Pulse Code Modulation (PCM)
Complete Theory Notes with Prerequisite Topics
These notes are organized for a teacher/student preparing for a classroom session on Pulse Code
Modulation. Part A covers the prerequisite background a student must know before this topic, and
Part B covers the complete PCM theory in depth.
Table of Contents
Part A: Prerequisite Topics
A.1 Signals and Systems Basics
A.2 Analog Communication and Modulation (AM, FM, PM)
A.3 The Sampling Theorem (Nyquist Criterion)
A.4 Analog-to-Digital Conversion Overview
Part B: Pulse Code Modulation (PCM) - Complete Theory
B.1 Introduction to PCM
B.2 Block Diagram of a PCM System
B.3 Sampling
B.4 Quantization
B.5 Quantization Error and Signal-to-Quantization-Noise Ratio
B.6 Companding (mu-law and A-law)
B.7 Encoding
B.8 Bit Rate and Bandwidth of PCM
B.9 PCM Transmitter
B.10 PCM Receiver (Decoding and Reconstruction)
B.11 Advantages and Disadvantages of PCM
B.12 Applications of PCM
B.13 Comparison: PCM vs DPCM vs Delta Modulation
B.14 Worked Numerical Examples
B.15 Summary and Key Formulas
Part A: Prerequisite Topics
Before teaching Pulse Code Modulation, students should already be comfortable with the following
foundational topics. Without these, PCM concepts such as sampling rate, quantization noise, and
bit rate will not make intuitive sense.
A.1 Signals and Systems Basics
A signal is any physical quantity that varies with time, space, or another independent variable and
carries information. In communication engineering, we mainly deal with time-varying electrical
signals.
Key concepts students must know:
• Continuous-time vs Discrete-time signals: A continuous-time signal is defined for every instant
of time (e.g., speech waveform). A discrete-time signal is defined only at specific time instants
(e.g., after sampling).
• Analog vs Digital signals: An analog signal can take any value within a continuous range. A
digital signal takes only a finite set of discrete values (usually represented in binary).
• Periodic and Aperiodic signals: A periodic signal repeats itself after a fixed interval called the
time period. Most real-world information signals (speech, video) are aperiodic but can still be
analyzed using frequency-domain tools.
• Frequency domain representation (Fourier concept): Any signal can be expressed as a
combination of sinusoids of different frequencies. The highest frequency component present in a
signal is called its bandwidth (denoted fm), and this value is essential for deciding the sampling
rate in PCM.
• Noise: Unwanted random disturbances added to a signal during transmission. Digital systems
like PCM are more robust to noise than analog systems.
A.2 Analog Communication and Modulation (AM, FM, PM)
Modulation is the process of varying a high-frequency carrier signal's characteristics (amplitude,
frequency, or phase) in accordance with a low-frequency message (baseband) signal, so that it can
be transmitted efficiently over long distances.
• Amplitude Modulation (AM): The amplitude of the carrier is varied in proportion to the message
signal.
• Frequency Modulation (FM): The frequency of the carrier is varied in proportion to the message
signal.
• Phase Modulation (PM): The phase of the carrier is varied in proportion to the message signal.
Studying analog modulation first gives students a reference point: PCM is fundamentally different
because instead of modulating a continuous carrier with a continuous message, it first converts the
message into a stream of binary digits, and then this binary stream may be transmitted using digital
modulation techniques (ASK, FSK, PSK) at a later stage.
A.3 The Sampling Theorem (Nyquist Criterion)
The Sampling Theorem is the single most important prerequisite for PCM. It states:
A band-limited continuous-time signal with maximum frequency component fm can be completely
reconstructed from its samples if it is sampled at a rate fs greater than or equal to twice the
maximum frequency:
fs ≥ 2 fm
This minimum rate, 2fm, is called the Nyquist rate. If the signal is sampled below this rate, a
distortion called aliasing occurs, where higher frequencies fold back and appear as lower
frequencies, corrupting the signal permanently.
Example: Human voice for telephone-quality speech is band-limited to about 4 kHz, so the standard
sampling rate used in telephony is 8 kHz (which is above the Nyquist rate of 8 kHz exactly, i.e., 2 x
4 kHz).
A.4 Analog-to-Digital Conversion Overview
Before diving into PCM specifically, students should understand the general concept that
converting an analog signal into a digital signal requires three sequential steps:
1 Sampling - converting continuous time into discrete time instants.
2 Quantization - converting continuous amplitude into a finite set of discrete amplitude levels.
3 Encoding - representing each quantized level as a binary codeword.
PCM is essentially the formal technique that implements exactly these three steps. Understanding
this overview before the detailed PCM lecture helps students see the 'big picture' before the
mathematical details.
Part B: Pulse Code Modulation - Complete
Theory
B.1 Introduction to PCM
Pulse Code Modulation (PCM) is a method used to digitally represent sampled analog signals. It
was developed by Alec Reeves in 1937 and became the standard technique for digital transmission
of analog information such as voice and video. In PCM, the analog signal is sampled, the sampled
amplitude values are quantized into a finite number of levels, and each level is encoded into a
binary codeword. The resulting stream of binary digits (bits) can then be transmitted, stored, or
processed digitally.
PCM is called a waveform coding technique because it directly represents the waveform's
amplitude values, as opposed to source coding techniques that exploit the statistical structure of
the source.
B.2 Block Diagram of a PCM System
A complete PCM communication system consists of the following stages:
1 Low Pass Filter (LPF): Band-limits the input analog signal to remove frequency components
above fm, preventing aliasing.
2 Sampler: Samples the band-limited signal at a rate fs ≥ 2fm, producing a Pulse Amplitude
Modulated (PAM) signal.
3 Quantizer: Rounds each sample's amplitude to the nearest of a fixed set of discrete levels.
4 Encoder: Converts each quantized level into a binary codeword of n bits.
5 Transmission Channel: Carries the binary bitstream (may include line coding, regeneration,
and digital modulation).
6 Regenerative Repeaters (for long distance): Detect and regenerate clean pulses to remove
accumulated noise and distortion.
7 Decoder (at receiver): Converts the received binary codewords back into quantized PAM
samples.
8 Reconstruction / Low Pass Filter: Smooths the PAM samples back into a continuous analog
waveform, recovering the original message.
Diagram summary: Analog Input → LPF → Sampler → Quantizer → Encoder → Digital Output
(Transmitter side). Digital Input → Decoder → Reconstruction Filter → Analog Output (Receiver
side).
B.3 Sampling
Sampling is the process of measuring the amplitude of a continuous-time signal at regular time
intervals Ts = 1/fs. The output of the sampler is a Pulse Amplitude Modulated (PAM) signal, i.e., a
sequence of pulses whose amplitudes match the original signal at the sampling instants.
Types of sampling:
• Ideal (Instantaneous) Sampling: Uses impulse functions; theoretical concept used for
mathematical analysis.
• Natural Sampling: The pulse tops follow the shape of the analog signal during the pulse
duration.
• Flat-top Sampling: The pulse amplitude is held constant (flat) at the sampled value for the
duration of the pulse; this is what is practically used before quantization, and it introduces a small
distortion called aperture effect, corrected using an equalizer at the receiver.
As established in the prerequisite section, the sampling rate must satisfy the Nyquist criterion fs ≥
2fm to avoid aliasing.
B.4 Quantization
Quantization is the process of mapping the continuous range of sample amplitudes into a finite
number of discrete amplitude levels. Since a sample can take infinitely many possible amplitude
values, but digital systems can only represent a finite number of values, each sample is rounded to
the nearest available quantization level.
Key quantities:
• If n bits are used to represent each sample, the number of quantization levels is L = 2n.
• The step size (resolution) is ∆ = Vpp / L, where Vpp is the peak-to-peak range of the signal.
Types of Quantization:
• Uniform Quantization: All quantization steps (∆) are of equal size across the entire amplitude
range. Simple to implement but gives poor resolution for low-amplitude signals.
• Non-uniform Quantization: Step size is smaller for low-amplitude signals and larger for
high-amplitude signals. This matches human speech statistics (most speech energy is at low
amplitudes) and is implemented using companding (see Section B.6).
B.5 Quantization Error and Signal-to-Quantization-Noise
Ratio
Because quantization rounds each sample to the nearest level, an error is introduced, called
quantization error or quantization noise. This is the fundamental and unavoidable source of
distortion in PCM (unlike sampling, which is theoretically lossless if the Nyquist criterion is met).
The maximum quantization error is half the step size:
Maximum error = ± ∆/2
The Signal-to-Quantization-Noise Ratio (SQNR), which measures how much stronger the signal is
compared to the quantization noise, is approximately:
SQNR (dB) ≈ 6.02n + 1.76 dB
This important formula shows that every additional bit used per sample improves the SQNR by
approximately 6 dB. This is why increasing the number of quantization bits (n) directly improves the
quality of the reconstructed signal.
B.6 Companding (mu-law and A-law)
Companding is a combination of the words 'compressing' and 'expanding'. It is a non-uniform
quantization technique where the signal is compressed (using a logarithmic function) before
uniform quantization at the transmitter, and expanded back at the receiver. This effectively gives
finer resolution to small-amplitude signals (like quiet speech) without needing extra bits, improving
the SQNR for weak signals.
• mu-law Companding: Used in North America and Japan for telephone systems (typically mu =
255).
• A-law Companding: Used in Europe and most of the rest of the world for telephone systems
(typically A = 87.6).
Both achieve a similar purpose: they compress the dynamic range of the signal so that quiet and
loud sounds are both represented with acceptable quality using the same number of bits.
B.7 Encoding
Encoding is the process of representing each quantized amplitude level as a unique binary
codeword. If there are L = 2n quantization levels, each level is assigned an n-bit binary code. For
example, with n = 8 bits, there are 256 possible quantization levels, and each sample is
represented by an 8-bit binary word (a byte).
Common encoding formats include natural binary coding, and sign-magnitude or folded binary
coding (often used together with companding schemes).
B.8 Bit Rate and Bandwidth of PCM
Once the sampling rate fs and the number of bits per sample n are known, the bit rate (also called
signaling rate) of the resulting PCM signal is:
Bit Rate (R) = fs × n bits per second
The minimum transmission bandwidth theoretically required for a PCM signal is related to the bit
rate:
Bandwidth (BPCM) ≥ R / 2 = (fs × n) / 2
This shows an important tradeoff in PCM: increasing the number of bits per sample improves signal
quality (higher SQNR) but also increases the required bit rate and transmission bandwidth.
B.9 PCM Transmitter
The transmitter side of a PCM system performs the operations in this order:
1 The analog message signal passes through a low-pass anti-aliasing filter, limiting its bandwidth
to fm.
2 The filtered signal is sampled at fs ≥ 2fm, producing a PAM signal.
3 Optionally, the PAM signal is compressed (companded) to give more resolution to weak signal
amplitudes.
n
4 The compressed samples are quantized into L = 2 discrete levels.
5 Each quantized sample is encoded into an n-bit binary codeword by the encoder.
6 The resulting serial bit stream is line-coded (e.g., using a format such as NRZ or Manchester)
for transmission over the channel.
B.10 PCM Receiver (Decoding and Reconstruction)
The receiver side performs the reverse operations:
1 The received binary bit stream is regenerated (cleaned of channel noise and distortion) using
regenerative repeaters along the transmission path.
2 The decoder converts each n-bit binary codeword back into its corresponding quantized
amplitude level, reconstructing a PAM-like signal (a 'staircase' waveform).
3 If companding was used at the transmitter, an expander reverses the compression.
4 The staircase signal is passed through a reconstruction (low-pass) filter, which smooths it into a
continuous analog waveform that approximates the original message signal.
Note: Because quantization is an irreversible, lossy process, the reconstructed signal at the
receiver is not perfectly identical to the original; it differs by the quantization error introduced at the
transmitter. This is the fundamental theoretical limitation of PCM.
B.11 Advantages and Disadvantages of PCM
Advantages:
• High noise immunity - since only two levels (0 and 1) need to be distinguished, PCM signals are
far more resistant to channel noise than analog signals.
• Signal can be regenerated (not just amplified) at repeater stations, removing accumulated noise
and distortion over long transmission distances.
• Easy to multiplex multiple PCM channels using Time Division Multiplexing (TDM).
• Compatible with digital signal processing, storage, encryption, and error correction techniques.
• Uniform format regardless of the type of source (voice, video, data), enabling integrated digital
networks.
Disadvantages:
• Requires a much larger transmission bandwidth compared to analog modulation of the same
message signal.
• Increased system complexity due to the need for sampling, quantizing, and encoding circuitry.
• Introduces unavoidable quantization noise/error, which is not present in ideal analog transmission
(theoretically).
B.12 Applications of PCM
• Digital telephone networks (e.g., the standard 64 kbps PCM voice channel using 8 kHz sampling
and 8-bit encoding).
• Compact Disc (CD) audio, which uses PCM with 44.1 kHz sampling and 16-bit encoding.
• Digital audio and video broadcasting.
• Satellite and space communication systems.
• Digital storage of any analog signal for later playback or processing.
B.13 Comparison: PCM vs DPCM vs Delta Modulation
Feature PCM DPCM Delta Modulation (DM)
Difference between Single bit indicating
What is encoded Actual sample amplitude
sample and prediction rise or fall
Bits per sample Multiple (e.g., 8) Fewer than PCM Only 1 bit
Bandwidth required Highest Lower than PCM Lowest
Complexity Moderate Higher (needs predictor) Simple
Typical use Telephone, CD audio Speech/video compression Simple voice/telemetry links
B.14 Worked Numerical Examples
Example 1: Telephone Voice Channel
A telephone voice signal is band-limited to fm = 4 kHz. It is sampled at the standard rate of fs = 8
kHz and encoded using n = 8 bits per sample. Find the bit rate.
Solution: Bit Rate = fs × n = 8000 × 8 = 64,000 bits per second = 64 kbps.
This is the well-known standard PCM voice channel rate used worldwide in digital telephony.
Example 2: SQNR Calculation
Find the SQNR for a PCM system using n = 10 bits per sample.
Solution: SQNR = 6.02n + 1.76 = 6.02(10) + 1.76 = 60.2 + 1.76 = 61.96 dB.
Example 3: Quantization Levels and Step Size
A signal has a peak-to-peak voltage range of 10 V and is quantized using n = 4 bits. Find the
number of quantization levels and the step size.
Solution: L = 24 = 16 levels. Step size ∆ = 10 V / 16 = 0.625 V.
B.15 Summary and Key Formulas
Quantity Formula
Nyquist Sampling Rate f_s >= 2 f_m
Number of Quantization Levels L = 2^n
Step Size Delta = V_pp / L
Maximum Quantization Error +/- Delta / 2
SQNR (approx.) 6.02n + 1.76 dB
Bit Rate R = f_s x n
Minimum Bandwidth B >= R / 2
End of Notes. These theory notes, combined with the earlier 4-period lesson plan, provide a
complete package for teaching Pulse Code Modulation from the ground up, including all necessary
prerequisite background.