Video Filter Coefficient Precision Allocation

The patent divides the filter coefficient transmission into two parts: a first set of coefficients is transmitted with higher precision for critical filtering operations, while a second set of coefficients is transmitted with lower precision for less critical operations. This segmentation allows the system to allocate bits strategically based on the functional requirements of different filter coefficients, improving overall filtering performance while reducing total bit requirements.

The patent applies different quantization precisions to different portions of the filter coefficients based on their specific requirements. Important coefficients that significantly impact filtering quality are transmitted with higher precision, while less important coefficients are transmitted with lower precision. This local differentiation of quality ensures optimal filtering performance for critical parameters while minimizing unnecessary bit usage for less critical parameters.

The patent pre-provides multiple filter coefficients in the video data stream before the actual filtering operation. The decoder receives and stores these pre-encoded filters, allowing the system to select and apply the appropriate filter based on current video characteristics without requiring complex real-time filter generation or synthesis. This preliminary preparation simplifies the decoding process while maintaining high filtering adaptability.

The patent transmits multiple copies of filter coefficients with different precision levels for the same filter type. Instead of requiring the decoder to generate filters dynamically or combine multiple complex filter designs, the system provides ready-to-use filter copies that can be directly applied. This copying approach reduces decoder complexity by eliminating the need for complex filter synthesis operations while maintaining the ability to adapt to different video conditions through selective filter application.