The Pro AV and Broadcast markets are rapidly migrating to 4K and 8K video with higher dynamic range and higher frame rates, as well as adopting IP-based transport protocols. This leads to increasing demands for bandwidth, storage, real-time low latency video processing and often compression, along with the ability to differentiate through video quality improvements and features. Building on programmable and adaptable platforms results in highly integrated and cost optimized products with tight coupling between embedded software and multichannel AV processing pipelines.
At ISE 2019, Xilinx and its Alliance Program Members are showcasing 4K, 8K, and AV over IP ready technologies on leading edge Xilinx platforms that will increase your ability to innovate, differentiate and accelerate time-to-market.
For more information please visit www.xilinx.com/proav
The following demos will be showcased in the Xilinx booth #15-D240
8K HDMI® 2.1
The industry’s first FPGA implementation of HDMI 2.1 running on the Xilinx ZCU106 evaluation kit with new HDMI 2.1 FMC (available soon). With data rates up to 48Gbps, HDMI 2.1 supports a range of higher video resolutions and frame rates including 8K60 and 4K120 for a more immersive experience in LED walls, large format displays and digital signage.
8K DisplayPort™ 1.4 with DSC
Presented by Hardent, their new core for VESA Display Stream Compression (DSC 1.2a) is integrated with the Xilinx DisplayPort 1.4 subsystem and shown running on the ZCU102 Evaluation Kit with an inrevium DisplayPort 1.4 FMC. DSC provides visually lossless picture quality for all types of content whilst increasing data transport capacity by up to 3x, enabling 8K60 and 4K120 for instance. DSC also offers ultra-low latency performance, reduces power consumption, system costs, and EMI and is the VESA standard for all kinds of DisplayPort sources and displays across the industry.
4K AV over 1GbE Module
Presented by Macnica Technology and intoPIX, this new system on module (SOM) creates a bridge between 4K HDMI and 1GbE using Xilinx HDMI 2.0 subsystem and Ethernet MACs, intoPIX TICO-XS mezzanine codec and Macnica SMPTE ST 2110-based IP transport. The 4K AV over 1GbE module can be placed onto a carrier card or used standalone in any connected pro AV system to provide a very low-latency, visually lossless way to carry 4K AV content over low-cost 1GbE infrastructure, reducing the cost of cabling that would otherwise be required, extending reach, and proving IT-based control capabilities. Because everything is provided on a module, the time to integrate, and therefore time-to-market, is dramatically shortened. By adopting an interoperable and standards-based approach for IP connectivity with ST 2110, audio, video and data streams can be carried on the same link, but processed and controlled separately on the network, whilst retaining all the necessary timing and synchronization information.
H.265/H.264 Low Latency Streaming
Based on the H.265/H.264 video codec unit (VCU) integrated in Xilinx Zynq UltraScale+ MPSoC EV devices, this demonstration shows simultaneous bidirectional encode and decode using new low-latency modes available in our VCU Targeted Reference Design. Based on ZCU106 evaluation boards, we show 4K content from HDMI sources being encoded, sent over Ethernet, decoded and presented to 4K HDMI displays. The Zynq UltraScale+ MPSoC, with its high integration of on-board codec, ARM® processors and GPU, embedded peripherals, AV-capable transceivers and high performance programmable logic provide the ideal AV processing platform for streaming endpoints, video conferencing, cameras, displays and many other pro AV applications.
Machine Learning for Digital Signage
Using Xilinx' DNN for machine learning with Deephi technology, this proof-of-concept demonstration highlights how cameras placed on digital signage could be used to detect people and faces, perform classification of viewer attributes such as gender and age, and present targeted ads and/or capture viewership metrics. The demonstration is shown on the ZCU104 evaluation kit for edge processing, but could also be run in servers on Alveo accelerator cards.