Bandwidth

Introduction

Bandwidth is one of the most frequently used terms in signal processing and communications, yet it has multiple definitions depending on context. At its core, bandwidth describes the range of frequencies that a signal occupies or that a system can pass. A wider bandwidth means more frequency content (more information capacity for communications, finer time resolution for radar) but also requires more resources (wider channels, faster sampling, more noise).

Understanding the various bandwidth definitions and their appropriate applications is essential because using the wrong definition can lead to system designs that either waste resources or fail to meet specifications.

Bandwidth Definitions

3-dB (Half-Power) Bandwidth

The most common engineering definition: the frequency range over which the power spectral density (or system gain squared) remains within 3 dB of its peak value:

$$B_{3dB} = f_2 - f_1$$

where $|H(jf_1)|^2 = |H(jf_2)|^2 = |H_{max}|^2 / 2$ (half the peak power).

At the 3-dB frequencies, the amplitude response drops to $1/\sqrt{2} \approx 0.707$ of its maximum. This is the standard definition for filters, amplifiers, and communication channels.

Null-to-Null (Main Lobe) Bandwidth

The frequency span between the first nulls (zeros) on either side of the main spectral lobe. For a rectangular pulse of duration $T$: $$B_{null} = 2/T$$

Commonly used for radar pulses and modulated signals.

Equivalent Noise Bandwidth (ENBW)

The width of an ideal rectangular filter that would pass the same total noise power as the actual system: $$B_N = \frac{\int_0^{\infty}|H(jf)|^2 df}{|H_{max}|^2}$$

Used in noise calculations because it directly relates system bandwidth to noise power: $P_{noise} = N_0 \cdot B_N$.

Percent Bandwidth (Fractional Bandwidth)

$$B_{\%} = \frac{f_H - f_L}{f_c} \times 100\%$$

where $f_c = (f_H + f_L)/2$ is the center frequency. Used for bandpass systems to express bandwidth relative to the operating frequency.

Essential Bandwidth

The frequency range containing a specified percentage (e.g., 99%) of the total signal energy: $$\int_0^{B_{ess}} |X(jf)|^2 df = 0.99 \int_0^{\infty}|X(jf)|^2 df$$

Bandwidth of Common Signals

The Time-Bandwidth Product

The product of signal duration and bandwidth is bounded below: $$T \cdot B \geq k$$

where $k$ is a constant that depends on the definitions used (typically around 1). This expresses the uncertainty principle: a signal cannot be simultaneously short in time and narrow in frequency.

A Gaussian pulse achieves the theoretical minimum time-bandwidth product ($TBP = 0.5$ using RMS definitions), making it the most "efficient" waveform.

Bandwidth and Information Capacity

Shannon\'s channel capacity theorem relates bandwidth to information rate: $$C = B \log_2(1 + SNR) \text{ bits/second}$$

Doubling the bandwidth doubles the capacity (for fixed SNR). This is why wideband technologies (5G, fiber optics) can carry more data than narrowband ones (telephone lines).

System Bandwidth vs Signal Bandwidth

System bandwidth describes the range of frequencies a system can process without significant attenuation. A system\'s 3-dB bandwidth must exceed the signal\'s bandwidth for undistorted transmission:

$$B_{system} \geq B_{signal}$$

If the system bandwidth is too narrow, high-frequency components are attenuated (the signal is distorted/smoothed).

Bandwidth in Digital Systems

For digital signals at bit rate $R_b$:

• Minimum (Nyquist) bandwidth: $B_{min} = R_b/2$ Hz (for binary signaling with ideal Nyquist pulse shaping)
• Practical bandwidth: $B \approx R_b$ Hz (for commonly used pulse shapes)
• Spectral efficiency: $\eta = R_b/B$ bits/s/Hz

Key Takeaways

• Bandwidth measures the frequency range occupied by a signal or passable by a system
• Multiple definitions exist: 3-dB, null-to-null, noise equivalent, essential — use the appropriate one
• Time-bandwidth product is bounded: short signals have wide bandwidth and vice versa
• Shannon capacity: $C = B\log_2(1+SNR)$ — bandwidth directly limits information rate
• System bandwidth must exceed signal bandwidth for faithful transmission
• The 3-dB (half-power) definition is the most commonly used in filter and amplifier specifications