New – Amazon EC2 VT1 Instances for Live Multi-stream Video Transcoding
Global demand for video content has been rapidly increasing and now has the major audiences of Internet and mobile network traffic. Over-the-top streaming services such as Twitch continue to see an explosion of content creators who are seeking live delivery with great image quality, while live event broadcasters are increasingly looking to embrace agile cloud…
Global demand for video content has been rapidly increasing and now has the major audiences of Internet and mobile network traffic. Over-the-top streaming services such as Twitch continue to see an explosion of content creators who are seeking live delivery with great image quality, while live event broadcasters are increasingly looking to embrace agile cloud infrastructure to reduce costs without sacrificing reliability, and efficiently scale with demand.
Today, I am happy to announce the general availability of Amazon EC2 VT1 instances that are designed to provide the best price performance for multi-stream video transcoding with resolutions up to 4K UHD. These VT1 instances feature Xilinx® Alveo™ U30 media accelerator transcoding cards with accelerated H.264/AVC and H.265/HEVC codecs and provide up to 30% better price per stream compared to the latest GPU-based EC2 instances and up to 60% better price per stream compared to the latest CPU-based EC2 instances.
Customers with their own live broadcast and streaming video pipelines can use VT1 instances to transcode video streams with resolutions up to 4K UHD. VT1 instances feature networking interfaces of up to 25 Gbps that can ingest multiple video streams over IP with low latency and low jitter. This capability makes it possible for these customers to fully embrace scalable, cost-effective, and resilient infrastructure.
Amazon EC2 VT1 Instance Type EC2 VT1 instances are available in three sizes. The accelerated H.264/AVC and H.265/HEVC codecs are integrated into Xilinx Zynq ZU7EV SoCs. Each Xilinx® Alveo™ U30 media transcoding accelerator card contains two Zynq SoCs.
Instance size
vCPUs
Xilinx U30 card
Memory
Network bandwidth
EBS-optimized bandwidth
1080p60 Streams per instance
vt1.3xlarge
12
1
24GB
Up to 3.12 Gbps
Up to 4.75 Gbps
8
vt1.6xlarge
24
2
48GB
6.25 Gbps
4.75 Gbps
16
vt1.24xlarge
96
8
192GB
25 Gbps
19 Gbps
64
The VT1 instances are suitable for transcoding multiple streams per instance. The streams can be processed independently in parallel or mixed (picture-in-picture, side-by-side, transitions). The vCPU cores help with implementing image processing, audio processing, and multiplexing. The Xilinx® Alveo™ U30 card can simultaneously output multiple streams at different resolutions (1080p, 720p, 480p, and 360p) and in both H.264 and H.265.
Each VT1 instance can be configured to produce parallel encoding with different settings, resolutions and transmission bit rate (“ABR ladders“). For example, a 4K UHD stream can be encoded at 60 frames per second with H.265 for high resolution display. Multiple lower resolutions can be encoded with H.264 for delivery to standard displays.
We provide a number of sample video processing pipelines for the VT1 instances. There are tutorials and code examples in the GitHub repository that cover how to tune the codecs for image quality and transcoding latency, call the runtime for the U30 cards directly from your own applications, incorporate video filters such as titling and watermarking, and deploy with container orchestration frameworks.
You can complement VT1 instances with AWS Media Services for reliable packaging and origination of transcoded content. To learn more, you can use a solution library of Live Streaming on AWS to build a live video workflow using these AWS services.
Available Now Amazon EC2 VT1 instances are now available in the US East (N. Virginia), US West (Oregon), Europe (Ireland), Asia Pacific (Tokyo) Regions. To learn more, visit the EC2 VT1 instance page. Please send feedback to the AWS forum for Amazon EC2 or through your usual AWS support contacts.
How Northpower used computer vision with AWS to automate safety inspection risk assessments
In this post, we share how Northpower has worked with their technology partner Sculpt to reduce the effort and carbon required to identify and remediate public safety risks. Specifically, we cover the computer vision and artificial intelligence (AI) techniques used to combine datasets into a list of prioritized tasks for field teams to investigate and…
In this post, we share how Northpower has worked with their technology partner Sculpt to reduce the effort and carbon required to identify and remediate public safety risks. Specifically, we cover the computer vision and artificial intelligence (AI) techniques used to combine datasets into a list of prioritized tasks for field teams to investigate and mitigate.
Architecture to AWS CloudFormation code using Anthropic’s Claude 3 on Amazon Bedrock
In this post, we explore some ways you can use Anthropic’s Claude 3 Sonnet’s vision capabilities to accelerate the process of moving from architecture to the prototype stage of a solution. Source
In this post, we explore some ways you can use Anthropic’s Claude 3 Sonnet’s vision capabilities to accelerate the process of moving from architecture to the prototype stage of a solution.
Control data access to Amazon S3 from Amazon SageMaker Studio with Amazon S3 Access Grants
In this post, we demonstrate how to simplify data access to Amazon S3 from SageMaker Studio using S3 Access Grants, specifically for different user personas using IAM principals. Source
In this post, we demonstrate how to simplify data access to Amazon S3 from SageMaker Studio using S3 Access Grants, specifically for different user personas using IAM principals.