It is a variation of delta modulation (DM) designed to overcome the limitations of fixed step size in standard DM systems. In essence, the step size used to approximate the input signal is not constant but dynamically adjusted based on the characteristics of the input signal. For instance, if a continuous sequence of the same sign is encountered, it signifies that the signal is rapidly changing, and the step size is increased. Conversely, if the output oscillates, indicating a slower signal change or noise, the step size is decreased. This adaptive approach ensures that the modulator can track the input signal more accurately, leading to improved signal-to-noise ratio (SNR) compared to its non-adaptive counterpart. This capability allows the system to effectively encode signals with a wide dynamic range without requiring a large bandwidth.
Its importance lies in its enhanced ability to efficiently encode signals with varying amplitudes and frequencies. By dynamically adjusting the quantization step size, it reduces the occurrence of slope overload distortion, a common issue in standard DM, where the modulator struggles to keep pace with rapidly changing signals. Furthermore, it also minimizes granular noise, which arises when the step size is too large for relatively flat signal segments. Historically, it offered a more sophisticated approach to speech encoding and other audio applications, offering a reasonable trade-off between complexity, bit rate, and signal quality. The improvements in SNR and dynamic range it offers can be particularly beneficial in noisy environments or where signal fidelity is paramount.
