Researchers have proposed a memory-centered motion estimation framework for VVC/H.266 that utilizes an Adaptive Winograd Coding Unit Partitioning accelerator to reduce redundant RDO processes. The FPGA-based design demonstrates a 0.63% bitrate reduction and 1.13 dB BD-PSNR improvement, providing a more efficient hardware path for real-time VVC encoding.
This hardware-centric approach directly addresses the computational complexity of VVC/H.266, which typically requires significantly more processing power than HEVC. By optimizing memory access through SRAM-based candidate block storage and reducing redundant calculations, the architecture provides a viable path for real-time 4K and 8K Ultra HD streaming on FPGA hardware. For the broader ecosystem, these efficiency gains suggest that the 50% bitrate reduction promised by VVC can be achieved without prohibitive hardware costs or latency. Watch for whether these Winograd-based acceleration techniques are integrated into commercial Xilinx Vivado IP cores for live broadcast encoders.
Researchers have introduced a memory-centered VVC encoding architecture featuring an Adaptive Winograd Coding Unit Partitioning accelerator. By streamlining motion estimation and reducing redundant calculations, the design achieves a 1.13 dB signal-to-noise ratio improvement. This hardware-centric approach enables efficient real-time 4K and 8K streaming, potentially lowering the costs associated with VVC implementation.
The architecture achieves a 1.13 dB improvement in Bjøntegaard Delta Peak Signal-to-Noise Ratio and a 0.63% bitrate reduction compared to existing motion estimation methods.
It utilizes an Adaptive Winograd Coding Unit Partitioning accelerator to predict partitions and bypass redundant rate-distortion optimization processes, while using a Hierarchical Carry Select Adder to accelerate Sum of Absolute Differences computations.
The design was implemented on Xilinx hardware, consuming 6,958 Look Up Tables and 19,750 Flip Flops, and was simulated using Xilinx Vivado.
VVC typically requires high processing power; this architecture optimizes memory access and reduces redundant calculations, providing a viable path for real-time 4K and 8K Ultra HD streaming on FPGA hardware without prohibitive costs.
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source
