Signal Representation
A modulation format represents an analog signal value using the relative density of a series of binary pulses. In pulse density modulation, the average value of the pulse train over a given time interval is proportional to the amplitude of the analog input. A high value is represented by a high density of logical ones, while a low value is represented by a high density of zeros.
Spectral Distribution
The frequency spectrum of the modulated signal shows that the quantization noise is pushed into higher frequency ranges. This noise shaping is achieved through delta sigma modulation, which keeps the signal of interest in the low frequency band. It allows high resolution measurements to be made with simple low resolution converters.
Analog Reconstruction
The recovery of the analog signal requires only a simple low pass filter. This filter removes the high frequency quantization noise, leaving the original low frequency signal. Such filtering reduces the complexity and cost of the analog interface circuit.
Digital Integration
The processing of these digital streams is highly efficient in modern microcontrollers and digital signal processors. For example, a digital micro-electro-mechanical microphone often outputs a pulse density modulation stream that can be directly filtered by a digital sinc filter. This direct digital interface reduces the susceptibility to electromagnetic interference and simplifies the routing of signals on a circuit board.