017 for Windows*
Significant Advantages in Bitrate Accuracy in Packet Constrained Environments
Image Placeholder
Comparison of average bitrate accuracy over 10 test streams, each 3600 frames long, showing dramatic improvement under highly constraining bitrate requirements, similar to online gaming. The left (blue) bar corresponds to average QVBR bitrate control constrained by Max Frame Size; the right (orange) bar corresponds to average bitrate from the new Low Delay BRC mode, likewise constrained by MaxFrameSize. In both sequences, MaxFrameSize for P-frames is equal to the average frame size; while MaxFrameSize for I frames is set to 3X the P-size. Error bars represent the maximum and minimum across all sequences. Tests were performed by Intel, and run using Intel® Media Server Studio 2017 R3 for Windows, on Windows 10 64-bit, Intel graphics driver 15.45, Intel ® 6th Generation Core™ i5-6260U Processor, 8 GB DDR3-1600, on an Intel NUC model NUC6i5SYH. Tests of 10 standard 1080p60 sequences from https://media.xiph.org/video/derf/  Gaming section, at 16, 20, and 24 mbps. See also footnotes below.

 

About Intel Media Server Studio

Intel Media Server Studio allows video delivery and cloud graphics solution providers to quickly create, optimize and debug media applications, enable real-time video transcoding in media delivery, cloud gaming, remote desktop, and more solutions targeting the latest Intel® Xeon® and Intel® Core™ processor-based platforms.**

More Resources and Get Help

 

OpenCL and the OpenCL logo are trademarks of Apple Inc. used by permission by Khronos

 

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: http://www.intel.com/performance Source: Intel measured as of June 2017. Optimization Notice: Intel's compilers may or may not optimize to the same degree for non-Intel microprocessors for optimizations that are not unique to Intel microprocessors. These optimizations include SSE2, SSE3, and SSSE3 instruction sets and other optimizations. Intel does not guarantee the availability, functionality, or effectiveness of any optimization on microprocessors not manufactured by Intel. Microprocessor-dependent optimizations in this product are intended for use with Intel microprocessors. Certain optimizations not specific to Intel microarchitecture are reserved for Intel microprocessors. Please refer to the applicable product User and Reference Guides for more information regarding the specific instruction sets covered by this notice.

For more complete information about compiler optimizations, see our Optimization Notice.