The Central Idea
Suppose a function repeats with period \(2\pi\). A Fourier series seeks a representation of the form
\[ \boxed{ f(x)\sim \frac{a_0}{2} +\sum_{n=1}^{\infty} \left(a_n\cos nx+b_n\sin nx\right) }. \]The constant term describes the average level of the function. The cosine and sine terms describe oscillations at integer multiples of a fundamental frequency. The coefficients \(a_n\) and \(b_n\) measure how strongly each oscillation is present.
Periodic Functions
A function \(f\) is periodic with period \(T>0\) when
\[ \boxed{f(x+T)=f(x)} \]for every \(x\) where the relationship is defined.
If \(T\) is the smallest positive period, it is called the fundamental period.
Fourier series are especially useful because a repeating function can be understood by studying just one complete period.
The Trigonometric Building Blocks
On \([-\pi,\pi]\), the standard Fourier basis consists of
\[ 1,\quad \cos x,\quad \sin x,\quad \cos 2x,\quad \sin 2x,\quad\ldots \]More compactly, the basis functions are
\[ 1,\qquad \cos(nx),\qquad \sin(nx), \qquad n=1,2,3,\ldots \]Each larger value of \(n\) corresponds to a more rapidly oscillating component.
Why Sines and Cosines Work
The trigonometric basis functions are orthogonal on a complete period. For positive integers \(m\) and \(n\),
\[ \int_{-\pi}^{\pi}\cos(mx)\cos(nx)\,dx=0 \qquad\text{when }m\ne n, \] \[ \int_{-\pi}^{\pi}\sin(mx)\sin(nx)\,dx=0 \qquad\text{when }m\ne n, \]and every sine is orthogonal to every cosine:
\[ \boxed{ \int_{-\pi}^{\pi}\sin(mx)\cos(nx)\,dx=0 }. \]This orthogonality lets us isolate one Fourier coefficient at a time.
The Remaining Orthogonality Integrals
For every positive integer \(n\),
\[ \int_{-\pi}^{\pi}\cos(nx)\,dx=0, \qquad \int_{-\pi}^{\pi}\sin(nx)\,dx=0. \]The squared norms are
\[ \int_{-\pi}^{\pi}\cos^2(nx)\,dx=\pi, \] \[ \int_{-\pi}^{\pi}\sin^2(nx)\,dx=\pi, \]while the constant function satisfies
\[ \int_{-\pi}^{\pi}1^2\,dx=2\pi. \]The constant has a different norm, which explains the conventional factor \(a_0/2\) in the Fourier series formula.
The Fourier Coefficients on \([-\pi,\pi]\)
For a sufficiently well-behaved function on \([-\pi,\pi]\), the standard Fourier coefficients are
\[ \boxed{a_0=\frac{1}{\pi}\int_{-\pi}^{\pi}f(x)\,dx}, \] \[ \boxed{a_n=\frac{1}{\pi}\int_{-\pi}^{\pi}f(x)\cos(nx)\,dx}, \qquad n\ge 1, \]and
\[ \boxed{b_n=\frac{1}{\pi}\int_{-\pi}^{\pi}f(x)\sin(nx)\,dx}, \qquad n\ge 1. \]Where the Coefficient Formula Comes From
Begin formally with
\[ f(x)\sim \frac{a_0}{2} +\sum_{n=1}^{\infty} \left(a_n\cos nx+b_n\sin nx\right). \]To isolate \(a_m\), multiply by \(\cos(mx)\) and integrate over \([-\pi,\pi]\). Orthogonality eliminates the constant term, all sine terms, and all cosine terms except the one with \(n=m\). The result is
\[ \int_{-\pi}^{\pi}f(x)\cos(mx)\,dx=a_m\pi. \]Therefore,
\[ \boxed{ a_m=\frac{1}{\pi} \int_{-\pi}^{\pi}f(x)\cos(mx)\,dx }. \]The same argument with \(\sin(mx)\) gives the formula for \(b_m\).
The Constant Term Is the Average Level
The average value of \(f\) over one period \([-\pi,\pi]\) is
\[ \frac{1}{2\pi}\int_{-\pi}^{\pi}f(x)\,dx. \]Since
\[ a_0=\frac{1}{\pi}\int_{-\pi}^{\pi}f(x)\,dx, \]we have
\[ \boxed{ \frac{a_0}{2} =\frac{1}{2\pi}\int_{-\pi}^{\pi}f(x)\,dx }. \]Thus the Fourier series separates the average level from the oscillatory components.
Partial Sums
An infinite Fourier series is usually approximated computationally by a finite partial sum. The \(N\)-th partial sum is
\[ \boxed{ S_N(x)=\frac{a_0}{2} +\sum_{n=1}^{N} \left(a_n\cos nx+b_n\sin nx\right) }. \]Increasing \(N\) includes more frequencies. When the Fourier series converges appropriately, the partial sums provide increasingly detailed approximations to the original function.
Fourier Series as Orthogonal Projection
The partial sum \(S_N\) is not merely an arbitrary trigonometric approximation. It is the orthogonal projection of \(f\) onto the finite-dimensional space
\[ \operatorname{span} \{1,\cos x,\sin x,\ldots,\cos Nx,\sin Nx\}. \]Consequently, among all trigonometric polynomials using those same frequencies, \(S_N\) minimizes the squared error
\[ \int_{-\pi}^{\pi}|f(x)-S_N(x)|^2\,dx. \]This is the least-squares interpretation of a Fourier partial sum.
Even Functions Simplify to Cosines
If \(f\) is even,
\[ f(-x)=f(x), \]then \(f(x)\sin(nx)\) is odd. Its integral over a symmetric interval is zero, so
\[ \boxed{b_n=0}. \]An even function therefore has a cosine series:
\[ f(x)\sim \frac{a_0}{2}+\sum_{n=1}^{\infty}a_n\cos(nx). \]Odd Functions Simplify to Sines
If \(f\) is odd,
\[ f(-x)=-f(x), \]then \(f(x)\cos(nx)\) is odd, so
\[ a_0=0 \qquad\text{and}\qquad \boxed{a_n=0}. \]An odd function therefore has a sine series:
\[ f(x)\sim \sum_{n=1}^{\infty}b_n\sin(nx). \]Symmetry Is More Than a Shortcut
Symmetry predicts which frequency components are allowed before any integration is performed. Even symmetry selects cosine components, while odd symmetry selects sine components.
Recognizing symmetry can therefore reduce computation and provide a strong error check. If an odd function produces a nonzero cosine coefficient, or an even function produces a nonzero sine coefficient, the calculation should be reexamined.
Example: Fourier Series for \(f(x)=x\)
Consider
\[ f(x)=x, \qquad -\pi<x<\pi, \]extended periodically with period \(2\pi\). Since \(f\) is odd,
\[ a_0=0, \qquad a_n=0. \]Only the sine coefficients remain:
\[ b_n=\frac{1}{\pi}\int_{-\pi}^{\pi}x\sin(nx)\,dx. \]Computing the Sine Coefficients for \(f(x)=x\)
The integrand \(x\sin(nx)\) is even, so
\[ b_n=\frac{2}{\pi}\int_0^{\pi}x\sin(nx)\,dx. \]Integrating by parts gives
\[ \int_0^{\pi}x\sin(nx)\,dx =-\frac{\pi(-1)^n}{n}. \]Therefore,
\[ \boxed{b_n=\frac{2(-1)^{n+1}}{n}}. \]The Resulting Sawtooth Series
The Fourier series for the periodic extension of \(f(x)=x\) is
\[ \boxed{ x\sim 2\sum_{n=1}^{\infty} \frac{(-1)^{n+1}}{n}\sin(nx) } \]for \(-\pi<x<\pi\). Writing the first few terms,
\[ x\sim 2\left( \sin x-\frac{1}{2}\sin 2x+ rac{1}{3}\sin 3x -\frac{1}{4}\sin 4x+\cdots \right). \]The coefficients decrease like \(1/n\), so higher-frequency terms contribute progressively smaller amplitudes.
Example: A Square Wave
Define the \(2\pi\)-periodic function by
\[ f(x)= \begin{cases} -1, & -\pi<x<0,\\ 1, & 0<x<\pi. \end{cases} \]This function is odd, so only sine terms appear. Its coefficients are
\[ b_n=\frac{2}{\pi}\int_0^{\pi}\sin(nx)\,dx. \]Evaluating the integral gives
\[ b_n=\frac{2}{n\pi}\left(1-(-1)^n\right). \]Only Odd Harmonics Remain
When \(n\) is even,
\[ 1-(-1)^n=0. \]When \(n\) is odd,
\[ 1-(-1)^n=2. \]Therefore the square wave has the Fourier series
\[ \boxed{ f(x)\sim \frac{4}{\pi} \left( \sin x+\frac{1}{3}\sin 3x+ rac{1}{5}\sin 5x+\cdots \right) }. \]The absence of even harmonics is a structural property of this particular symmetry, not a universal rule for all odd functions.
Frequency and Harmonics
For a \(2\pi\)-periodic function, the terms \(\sin(nx)\) and \(\cos(nx)\) oscillate \(n\) times as fast as \(\sin x\) and \(\cos x\).
The \(n=1\) component is the fundamental harmonic. The terms with \(n=2,3,4,\ldots\) are higher harmonics.
Fourier coefficients therefore provide a frequency-domain description of a periodic function even though the original function is specified in the variable \(x\), often interpreted as time or position.
Amplitude and Phase Form
A pair of terms at the same frequency,
\[ a_n\cos(nx)+b_n\sin(nx), \]can be combined into one shifted cosine:
\[ a_n\cos(nx)+b_n\sin(nx) =A_n\cos(nx-\phi_n), \]where
\[ \boxed{A_n=\sqrt{a_n^2+b_n^2}}. \]The phase \(\phi_n\) can be chosen so that
\[ a_n=A_n\cos\phi_n, \qquad b_n=A_n\sin\phi_n. \]This form separates the strength of a frequency component from its phase shift.
General Period \(2L\)
Not every periodic function has period \(2\pi\). Suppose the period is \(2L\) and we work on \([-L,L]\). The natural basis becomes
\[ \cos\left(\frac{n\pi x}{L}\right), \qquad \sin\left(\frac{n\pi x}{L}\right). \]The Fourier series is
\[ \boxed{ f(x)\sim \frac{a_0}{2} +\sum_{n=1}^{\infty} \left[ a_n\cos\left(\frac{n\pi x}{L}\right) +b_n\sin\left(\frac{n\pi x}{L}\right) \right] }. \]Coefficients for General Period \(2L\)
The corresponding coefficients are
\[ \boxed{ a_0=\frac{1}{L}\int_{-L}^{L}f(x)\,dx }, \] \[ \boxed{ a_n=\frac{1}{L}\int_{-L}^{L} f(x)\cos\left(\frac{n\pi x}{L}\right)\,dx }, \]and
\[ \boxed{ b_n=\frac{1}{L}\int_{-L}^{L} f(x)\sin\left(\frac{n\pi x}{L}\right)\,dx }. \]Setting \(L=\pi\) recovers the standard \([-\pi,\pi]\) formulas.
General Period \(T\)
If the fundamental period is \(T\), define the fundamental angular frequency
\[ \boxed{\omega_0=\frac{2\pi}{T}}. \]Then a Fourier series can be written
\[ f(t)\sim \frac{a_0}{2} +\sum_{n=1}^{\infty} \left[a_n\cos(n\omega_0 t)+b_n\sin(n\omega_0 t)\right]. \]The frequencies are integer multiples \(n\omega_0\) of the fundamental.
Half-Range Expansions
Sometimes a function is given only on \([0,L]\). We can still build a Fourier series by extending it to \([-L,L]\).
An even extension produces a Fourier cosine series, while an odd extension produces a Fourier sine series.
This flexibility is especially useful in boundary-value problems because the choice of sine or cosine basis can be matched to boundary conditions.
Half-Range Cosine Series
For an even extension from \([0,L]\), the cosine coefficients can be computed directly as
\[ a_0=\frac{2}{L}\int_0^L f(x)\,dx, \] \[ \boxed{ a_n=\frac{2}{L}\int_0^L f(x)\cos\left(\frac{n\pi x}{L}\right)\,dx }. \]The resulting series is
\[ f(x)\sim \frac{a_0}{2} +\sum_{n=1}^{\infty} a_n\cos\left(\frac{n\pi x}{L}\right). \]Half-Range Sine Series
For an odd extension from \([0,L]\), the sine coefficients are
\[ \boxed{ b_n=\frac{2}{L}\int_0^L f(x)\sin\left(\frac{n\pi x}{L}\right)\,dx }. \]The resulting series is
\[ f(x)\sim \sum_{n=1}^{\infty} b_n\sin\left(\frac{n\pi x}{L}\right). \]The original function on \([0,L]\) is the same in both approaches; the difference is how it is extended beyond that interval.
Convergence at Smooth Points
Under standard Fourier convergence conditions, if a periodic function is sufficiently regular near a point \(x\) and is continuous there, its Fourier series converges to
\[ \boxed{f(x)}. \]Piecewise smooth functions are a common practical setting. They may have a finite number of corners or jump discontinuities while still possessing a useful Fourier representation.
Convergence at a Jump
At a jump discontinuity, the Fourier series does not generally converge to either one-sided value. Instead, under standard conditions, it converges to the midpoint of the jump:
\[ \boxed{ \frac{f(x^-)+f(x^+)}{2} }. \]For the square wave, the left and right limits at a jump are \(-1\) and \(1\), so the Fourier series converges there to
\[ \frac{-1+1}{2}=0. \]Why the Symbol \(\sim\) Is Useful
We often write
\[ f(x)\sim \frac{a_0}{2} +\sum_{n=1}^{\infty} \left(a_n\cos nx+b_n\sin nx\right) \]while developing a Fourier series. The symbol reminds us that a series can represent the function in a convergence sense without necessarily agreeing with the originally assigned point value at every discontinuity.
Once the convergence behavior is known, equality may be stated on the points where the series and function agree.
The Gibbs Phenomenon
Near a jump discontinuity, Fourier partial sums develop oscillatory overshoots and undershoots. Increasing the number of terms makes the oscillations narrower, but the peak overshoot does not simply disappear pointwise at the jump.
This behavior is called the Gibbs phenomenon. It is an important reminder that adding more Fourier modes improves global approximation without making every local feature behave monotonically.
Coefficient Decay and Smoothness
The rate at which Fourier coefficients decrease often reflects how smooth the function is. Roughly speaking, smoother functions tend to have faster-decaying coefficients, while jumps and sharp features tend to produce slower decay.
This connection is important because coefficient decay tells us how many modes may be needed for a useful approximation and how efficiently a function can be represented spectrally.
Parseval's Identity
Orthogonality also creates an energy relationship. For a function represented by a Fourier series on \([-\pi,\pi]\), Parseval's identity takes the form
\[ \boxed{ \frac{1}{\pi}\int_{-\pi}^{\pi}|f(x)|^2\,dx =\frac{a_0^2}{2} +\sum_{n=1}^{\infty}(a_n^2+b_n^2) }. \]The total squared magnitude of the function is therefore distributed among its Fourier coefficients.
Interpreting Parseval's Identity
In vector language, Parseval's identity is analogous to the Pythagorean theorem: the squared length of an object equals the sum of the squared lengths of its orthogonal components.
In signal language, it says that energy measured in the original domain can be accounted for by energy in the frequency components.
This principle becomes even more central when we move from Fourier series to Fourier transforms.
Using Parseval to Evaluate a Series
For \(f(x)=x\) on \([-\pi,\pi]\), we found
\[ a_0=0, \qquad a_n=0, \qquad b_n=\frac{2(-1)^{n+1}}{n}. \]Parseval's identity gives
\[ \frac{1}{\pi}\int_{-\pi}^{\pi}x^2\,dx =\sum_{n=1}^{\infty}\frac{4}{n^2}. \]Since
\[ \frac{1}{\pi}\int_{-\pi}^{\pi}x^2\,dx =\frac{2\pi^2}{3}, \]we obtain
\[ \boxed{ \sum_{n=1}^{\infty}\frac{1}{n^2}=\frac{\pi^2}{6} }. \]Thus a function expansion can reveal a nontrivial numerical series identity.
Complex Exponential Form
Euler's formula,
\[ e^{ix}=\cos x+i\sin x, \]lets us combine sine and cosine components into complex exponentials. A \(2\pi\)-periodic Fourier series can be written
\[ \boxed{ f(x)\sim \sum_{n=-\infty}^{\infty}c_n e^{inx} }. \]This form is algebraically compact and is especially useful in differential equations, signal processing, and later transform methods.
The Complex Fourier Coefficients
The complex coefficients are
\[ \boxed{ c_n=\frac{1}{2\pi} \int_{-\pi}^{\pi}f(x)e^{-inx}\,dx }. \]For \(n>0\), the relationship with the real coefficients is
\[ c_n=\frac{a_n-i b_n}{2}, \qquad c_{-n}=\frac{a_n+i b_n}{2}, \]while
\[ c_0=\frac{a_0}{2}. \]Real Functions and Conjugate Symmetry
If \(f(x)\) is real-valued, its complex Fourier coefficients satisfy
\[ \boxed{c_{-n}=\overline{c_n}}. \]The negative- and positive-frequency coefficients therefore contain paired information. Their combination produces the real sine and cosine representation.
Fourier Series and Differential Equations
Trigonometric functions are natural building blocks for many differential equations because differentiating sines and cosines preserves the same family:
\[ \frac{d}{dx}\sin(nx)=n\cos(nx), \] \[ \frac{d}{dx}\cos(nx)=-n\sin(nx). \]In suitable settings, a differential equation can therefore be converted into equations for individual Fourier coefficients. This is one reason Fourier series are central to boundary-value problems and partial differential equations.
Fourier Series and Boundary Conditions
Sine and cosine families naturally satisfy different endpoint behaviors. For example,
\[ \sin\left(\frac{n\pi x}{L}\right)=0 \quad\text{at }x=0\text{ and }x=L. \]This makes sine expansions natural for many problems with zero-value boundary conditions. Cosine expansions often arise when derivative conditions are imposed at endpoints.
A later lesson on boundary-value problems will develop this connection more systematically.
Fourier Series and Signal Analysis
If \(f(t)\) represents a periodic signal, its Fourier coefficients identify the frequencies contributing to that signal. Large coefficients indicate strong components; small coefficients indicate weak components.
This viewpoint supports frequency analysis, filtering, compression, noise reduction, and system response. Later lessons will separate these applications from the underlying mathematics developed here.
Fourier Series and Spectral Thinking
The word spectrum refers to how an object is distributed among components associated with frequencies or other characteristic modes. For a Fourier series, the sequence
\[ \{a_n,b_n\}_{n=1}^{\infty} \]or equivalently \(\{c_n\}\), provides a discrete spectrum.
The original function describes behavior in the physical or time domain. The coefficients describe the same object through its frequency components.
Discrete Frequencies versus a Continuous Spectrum
A Fourier series is built for periodic behavior, so its allowed frequencies are separated into discrete harmonics:
\[ \omega_n=n\omega_0. \]For nonperiodic functions, a discrete list of harmonics is often too restrictive. The next major step is to allow a continuum of frequencies. That transition leads to the Fourier transform.
Computing Fourier Coefficients Numerically
When the defining integrals cannot be evaluated conveniently by hand, numerical integration can approximate the coefficients. For example,
\[ a_n=\frac{1}{\pi}\int_{-\pi}^{\pi}f(x)\cos(nx)\,dx \]can be estimated using quadrature methods.
If the function is available only at sampled points, discrete Fourier methods become more natural. Those methods will be developed later in this section.
Truncation as a Practical Model
Real computations use finitely many coefficients. A truncated Fourier model
\[ S_N(x)=\frac{a_0}{2} +\sum_{n=1}^{N}\left(a_n\cos nx+b_n\sin nx\right) \]retains frequencies up to the \(N\)-th harmonic.
The choice of \(N\) is a modeling decision. A small \(N\) produces a simpler, smoother approximation. A large \(N\) can capture finer structure but may be more sensitive to data quality and may require more computation.
A Frequency Cutoff
Truncating a Fourier series at \(N\) can be interpreted as retaining lower frequencies and discarding all higher ones. In that sense, ordinary truncation acts like an idealized low-pass operation on the expansion.
This observation provides an early connection to filtering. A later lesson will develop filtering and signal analysis directly rather than treating truncation only as an approximation device.
Common Error: Forgetting the Constant Factor \(1/2\)
The standard series is
\[ \frac{a_0}{2} +\sum_{n=1}^{\infty}(a_n\cos nx+b_n\sin nx), \]not
\[ a_0+\sum_{n=1}^{\infty}(a_n\cos nx+b_n\sin nx). \]The factor \(1/2\) appears because the constant basis function has squared norm \(2\pi\), while each nonconstant sine or cosine has squared norm \(\pi\).
Common Error: Using the Wrong Interval Formula
The formulas
\[ a_n=\frac{1}{\pi}\int_{-\pi}^{\pi}f(x)\cos(nx)\,dx \]and
\[ b_n=\frac{1}{\pi}\int_{-\pi}^{\pi}f(x)\sin(nx)\,dx \]are specific to a \(2\pi\)-periodic representation on \([-\pi,\pi]\). For period \(2L\), both the normalization factor and the trigonometric arguments change.
Common Error: Ignoring Symmetry
Computing every coefficient from scratch can create unnecessary algebra and more opportunities for mistakes.
Before integrating, test whether the function is even, odd, or has another symmetry. An even function has no sine coefficients; an odd function has no constant or cosine coefficients.
Common Error: Assuming Convergence to the Assigned Jump Value
At a jump, a Fourier series normally converges to
\[ \frac{f(x^-)+f(x^+)}{2}, \]under standard convergence conditions. This may differ from whatever value was assigned to \(f(x)\) at the jump itself.
Changing the value at one isolated point does not change the Fourier coefficients, because definite integrals are unaffected by a single point.
Common Error: Treating a Partial Sum as the Full Series
The quantity
\[ S_N(x)=\frac{a_0}{2} +\sum_{n=1}^{N}(a_n\cos nx+b_n\sin nx) \]is a finite approximation. It should not be confused with the infinite series.
A graph of \(S_N\) can be extremely informative, but it represents the behavior of the truncated model, not automatically the exact pointwise behavior of the infinite expansion.
Common Error: Losing the Harmonic Index
In the coefficient formulas, the same index \(n\) must be used consistently:
\[ a_n=\frac{1}{\pi}\int_{-\pi}^{\pi}f(x)\cos(nx)\,dx. \]Replacing \(\cos(nx)\) with \(\cos x\) computes only the first cosine harmonic, not the general coefficient.
A Reliable Procedure for Finding a Fourier Series
- Identify the period and choose one complete interval.
- Write the coefficient formulas appropriate to that interval.
- Check for even, odd, or other useful symmetry.
- Compute \(a_0\).
- Compute the general cosine coefficient \(a_n\).
- Compute the general sine coefficient \(b_n\).
- Simplify the coefficients, noting whether some indices vanish.
- Write the Fourier series and, when useful, its first several terms.
- State the expected behavior at discontinuities.
- Use a partial sum, symmetry check, or independent calculation to verify the result.
A Reliable Procedure for General Period \(2L\)
- Set the working interval to \([-L,L]\).
- Use frequencies \(n\pi/L\).
- Use the normalization factor \(1/L\).
- Exploit symmetry before integrating.
- Compute coefficients in symbolic form whenever possible.
- Check that the final series has the required period \(2L\).
Example: Full Fourier Workflow
Find the Fourier series for
\[ f(x)=x, \qquad -\pi<x<\pi, \]with periodic extension.
First, \(f\) is odd, so
\[ a_0=0, \qquad a_n=0. \]Then
\[ b_n=\frac{2}{\pi}\int_0^{\pi}x\sin(nx)\,dx. \]Integration by parts gives
\[ b_n=\frac{2(-1)^{n+1}}{n}. \]Therefore,
\[ \boxed{ f(x)\sim 2\sum_{n=1}^{\infty}\frac{(-1)^{n+1}}{n}\sin(nx) }. \]At the periodic jump points \(x=(2k+1)\pi\), the series converges to the midpoint of the left and right limits, which is \(0\).
Example: Fourier Series of a Constant Function
Let
\[ f(x)=C \]on \([-\pi,\pi]\). Then
\[ a_0=\frac{1}{\pi}\int_{-\pi}^{\pi}C\,dx=2C. \]For \(n\ge 1\), orthogonality gives
\[ a_n=0, \qquad b_n=0. \]Thus
\[ \boxed{f(x)=\frac{a_0}{2}=C}. \]A pure constant contains only the zero-frequency component.
Example: A Single Harmonic
Suppose
\[ f(x)=3\cos(4x)-2\sin(4x). \]This function is already in Fourier form. All coefficients vanish except
\[ a_4=3, \qquad b_4=-2. \]The amplitude of the fourth harmonic is
\[ \boxed{A_4=\sqrt{3^2+(-2)^2}=\sqrt{13}}. \]This example shows how the coefficient pair \((a_n,b_n)\) describes one frequency component.
Practice 1: Identify the Period
Find the fundamental period of
\[ f(x)=\cos(3x). \]Practice 2: Use Symmetry
For
\[ f(x)=x^3, \qquad -\pi<x<\pi, \]determine immediately which types of Fourier coefficients must be zero.
Practice 3: Constant Coefficient
Compute \(a_0\) for
\[ f(x)=2+x, \qquad -\pi<x<\pi. \]Practice 4: A Known Harmonic
For
\[ f(x)=5\cos(2x)+7\sin(3x), \]identify all nonzero Fourier coefficients in the standard \(2\pi\)-periodic form.
Practice 5: General Period
A function has period \(8\). Find its fundamental angular frequency \(\omega_0\).
Practice 6: Jump Convergence
Suppose a periodic function has
\[ f(x_0^-)=2, \qquad f(x_0^+)=8. \]Under the standard Fourier convergence conditions, what value does its Fourier series approach at \(x_0\)?
Practice 7: Partial Sum
Write the second partial sum \(S_2(x)\) when
\[ a_0=4, \quad a_1=1, \quad b_1=-2, \quad a_2=3, \quad b_2=0. \]Practice 8: Identify the Error
A student says that because \(f(x)=x\) is odd on \([-\pi,\pi]\), all of its Fourier coefficients must be zero. Explain the error.
Practice 9: Complex Coefficient Symmetry
If a real-valued periodic function has
\[ c_3=2-5i, \]find \(c_{-3}\).
Practice 10: Parseval Interpretation
In words, explain what Parseval's identity says about a function and its Fourier coefficients.
Practice Answers
Practice 1
The cosine function repeats when its argument increases by \(2\pi\). Thus
\[ 3T=2\pi, \]so
\[ \boxed{T=\frac{2\pi}{3}}. \]Practice 2
Since \(x^3\) is odd,
\[ \boxed{a_0=0\quad\text{and}\quad a_n=0}. \]The sine coefficients \(b_n\) may be nonzero.
Practice 3
\[ a_0=\frac{1}{\pi}\int_{-\pi}^{\pi}(2+x)\,dx. \]The odd term \(x\) integrates to zero, so
\[ a_0=\frac{1}{\pi}(4\pi)=\boxed{4}. \]Therefore the constant term in the Fourier series is \(a_0/2=2\).
Practice 4
The expression is already in Fourier form. The only nonzero coefficients are
\[ \boxed{a_2=5, \qquad b_3=7}. \]Also \(a_0=0\).
Practice 5
With \(T=8\),
\[ \omega_0=\frac{2\pi}{T} =\frac{2\pi}{8} =\boxed{\frac{\pi}{4}}. \]Practice 6
The Fourier series converges to the midpoint of the jump:
\[ \frac{2+8}{2}=\boxed{5}. \]Practice 7
Since \(a_0/2=2\),
\[ \boxed{ S_2(x) =2+\cos x-2\sin x+3\cos(2x) }. \]Practice 8
Odd symmetry forces the constant and cosine coefficients to vanish, but it does not force the sine coefficients to vanish. For \(f(x)=x\),
\[ \boxed{ b_n=\frac{2(-1)^{n+1}}{n} }, \]so the function has a nontrivial sine series.
Practice 9
A real-valued function satisfies \(c_{-n}=\overline{c_n}\). Therefore,
\[ \boxed{c_{-3}=2+5i}. \]Practice 10
Parseval's identity says that the total squared magnitude, or energy, of a function can be accounted for by the squared magnitudes of its orthogonal Fourier components. It is the function-space analogue of decomposing squared vector length among perpendicular coordinates.
What to Remember
A Fourier series expresses periodic behavior in orthogonal trigonometric modes:
\[ \boxed{ f(x)\sim \frac{a_0}{2} +\sum_{n=1}^{\infty}(a_n\cos nx+b_n\sin nx) }. \]- \(a_0/2\) is the average level of the function;
- \(a_n\) measures cosine content at the \(n\)-th harmonic;
- \(b_n\) measures sine content at the \(n\)-th harmonic;
- even symmetry eliminates sine coefficients;
- odd symmetry eliminates the constant and cosine coefficients;
- partial sums are least-squares projections onto finite trigonometric spaces;
- at a jump, the series converges to the midpoint of the one-sided limits under standard conditions;
- Parseval's identity relates total squared magnitude to coefficient magnitudes;
- the complex exponential form packages the same information more compactly.
The most important structural idea is that a periodic function can be understood through its harmonics. Fourier coefficients convert a description in the original variable into a discrete frequency description.
Where We Go Next
Fourier series describe periodic functions using a discrete set of harmonic frequencies. Many important signals and functions are not periodic, so they cannot be represented naturally by a single discrete harmonic ladder tied to one fundamental period.
The next lesson, Fourier Transforms, extends the same projection and frequency ideas from discrete harmonics to a continuous range of frequencies. That transition turns the Fourier series viewpoint into one of the central transform methods of analysis, differential equations, physics, engineering, and signal processing.