Processing Technique
Method for redistributing the power of quantization errors to frequencies where they are less detrimental to the signal. The quantizer noise shaping process uses a feedback loop to push the error energy out of the baseband and into a higher frequency range. This technique allows for higher effective resolution within the signal bandwidth than the bit depth of the converter would otherwise permit.
Feedback Architecture
Architecture of the shaper includes a filter that processes the difference between the input and the quantized output. By carefully selecting the coefficients of this filter, quantizer noise shaping minimizes the noise floor in the region of interest. The technique is a fundamental component of sigma delta converters, which rely on oversampling to provide wide dynamic range.
Performance Measurement
Evaluation of the noise power spectral density confirms the effectiveness of the shaping. A high performance analyzer identifies the frequency range where quantizer noise shaping has successfully lowered the noise floor.
Stability Boundary
Loop stability is the primary constraint when designing high order noise shapers. If the input signal exceeds the stable range of the quantizer noise shaping loop, the system may oscillate or saturate, resulting in massive distortion. Verification of gain margins across all expected input levels and temperatures ensures reliable operation of the loop in the field.