MainConcept began optimizing professional video codecs for Arm more than five years ago, anticipating an industry that would increasingly operate across multiple compute architectures. Cloud providers are expanding Arm-based infrastructure, NVIDIA is pairing its rack-scale AI systems with Arm-based CPUs, and video operators are evaluating CPUs, GPUs, and dedicated VPUs for different workloads. For broadcast, streaming, and media platforms, that creates a planning problem: the hardware that offers the best performance or economics today may not be the hardware they want to use tomorrow.
EVA brings MainConcept's years of codec optimization and hardware integration work together behind a single API. Customers build their video application once and use the same interface across MainConcept software codecs and supported hardware. As requirements change, they can move between x86 and Arm, GPUs and VPUs, or software processing without building and maintaining a separate codec integration for each platform.
Moving a professional video codec from x86 to Arm requires significant engineering investment. MainConcept's early focus on Arm optimization and multi-platform integration has absorbed much of that specialized engineering, leaving customers free to focus on their applications and to choose the infrastructure that meets their needs.
On NVIDIA Arm-based Linux systems, EVA can keep decode, scaling, color conversion and encode stages within the GPU pipeline, drastically reducing repeated transfers between CPU and GPU memory. NETINT support extends EVA's common API model to dedicated VPUs, adding another processing architecture without requiring a separate application-level codec integration.
Faster Time to Market
MainConcept estimates that EVA can reduce implementation time and associated development costs by over 80% compared with independently integrating multiple hardware and software codecs. Beyond the initial integration, a common API also reduces the amount of platform-specific code that must be maintained as hardware options and deployment requirements change.
"Video infrastructure is entering a period where there is no single obvious compute platform," said Deacon Johnson, President and CEO of MainConcept. "Arm is expanding, GPU architectures continue to evolve, and dedicated VPUs are becoming viable at scale. Customers should not have to rebuild their video stack every time the hardware changes, which is why EVA is already running across broadcast and streaming deployments worldwide. Years of platform-specific codec work means our customers can keep one integration in place no matter how the underlying infrastructure shifts."
Hardware Flexibility Through a Common API
EVA supports MainConcept software codecs for AVC/H.264, HEVC/H.265, AV1, and JPEG XS, plus the Universal Colorspace Converter, alongside hardware platforms from AMD, Intel, NVIDIA, and Qualcomm GPUs and NETINT VPUs, across Windows and Linux on both x86 and Arm.
Applications detect supported hardware and codec capabilities at runtime and select GPU, VPU or software processing through the same API. EVA is deployed in production today across broadcast, streaming, playout, and government video environments.
Meet MainConcept at IBC2026
MainConcept will demonstrate the latest Easy Video API capabilities at IBC 2026, September 11 - 14 in Amsterdam, Booth 1.C15.
Learn more about Easy Video API at mainconcept.com/easy-video-api.
About MainConcept
Since 1993, MainConcept has provided best-of-breed video and audio codec solutions that fuel creativity and business globally for professional video production, broadcast, medical, surveillance and more. The company's software development kits and plugins are designed to meet an ever-expanding list of use cases. MainConcept codecs are used by the likes of Adobe, AVID, Blackmagic Design, Dalet, Grass Valley, Nikon, Ross, Sony, Techex, Telestream, and Wowza.
Photos: (Click photo to enlarge)
Source: MainConcept
Read Full Story - MainConcept Easy Video API Extends Full Transcoding to Arm and NETINT VPUs | More news from this source
Press release distribution by PRLog
