Chapter 3 / 3.7

3.7 REVISION QUESTIONS

1. What is meant by source coding in a digital communication system, and how does it differ from channel coding in terms of objectives and design constraints?

1. Explain the fundamental trade-off between bit rate, fidelity, and complexity in source coding. Why can no single coding technique be optimal for all applications?

1. Distinguish clearly between waveform coding, model-based coding, and hybrid waveform/model-based coding, giving one example of each and stating the key principle exploited by each approach.

1. State the Nyquist sampling theorem and explain, with an example, how aliasing arises when the sampling criterion is violated.

1. Define quantization noise. Why does increasing the number of quantization levels improve signal quality, and why does this improvement exhibit diminishing returns?

1. Briefly describe pulse code modulation (PCM).

1. Why is a sampling rate of 8 kHz used for narrowband telephony, and how does this lead to the standard 64 kbps PCM voice channel?

1. Explain the purpose of companding in PCM systems. How do μ-law and A-law companding improve perceived speech quality without increasing bit rate?

1. Describe the principle of adaptive differential pulse-code modulation (ADPCM). Why can differences between successive samples often be encoded using fewer bits than absolute sample values? Compare PCM and ADPCM in terms of bit rate, complexity, and typical applications.

1. Briefly describe delta modulation (DM). How does it differ conceptually from PCM and DPCM?

1. Explain slope overload distortion and granular noise in delta modulation. How do step size and sampling rate influence these effects?

1. What is meant by adaptive delta modulation or continuously variable slope delta modulation (CVSD)? How does step-size adaptation improve performance relative to basic delta modulation?

1. Why has CVSD found long-standing use in military and tactical communication systems, despite the availability of more efficient speech coders?

1. Briefly compare PCM and DM.

1. Briefly describe entropy coding.

1. Briefly describe predictive coding.

1. Explain the operating principle of a channel vocoder. Why is vocoded speech intelligible at very low bit rates, yet perceived as unnatural?

1. Contrast channel vocoders with formant vocoders. What modeling assumption allows formant vocoders to operate at lower bit rates?

1. Describe the speech production model underlying linear predictive coding (LPC). What roles do the excitation signal and the vocal-tract filter play?

1. What is meant by analysis-by-synthesis in CELP coders, and why does it result in improved subjective quality compared with earlier LPC vocoders?

1. Explain why MELP remains relevant in HF military communications, while CELP-based coders dominate civilian telephony.

1. Define information content and entropy. How does entropy quantify the theoretical lower bound on average code length?

1. Explain why fixed-length coding is inefficient for sources with unequal symbol probabilities.

1. Describe the principles of Huffman coding. Why can Huffman codes approach but not exceed the entropy limit?

1. Explain how arithmetic coding differs fundamentally from Huffman coding. Why can arithmetic coding achieve fractional bits per symbol on average?

1. Why are entropy coders typically applied after predictive or transform coding rather than directly to raw samples?

1. Explain how predictive coding reduces redundancy in correlated sources. Why is the prediction error often easier to compress than the original signal?

1. Describe the principle of transform coding. Why does concentrating signal energy into a small number of coefficients enable efficient compression?

1. What role does the discrete cosine transform (DCT) play in image and video compression standards?

1. Explain the operating principle of dictionary coding using Lempel–Ziv methods. How does dictionary coding exploit redundancy differently from entropy coding?

1. Why are dictionary coders typically used for lossless compression, whereas transform coding is commonly used for lossy compression?

1. For a bandwidth-limited HF voice link subject to fading and interference, which source-coding approach would you choose and why?

1. For a broadband multimedia application over a reliable packet-switched network, which combination of source-coding techniques is most appropriate?

1. Explain how modern communication systems integrate waveform coding, predictive coding, transform coding, and entropy coding within layered architectures.