Capella announces Split and Stitch, AI-assisted tools, and other new capabilities to be demonstrated for the first time at IBC 2024

Santa Clara, 13 September 2024: At IBC 2024, Capella Systems (Hall 7, D37), the creator of industrial-strength transcoding and VOD software, will demonstrate its next-generation of products. Capella’s transcoding suite, Cambria FTC, enables maximum productivity with high-density processing and lower total computing costs. New AI features allow customers to automatically create more metadata and reduce manual operations.

New at IBC: Split and Stitch - dynamic, distributed optimization

Distributed computing is the way forward. We can now analyze source material and split it into chunks, distributing them across multiple computing nodes. Jobs can be dynamically split across GPU and CPU processing, based on optimizations between performance and quality, for the most efficient workflow. This ‘split and stitch’ approach means we can dynamically allocate transcoding jobs based on customers’ priorities at the time. It also allows them to take advantage of low-cost CPU spot instances in the cloud.

New at IBC: AI facial recognition and AI-assisted tools

We now employ AI to save time creating metadata for media content. More metadata means more automation and more efficiency. Our AI detects and recognizes faces, blurring them where appropriate, and allows automated processing based on content. We also have advanced scene-change detection, video complexity analysis, speech detection, and sports score detection. We can even identify specific images to create ad insertion points automatically for live-streaming content, maximizing revenue. We have incorporated Open AI/Whisper-based Speech-to-Text and upgraded OCR with improved quality that can also detect ad break thresholds and other video events.

New at IBC: Massive GPU optimization

GPUs continue to transform the computing landscape. We always want to make the best use of available resources, so we have reworked our code to maximize productivity when running on GPU nodes. Used in conjunction with Split and Stitch (above), this gives unprecedented flexibility and a massive efficiency boost. A single Cambria Stream can process up to 25 channels of multi-layer encoding at once, reducing the total computing cost significantly compared to CPU-based encoding.

New at IBC: Cloud-ready workflows with Docker and Kubernetes 

Workflows can be unnecessarily complicated, but we make it easy. Our customers can deploy our products anywhere, quickly plugging Capella into their existing workflows. Our software runs on Windows or Linux, on-premise or in the cloud. We are now deployable via Docker and scalable via Kubernetes. Our modern REST-based API allows users full control of our products from their workflow engines, enabling them to orchestrate with maximum efficiency.

Capella CEO Ikuyo Yamada says, “We have spent the last year refining our Cambria suite to make it more efficient than ever. It’s the ideal transcoding platform for raising productivity and lowering costs. With cutting-edge, scalable architecture, we can respond quickly to clients’ changing needs, tightly integrating with existing workflows on-premise, in the cloud, or both.”

Focus on the future

At IBC 2024, we will show that Capella’s modern transcoding engine, Cambria FTC, is unbeatable for efficiency, maintainability, and support. Encoding/transcoding is a crowded field, but no other company has Capella’s combination of experience, performance, and flexibility. Our increasing use of rich metadata will enable users to increase the life-long value of their media. With our unparalleled quality, flexibility, and ability to integrate with modern workflows, Capella is the credible choice for today’s and tomorrow’s Live and VOD encoding facilities.