7. Convolution Theorem for Fourier Transforms | 傅里叶变换的卷积定理

📚 7. Convolution Theorem for Fourier Transforms | 傅里叶变换的卷积定理

Within the IB Mathematics: Analysis and Approaches HL course, the study of integral transforms often begins with Fourier series. Extending these ideas, the Fourier transform handles non‑periodic functions and forms the backbone of modern signal processing, differential equations and many areas of physics. One of its most elegant and practical properties is the convolution theorem, which reveals a deep algebraic link between convolution in the time domain and multiplication in the frequency domain. Mastery of this theorem not only simplifies calculations but also provides powerful insight into linear systems and filtering.

在 IB 数学：分析与方法 HL 课程中，从傅里叶级数出发很自然地会接触到积分变换。傅里叶变换处理非周期函数，成为现代信号处理、微分方程和物理学诸多领域的基石。其最优雅、最实用的性质之一就是卷积定理——它揭示了时域卷积与频域乘法之间的深层代数联系。掌握这个定理不仅能简化大量计算，还能为理解线性系统和滤波提供强有力的视角。

---

1. Recalling the Fourier Transform | 回顾傅里叶变换的定义

A common convention for the Fourier transform of an absolutely integrable function f(t) is given by F(ω) = ∫_-∞^∞ f(t) e^-iωt dt. The inverse transform then reconstructs f(t) as (1/2π)∫_-∞^∞ F(ω) e^iωt dω. Other normalisations exist, but this asymmetric form highlights the 2π factor that will resurface in the convolution theorem.

傅里叶变换的一种常用约定是：F(ω) = ∫_-∞^∞ f(t) e^-iωt dt，其中函数 f(t) 绝对可积。相应地，逆变换为 (1/2π)∫_-∞^∞ F(ω) e^iωt dω。这种不对称形式突显了因子 2π，它将在卷积定理中以对称的方式重新出现。

When the independent variable represents time, ω is interpreted as angular frequency. In IB language, we are mapping from the time domain to the frequency domain, and the transform is linear: F{a f + b g} = aF{f} + bF{g}.

当自变量代表时间时，ω 表示角频率。用 IB 课程的语言来说，我们把函数从时域映射到频域，并且傅里叶变换是线性的：F{a f + b g} = aF{f} + bF{g}。

---

2. Defining Convolution | 卷积的定义

For two functions f(t) and g(t), their convolution is a new function defined by the integral (f * g)(t) = ∫_-∞^∞ f(τ) g(t - τ) dτ. Geometrically, this operation slides one function across the other, multiplies the overlapping values and integrates the product. Commutativity (f * g = g * f), associativity and distributivity over addition make convolution an algebraic structure similar to multiplication.

对于两个函数 f(t) 与 g(t)，它们的卷积定义为 (f * g)(t) = ∫_-∞^∞ f(τ) g(t - τ) dτ。在几何上，该操作使一个函数滑过另一个，将重叠部分的乘积积分。卷积满足交换律 (f * g = g * f)、结合律以及关于加法的分配律，因而在代数上颇似乘法。

---

3. The Convolution Theorem: Main Statement | 卷积定理的正式陈述

Under mild integrability conditions, the Fourier transform of a convolution equals the product of the individual Fourier transforms:

F{ (f * g)(t) } = F(ω) G(ω).

在适当的可积条件下，卷积的傅里叶变换等于各自傅里叶变换的乘积：

F{ (f * g)(t) } = F(ω) G(ω)。

Here, F(ω) = F{f} and G(ω) = F{g}. The beauty of the theorem is that a complicated integral (convolution) in the time domain turns into simple pointwise multiplication in the frequency domain.

这里 F(ω) = F{f}，G(ω) = F{g}。定理的美妙之处在于：时域中复杂的卷积积分在频域中变成了朴素的逐点乘法。

---

4. Proof Outline via Fubini’s Theorem | 借助富比尼定理的证明框架

Start with the definition of the Fourier transform of the convolution:

F{ f * g } = ∫_-∞^∞ [∫_-∞^∞ f(τ) g(t - τ) dτ] e^-iωt dt.

Assuming absolute integrability, the order of integration can be swapped:

= ∫_-∞^∞ f(τ) [∫_-∞^∞ g(t - τ) e^-iωt dt] dτ.

The inner integral, after the substitution u = t - τ, becomes e^-iωτ G(ω). Factorising this out leaves ∫_-∞^∞ f(τ) e^-iωτ dτ = F(ω), completing the proof.

从卷积的傅里叶变换定义出发：

F{ f * g } = ∫_-∞^∞ [∫_-∞^∞ f(τ) g(t - τ) dτ] e^-iωt dt。

在绝对可积的假设下，交换积分顺序得到：

= ∫_-∞^∞ f(τ) [∫_-∞^∞ g(t - τ) e^-iωt dt] dτ。

做代换 u = t - τ，内层积分化为 e^-iωτ G(ω)。将其提出，剩下的积分正是 ∫_-∞^∞ f(τ) e^-iωτ dτ = F(ω)，从而证毕。

---

5. The Dual Form: Product to Convolution | 对偶形式：时域乘积的傅里叶变换

By symmetry of the transform pair, the theorem also works in reverse. The Fourier transform of a product of two functions is (1/2π) times the convolution of their individual transforms:

F{ f(t) g(t) } = (1/2π) (F * G)(ω).

This dual statement is crucial for analysing amplitude modulation or windowing in signal processing.

由于变换具有对称性，该定理也可反向应用。两个函数乘积的傅里叶变换等于各自变换的卷积再乘以 (1/2π)：

F{ f(t) g(t) } = (1/2π) (F * G)(ω)。

这对偶形式在分析调幅或加窗等信号处理问题中起着不可或缺的作用。

---

广告 / Sponsored

6. Handling Alternative Normalisations | 处理不同归一化约定

If the Fourier transform is defined symmetrically with a factor 1/√2π in both forward and inverse transforms, the convolution theorem becomes equally symmetric: F{f * g} = √2π F{f} · F{g}. Students must always check which definition is being used — small coefficient differences can cause confusion in IB exploration style problems.

若傅里叶变换使用对称归一化（正、逆变换均含因子 1/√2π），卷积定理也变得完全对称：F{f * g} = √2π F{f} · F{g}。同学们务必核对教材所采用的定义——系数上的细微差别在 IB 探索型问题中常常导致混乱。

The key is consistency: once a convention is chosen, the inverse transform and convolution theorem must be stated accordingly. A wise strategy in IB internal assessment is to declare the chosen definition at the very beginning.

关键是保持一致性：一旦选定一种约定，逆变换与卷积定理都必须相应地陈述。在 IB 内部评估中，一个明智的做法是在开头就明确所用的定义。

---

7. Application to Linear Differential Equations | 在线性微分方程中的应用

Because differentiation in the time domain becomes multiplication by iω in the frequency domain, the convolution theorem helps solve linear ODEs with a forcing term. Taking the Fourier transform of both sides of an equation like y'' + 2y' + y = f(t) yields an algebraic relation for Y(ω) = H(ω) F(ω), where H(ω) is the system’s frequency response. Applying the inverse transform then gives the solution as a convolution: y(t) = (h * f)(t).

由于时域的求导操作在频域变为乘以 iω，卷积定理可用于求解带激励项的线性常微分方程。对形如 y'' + 2y' + y = f(t) 的方程两边取傅里叶变换，得到 Y(ω) = H(ω) F(ω)，其中 H(ω) 是系统的频率响应。再施以逆变换，解就写成卷积形式：y(t) = (h * f)(t)。

This technique underpins an entire family of “Green’s function” methods and is directly applicable to IB extended essay topics on vibrations or electric circuits.

这一手法是整个格林函数方法族的基础，并可直接应用于 IB 拓展论文中涉及振动或电路的主题。

---

8. Signal Filtering and Impulse Response | 信号滤波与冲激响应

In signal processing, a linear time‑invariant (LTI) system is completely described by its impulse response h(t). The output y(t) for an input x(t) is the convolution y = h * x. By the convolution theorem, the output spectrum is simply Y(ω) = H(ω) X(ω). Designing a filter thus becomes a matter of choosing an appropriate transfer function H(ω) and implementing it via convolution in the time domain.

在信号处理中，线性时不变 (LTI) 系统完全由其冲激响应 h(t) 描述。对输入 x(t) 的输出即为卷积 y = h * x。根据卷积定理，输出频谱简单地为 Y(ω) = H(ω) X(ω)。因此，滤波器设计实质上就是选取合适的传递函数 H(ω)，并通过时域卷积加以实现。

For IB students, experimenting with low‑pass, high‑pass or band‑pass filters provides a concrete context for applying the convolution theorem, often using simple rect or sinc functions.

对 IB 同学而言，用矩形函数或 sinc 函数制作低通、高通或带通滤波器，是应用卷积定理的具体实例，极具直观性。

---

9. Connection with Discrete Fourier Transforms | 与离散傅里叶变换的联系

The convolution theorem also holds in the discrete setting, where it becomes the statement that circular convolution of two sequences corresponds to pointwise multiplication of their DFTs. This is the engine behind fast convolution algorithms using the FFT, allowing large‑scale filtering to be performed efficiently — a topic that can make an excellent IB Mathematics HL internal exploration.

卷积定理在离散情境下同样成立：两个序列的循环卷积对应于它们离散傅里叶变换的逐点乘积。这正是利用 FFT 进行快速卷积算法的核心，使得大规模滤波能够高效实现——这一主题非常适合作为 IB 数学高水平课程内部评估的选题。

Understanding the continuous Fourier transform theorem thus serves as the theoretical foundation for its discrete cousin, which is ubiquitous in computational mathematics.

因此，理解连续傅里叶变换的卷积定理，也为无处不在的计算数学中离散版本的学习奠定了理论基础。

---

10. Comparison with Fourier Series Convolution | 与傅里叶级数卷积的比较

In the periodic world, convolution is often defined over one period, and the corresponding Fourier series coefficients obey an analogous product rule: the convolution of two periodic functions has Fourier coefficients equal to the product of the individual coefficients (up to a factor). While the transform handles continuous spectra, the series handles discrete spectra — the algebraic structure remains remarkably similar.

在周期性函数的世界里，卷积往往定义在一个周期上，相应的傅里叶级数系数也满足类似乘积法则：两个周期函数卷积的傅里叶系数等于各自系数的乘积（可能相差一个因子）。尽管变换处理连续谱、级数处理离散谱，其代数结构却惊人地一致。

This parallel helps IB students who have already studied Fourier series to see a natural generalisation and to appreciate the power of the convolution concept across different domains.

这种类比可帮助已学过傅里叶级数的 IB 学生看到自然的推广，并体会卷积概念在不同领域中的强大威力。

---

11. Common Pitfalls and How to Avoid Them | 常见误区与避免策略

Forgetting the 2π factor. Many students apply the convolution theorem without adjusting for the chosen normalisation. Always write the theorem in the form matching your definition and re‑derive it if uncertain.

忽略因子 2π。 很多同学在不调整归一化约定的情况下直接套用定理。务必将定理写成与所选定义一致的形式，拿不准时重新推导一遍。

Mixing domains. The product F(ω)G(ω) lives in the frequency domain; do not confuse it with the original time‑domain product.

混淆域。 F(ω)G(ω) 是频域的乘积，不要与原始的时域乘积弄混。

Ignoring convergence conditions. Although IB exam questions rarely probe delicate convergence, in an exploration it is wise to note that both f and g should be absolutely integrable for the standard proof to hold.

忽略收敛条件。 虽然 IB 考试很少深究微妙的收敛性，但在撰写探索性作业时指出 f 与 g 均需绝对可积方能使标准证明成立，会体现出严谨的数学素养。

---

12. Practice and Deeper Understanding | 练习与深化理解

To truly internalise the convolution theorem, IB learners should attempt a few concrete examples: compute the convolution of two rectangular pulses, then verify that its Fourier transform equals the product of the individual sinc transforms. Such hands‑on work reveals the elegance of the theorem and cements the link between integral transform theory and real‑world applications.

为了真正内化卷积定理，IB 学习者应当亲手演算几个具体例子：计算两个矩形脉冲的卷积，然后验证其傅里叶变换确实等于各自 sinc 变换的乘积。这种动手实操能让定理的优雅性跃然纸上，并强化积分变换理论与实际应用之间的联系。

Pursuing this topic further can open doors to Green’s functions, partial differential equations, probability theory and even modern machine learning architectures, making it a rewarding addition to any ambitious IB mathematics portfolio.

进一步深究这个主题，还可通向格林函数、偏微分方程、概率论乃至现代机器学习架构，使之成为任何有抱负的 IB 数学作品集中极具价值的补充。

---

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）