Digital Requantization Using LSB Correction to Reduce Output Ripple

The patent applies periodic action by using dithering sequences that are periodically applied to the quantization process. The dither signal is added periodically to the input signal before quantization, and the resulting quantization errors are distributed over time in a periodic manner. This periodic dithering approach improves the effective resolution of the quantizer while the periodic nature helps in designing filters that can remove the resulting periodic ripple components at known frequencies.

The patent changes the parameter of quantization step size dynamically by using variable step-size dithering. Instead of using a fixed quantization step, the system adjusts the step size based on the signal characteristics and operating conditions. This allows optimization of the trade-off between quantization noise reduction and ripple generation, improving precision while controlling the harmful ripple effects through adaptive parameter adjustment.

The patent applies segmentation by dividing the high-resolution quantization task into multiple lower-resolution steps. Instead of using a single high-resolution waveform generator, the system performs sequential quantization in multiple stages, each operating at lower resolution. The dithering technique is used to distribute quantization errors across these stages, achieving effective high-resolution control through multiple coarse quantization steps rather than one fine quantization step.

The patent introduces dithering signals as intermediary elements between the low-resolution quantizer and the desired high-resolution output. These dithering signals act as mediators that modify the quantization process, allowing a low-resolution waveform generator to effectively produce high-resolution control signals. The dithering intermediaries add controlled noise that, when processed through the quantization and filtering stages, yields improved precision equivalent to higher-resolution hardware.