📚 The Two-Sided Fourier Series | 双边傅里叶级数
In signal processing, differential equations and modern physics, representing a periodic function as a sum of simple oscillating functions is a fundamental technique. The classical Fourier series uses sines and cosines, but an alternative formulation—the two-sided or complex exponential Fourier series—offers a more compact and algebraically elegant representation. By allowing both positive and negative frequency indices in a single summation, this form unifies the amplitude and phase information of each harmonic and becomes the natural language for the frequency spectrum of a signal. This article explores the definition, derivation, properties and applications of the two-sided Fourier series, bridging the gap between the real trigonometric series and its complex counterpart.
在信号处理、微分方程和现代物理学中，将周期函数表示为简单振荡函数之和是一项基本技术。经典的傅里叶级数使用正弦和余弦函数，但另一种形式——双边或复指数傅里叶级数——提供了一种更紧凑且代数上更优雅的表示。通过允许在一个求和式中同时出现正、负频率指标，这种形式统一了每个谐波的振幅和相位信息，并成为信号频谱的自然语言。本文探讨双边傅里叶级数的定义、推导、性质和应用，架起实数三角级数与其复数对应形式之间的桥梁。
Published by TutorHao | Mathematics Revision Series | aleveler.com
1. Periodic Functions and Orthogonal Bases | 周期函数与正交基
basis consists of {1, cos(nω₀t), sin(nω₀t)} for integer n ≥ 1. These functions satisfy orthogonality relations: the integral over [0, T] of the product cos(nω₀t)cos(mω₀t) is T/2 if n=m and 0 otherwise; similarly for sin(nω₀t)sin(mω₀t); and cos(nω₀t)sin(mω₀t) always integrates to 0. This orthogonality allows independent determination of each coefficient.
若对所有 t 有 f(t+T) = f(t)，则函数 f(t) 是周期为 T 的周期函数。基频为 ω₀ = 2π/T。傅里叶分析的核心思想是将 f(t) 表示为一组基本周期函数的线性组合，这些函数在一个周期内相互正交。对于实值函数，自然基底为 {1, cos(nω₀t), sin(nω₀t)}（整数 n ≥ 1）。这些函数满足正交关系：在一个周期 [0, T] 上，cos(nω₀t)cos(mω₀t) 的积分当 n=m 时为 T/2，否则为 0；sin(nω₀t)sin(mω₀t) 类似；而 cos(nω₀t)sin(mω₀t) 的积分始终为 0。这种正交性使得每个系数可被独立确定。
2. Real Trigonometric Fourier Series | 实数三角傅里叶级数
The real Fourier series of a periodic function f(t) with period T is written as:
f(t) = a₀/2 + Σₙ₌₁∞ [aₙ cos(nω₀t) + bₙ sin(nω₀t)]
where a₀/2 represents the average value (DC component), and the coefficients are given by:
aₙ = (2/T) ∫ₐT f(t) cos(nω₀t) dt, bₙ = (2/T) ∫ₐT f(t) sin(nω₀t) dt
This form clearly separates even and odd components, but the algebra becomes unwieldy when we need to combine sines and cosines, for example in solving linear differential equations or when multiplying series. Moreover, the presence of two separate coefficient sequences (aₙ, bₙ) obscures the symmetry of the frequency spectrum.
周期为 T 的周期函数 f(t) 的实数傅里叶级数写作：
f(t) = a₀/2 + Σₙ₌₁∞ [aₙ cos(nω₀t) + bₙ sin(nω₀t)]
其中 a₀/2 表示平均值（直流分量），系数由下式给出：
aₙ = (2/T) ∫ₐT f(t) cos(nω₀t) dt, bₙ = (2/T) ∫ₐT f(t) sin(nω₀t) dt
这种形式清楚地分离了偶分量和奇分量，但代数运算变得繁琐，例如在求解线性微分方程或级数相乘时。此外，两个独立的系数序列 (aₙ, bₙ) 掩盖了频谱的对称性。
3. Motivating the Complex Exponential Form | 引入复指数形式的动机
Euler' tituting these into the real series produces a sum of terms involving eⁱⁿω₀t and e⁻ⁱⁿω₀t. If we define a new sequence of coefficients cₙ that includes both positive and negative indices n, the entire series collapses into a single compact sum. This two-sided representation is not only mathematically elegant but also reveals that the spectrum extends symmetrically to negative frequencies, a concept essential in communications theory and quantum mechanics.
欧拉公式 eⁱⁿω₀t = cos(nω₀t) + i sin(nω₀t) 及其共轭 e⁻ⁱⁿω₀t = cos(nω₀t) − i sin(nω₀t) 允许我们用复指数替代三角函数。由于 cos(nω₀t) = (eⁱⁿω₀t + e⁻ⁱⁿω₀t)/2 且 sin(nω₀t) = (eⁱⁿω₀t − e⁻ⁱⁿω₀t)/(2i)，将这些代入实数级数便产生包含 eⁱⁿω₀t 和 e⁻ⁱⁿω₀t 的项之和。如果我们定义一个新的系数序列 cₙ，使其同时包含正、负 n，整个级数便合并为一个紧凑的和式。这种双边表示不仅在数学上优雅，还揭示了频谱对称地延伸到负频率，这一概念在通信理论和量子力学中至关重要。
4. The Orthogonal Basis of Complex Exponentials | 复指数正交基
The set of functions {eⁱⁿω₀t : n ∈ ℤ} forms an orthogonal basis on any interval of length T. Their inner product is defined as the integral of the product of one function and the complex conjugate of another over one period:
⟨eⁱⁿω₀t, eⁱᵐω₀t⟩ = ∫ₐT eⁱⁿω₀t e⁻ⁱᵐω₀t dt = ∫ₐT eⁱ⁽ⁿ⁻ᵐ⁾ω₀t dt
herwise, because the average of a complex exponential over an integer number of cycles is zero. Thus, the exponentials are orthogonal with respect to the Hermitian inner product. This orthogonality property is the key to extracting the coefficients cₙ.
函数集 {eⁱⁿω₀t : n ∈ ℤ} 在任意长度为 T 的区间上构成一组正交基。其内积定义为在一个周期内一个函数与另一个函数的复共轭之积的积分：
⟨eⁱⁿω₀t, eⁱᵐω₀t⟩ = ∫ₐT eⁱⁿω₀t e⁻ⁱᵐω₀t dt = ∫ₐT eⁱ⁽ⁿ⁻ᵐ⁾ω₀t dt
当 n = m 时该积分等于 T，否则等于 0，因为复指数在整数个周期上的平均值为零。因此，这些指数函数在埃尔米特内积下是正交的。这一正交性质是提取系数 cₙ 的关键。
5. Definition: The Two-Sided Fourier Series | 定义：双边傅里叶级数
For a T-periodic function f(t) that is integrable over one period, the two-sided (complex exponential) Fourier series representation is:
f(t) = Σₙ₌₋∞∞ cₙ eⁱⁿω₀t
where each complex coefficient cₙ is given by the analysis equation:
cₙ = (1/T) ∫ₐT f(t) e⁻ⁱⁿω₀t dt
or [0, T]. The series sums over all integers n, both positive and negative, hence the name “two-sided.” e conjugate symmetry c₋ₙ = cₙ* (complex conjugate), ensuring the overall sum remains real.
对于一个周期内可积的 T-周期函数 f(t)，双边（复指数）傅里叶级数表示为：
f(t) = Σₙ₌₋∞∞ cₙ eⁱⁿω₀t
其中每个复系数 cₙ 由分析方程给出：
cₙ = (1/T) ∫ₐT f(t) e⁻ⁱⁿω₀t dt
积分在任意完整周期上进行，例如 [−T/2, T/2] 或 [0, T]。该级数对所有整数 n（正和负）求和，因此得名“双边”。直流分量对应 n = 0：c₀ = (1/T) ∫ₐT f(t) dt，即平均值。对于实值 f(t)，系数满足共轭对称性 c₋ₙ = cₙ*（复共轭），以确保总和仍为实数。
6. Relating Real and Complex Coefficients | 实数系数与复系数的关系
By comparing the real trigonometric form with the two-sided form and using Euler’s identities, we obtain the conversion formulas. For n > 0:
cₙ = (aₙ − i bₙ)/2, c₋ₙ = (aₙ + i bₙ)/2
ₙ = cₙ + c₋ₙ, and bₙ = i(cₙ − c₋ₙ). The amplitude of the n-th harmonic, often denoted Aₙ = √aₙ² + bₙ², is related to the magnitude of the complex coefficients by |cₙ| = |c₋ₙ| resentation folds the amplitude and phase into a single complex number for each frequency.
通过比较实数三角形式和双边形式并利用欧拉恒等式，我们得到转换公式。对于 n > 0：
cₙ = (aₙ − i bₙ)/2, c₋ₙ = (aₙ + i bₙ)/2
当 n = 0 时：c₀ = a₀/2。反之，aₙ = cₙ + c₋ₙ，bₙ = i(cₙ − c₋ₙ)。第 n 次谐波的振幅常记为 Aₙ = √aₙ² + bₙ²，与复系数模长的关系为 |cₙ| = |c₋ₙ| = Aₙ/2。该谐波的相位角 φₙ（满足 aₙ cos + bₙ sin = Aₙ cos(nω₀t − φₙ)）当 n > 0 时由 φₙ = arg(cₙ) 给出（符号约定可能取决于定义）。这展示了复表示如何将振幅和相位折叠为每个频率的单一复数。
7. Amplitude and Phase Spectrum | 振幅谱与相位谱
Plotting |cₙ| versus n (or versus frequency nω₀) produces the amplitude spectrum. Because |c₋ₙ| = |cₙ| , giving a symmetric two-sided spectrum. In practice, the one-sided spectrum (only n ≥ 0) is often displayed for physical interpretation, but the two-sided spectrum is preferred in calculation and when dealing with complex-valued signals or modulations.
绘制 |cₙ| 对 n（或对频率 nω₀）的图即得振幅谱。由于 |c₋ₙ| = |cₙ|，实信号的振幅谱是 n 的偶函数。相位谱绘制 arg(cₙ) 对 n，为奇函数：arg(c₋ₙ) = −arg(cₙ)（假设 cₙ 的辐角对应于余弦表示的相位延迟）。这些图延伸至负 n，给出对称的双边谱。实际中，单边谱（仅 n ≥ 0）常用于物理解释，但双边谱在计算以及处理复值信号或调制时更受青睐。
8. Convergence and Dirichlet Conditions | 收敛性与狄利克雷条件
onditions: (1) f is absolutely integrable over one period; (2) f has a finite number of maxima and minima in any finite interval; (3) f has a finite number of discontinuities in any finite interval, all of which are finite. At a jump discontinuity, the series converges to the average of the left- and right-hand limits. Gibbs' phenomenon—overshoot near discontinuities—persists in the complex form as well. The complex coefficients cₙ decay as |n| the rate of decay depends on the smoothness of f. If f has k continuous derivatives, then |cₙ| = O(1/|n|ᵏ⁺¹).
与三角级数一样，若 f 满足狄利克雷条件，双边傅里叶级数在连续点收敛到 f(t)：(1) f 在一个周期内绝对可积；(2) 在任何有限区间内 f 只有有限个极大值和极小值；(3) 在任何有限区间内 f 只有有限个不连续点，且均为有限跳跃。在跳跃间断点处，级数收敛到左右极限的平均值。吉布斯现象——间断点附近的过冲——在复数形式中同样存在。复系数 cₙ 随 |n| 增大而衰减；衰减速率取决于 f 的光滑性。若 f 有 k 阶连续导数，则 |cₙ| = O(1/|n|ᵏ⁺¹)。
9. Example: Square Wave | 示例：方波
y f(t) = A for 0 < t < T/2 and f(t) = −A for −T/2 < t < 0. The two-sided coefficients are:
cₙ = (1/T) [ ∫₀ᵀ⁄² A e⁻ⁱⁿω₀t dt + ∫₋ₜ⁄₂₀ (−A) e⁻ⁱⁿω₀t dt ]
4A/(π(2k+1))) sin((2k+1)ω₀t). Note that the complex coefficients are purely imaginary, reflecting the odd symmetry. The amplitude spectrum |cₙ| is proportional to 1/|n| for odd n, illustrating the slow decay due to the discontinuity.
考虑一个奇对称的周期为 T 的方波，定义为 0 < t < T/2 时 f(t) = A，−T/2 < t < 0 时 f(t) = −A。双边系数为：
cₙ = (1/T) [ ∫₀ᵀ⁄² A e⁻ⁱⁿω₀t dt + ∫₋ₜ⁄₂₀ (−A) e⁻ⁱⁿω₀t dt ]
计算得 n ≠ 0 时 cₙ = (A/(iπn)) (1 − (−1)ⁿ)，c₀ = 0。因此 n 为奇数时 cₙ = 2A/(iπn)，偶数时为 0。级数为 f(t) = Σₙ 奇数 (2A/(iπn)) eⁱⁿω₀t。合并 n 与 −n 项即得熟悉的正弦级数 f(t) = Σₖ₌₀∞ (4A/(π(2k+1))) sin((2k+1)ω₀t)。注意复系数为纯虚数，反映了奇对称性。振幅谱 |cₙ| 对奇数次正比于 1/|n|，显示了因间断导致的缓慢衰减。
10. Parseval’s Theorem and Energy | 帕塞瓦尔定理与能量
Parseval’s theorem for the two-sided Fourier series relates the average power of the signal to the sum of squared magnitudes of the coefficients:
(1/T) ∫ₐT |f(t)|² dt = Σₙ₌₋∞∞ |cₙ|²
onjugate symmetry, it reduces to the one-sided version: (1/T) ∫ₐT f(t)² dt = c₀² + 2 Σₙ₌₁∞ |cₙ| ions of all harmonics, and it plays a central role in filter design and spectral analysis.
双边傅里叶级数的帕塞瓦尔定理将信号的平均功率与系数模长的平方和联系起来：
(1/T) ∫ₐT |f(t)|² dt = Σₙ₌₋∞∞ |cₙ|²
此恒等式对于理解能量在频率分量间的分布十分宝贵。对于实函数 f，利用共轭对称性可化简为单边形式：(1/T) ∫ₐT f(t)² dt = c₀² + 2 Σₙ₌₁∞ |cₙ|²。帕塞瓦尔定理提供了一种通过时域积分或求和所有谐波贡献来计算总功率的方法，在滤波器设计和频谱分析中占据核心地位。
11. Applications in Linear Systems and Communication | 在线性系统与通信中的应用
The two-sided Fourier series is particularly advantageous when analyzing the response of a linear time-invariant (LTI) system to a periodic input. Since complex exponentials are eigenfunctions of LTI systems, the output for each term cₙ eⁱⁿω₀t is simply cₙ H(nω₀) eⁱⁿω₀t, where H(ω) is the system's frequency response. The superposition principle then yields the output series. In amplitude modulation (AM) and other communication schemes, the two-sided spectrum naturally represents the shifted replicas of the baseband signal around the carrier frequency, with negative frequencies providing a mirror image that completes the physical description of sidebands.
在分析线性时不变（LTI）系统对周期输入的响应时，双边傅里叶级数尤为有利。由于复指数是 LTI 系统的特征函数，每一项 cₙ eⁱⁿω₀t 的输出即为 cₙ H(nω₀) eⁱⁿω₀t，其中 H(ω) 为系统的频率响应。叠加原理进而给出输出级数。在幅度调制（AM）及其他通信方案中，双边频谱自然地表示基带信号在载频附近的移位复制品，负频率提供镜像，从而完成对边带的物理描述。
12. Summary and Final Remarks | 总结与结语
Σₙ cₙ eⁱⁿω₀t, distils the essence of harmonic analysis into a single compact sum over all integer frequencies. It eliminates the need for separate sine and cosine terms, simplifies algebraic manipulations, and directly yields the amplitude and phase spectrum via complex coefficients. While the series converges to the original function under mild conditions, its true power lies in its capacity to unify time-domain and frequency-domain perspectives. For IB Mathematics students, mastering this form opens the door to deeper studies in Fourier transforms, signal processing, and applied differential equations.
双边傅里叶级数 f(t) = Σₙ cₙ eⁱⁿω₀t 将谐波分析的精髓浓缩为一个跨所有整数频率的单一紧凑求和。它消除了对单独正弦和余弦项的需求，简化了代数运算，并通过复系数直接给出了振幅和相位谱。虽然级数在温和条件下收敛到原函数，但其真正的威力在于统一时域和频域视角的能力。对于 IB 数学学生而言，掌握此形式为深入学习傅里叶变换、信号处理和应用微分方程打开了大门。
Published by TutorHao | Mathematics Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图