Open Visual Cloud is an open source project by Intel that includes a set of predefined pipeline references for visual cloud usage cases.
The project was born to address the complexity of deployment of a cloud service aimed at providing visual experiences. Open Visual Cloud provides interoperable…building blocks that allow developers to deliver advanced visual experiences to end users.
The basic ingredients of Open Visual Cloud
Open Visual Cloud provides the four main building blocks of visual cloud services. It also helps bridge the industry’s technological gaps in the five key services: media processing, media analytics, immersive media, cloud graphics and cloud gaming.
The four core Building Blocks to create a visual cloud service are as follows.
Encoding: compression of video data to reduce its size, including the new SVT-AV1 codec just released.
Decoding: the process of decompression of encoded video data.
Inference: data analysis using deep learning techniques to identify patterns or predictions.
Render: the process of generating graphic images, using raytracing or rasterization techniques.
Services enabled by Open Visual Cloud
There are five services, as we said, enabled by Open Visual Cloud.
Media Processing and Delivery includes technologies such as coding and decoding and video streaming. For media usage such as live streaming, broadcast and l’over-the-top (OTT).
Developers can also create advanced media analytics applications, incorporating technologies such as artificial intelligence and exploiting IoT edge devices. For example, for processing feeds from security cameras or traffic cameras in a smart city.
L’Immersive Media refers to experiences of virtual reality and augmented reality. In this case, among the possible use cases there is 360 degree live streaming, immersive entertainment experiences and much more.
The Cloud Gaming seems to be the future. Online games are starting to move to a cloud streaming model just like music, movies and TV programs.
Cloud Graphics allows you to work remotely and collaboratively on 3D graphics applications. Thanks to Open Visual Cloud, rendering and complex visualizations can be managed by high-performance servers in a virtualized environment.
Visual Cloud for a fast deployment
Open Visual Cloud reference pipeline is made up of highly optimized software ingredients for multimedia applications, artificial intelligence and rendering. They allow developers to quickly evaluate, optimize and distribute new services and applications. Without having to invest heavily in optimizing low-level software.
In addition, by adopting the most popular open source projects and frameworks, developers can benefit from the maximum reuse in existing investments. Like FFMPEG, GStreamer, OpenVINO, TensorFlow and other deep learning frameworks.
Open Visual Cloud has a modular approach that allows developers to simply exploit plug-ins when they need to migrate workloads from a Cpu to an accelerator. Maintaining top level software consistent along the pipeline.
These end-to-end pipeline references, Intel explains, are supplied as Docker containers. There are two reference pipes already available in Open Visual Cloud. They are: Content Delivery Network (CDN) – Transcode Sample and Intelligent Ad-insertion Sample. New pipeline, Intel assures, will be added regularly, as well as there will be continuous improvements and innovations.
More information about Open Visual Cloud is available at this link.
In addition to Open Visual Cloud, the latest NAB Show Intel has also presented with Netflix the next generation SVT-AV1 video codec.