JAWAHARLAL NEHRU TECHNOLOGICAL UNIVERSITY HYDERABAD B.Tech III DIGITAL COMMUNICATIONS Previous Question Paper (March - 2024 - R16)

The sampling theorem for a band-pass signal states that if a signal is band-limited to a frequency range between f_L and f_H, with a bandwidth B = f_H - f_L, it can be uniquely reconstructed from its samples if the sampling frequency f_s satisfies the condition:

Specifically, for a signal where f_H = nB, the minimum sampling rate is 2B, provided the signal occupies a frequency band that allows for this band-pass sampling condition.

The primary difference lies in the step size used during the quantization process. In uniform quantization, the step size (Δ) is constant across the entire dynamic range of the input signal. This approach is simple to implement but results in a constant signal-to-quantization-noise ratio (SQNR) that is only optimal for signals with a uniform probability distribution.

In contrast, non-uniform quantization utilizes a variable step size. Smaller step sizes are used for low-amplitude signal levels, while larger step sizes are employed for high-amplitude levels. This is typically achieved by passing the signal through a compressor before uniform quantization. This method is used to maintain a relatively constant SQNR over a wide range of input signal amplitudes, making it highly effective for speech signals where low-amplitude samples are more frequent than high-amplitude ones.