Neurealm developed a display port-based multi-channel video processing unit on FPGA for a Tier-1 audio and imaging semiconductor customer, supporting dual DisplayPort inputs with multi-stream transport and four DP outputs.

Business Context

A global Tier-1 company that designs and manufactures cinema, television, broadcast, and entertainment imaging products needed a high-performance video processing solution capable of ingesting two DisplayPort 1.2 inputs with Multi-Stream Transport (MST), applying custom video acceleration/processing algorithms (e.g., rotation, filtering), and routing four DisplayPort outputs. This required a complex, timing-critical FPGA-based architecture and custom logic to meet throughput and synchronization goals.

Solutions

Neurealm selected the appropriate FPGA (Xilinx Kintex Ultrascale XCKU095), finalized pin-out and microarchitecture, and integrated custom RTL blocks with vendor IP to implement the processing pipeline. They coded and optimized custom algorithms in HDL, completed synthesis and timing closure, and performed system-level testing to deliver a robust multi-channel video processing unit meeting performance and functional requirements.