Fourier Series Flashcards
fourier series is very useful for
describing phenomena with any kind of repeating, periodic behaviour.
Almost any kind of
periodic signal can be written as
sums of sines and cosines, or alternatively as exponentials.
for understanding fourier series, it is helpful to think of them as
a decomposition in
terms of orthogonal sets of vectors.
in the case of fourier series, the vectors are
functions
in the case of fourier series, the orthogonality is defined in terms of
a generalisation of the inner product of two vectors
inner product of vectors
<a|b>=a^cross b = sum for i=1 to N of ai^*bi
where cross denotes the hermitian conjugate and we have made us of Dirac notation
inner product of functions f and g
<f|g>= integral dx f(x)^*g(x)
for a scalar, the hermitian conjugate and the complex conjugate are
identical
proof for the functions sin(2πrx/L) and cos(2πrx/L) being the orthogonal basis vectors in the space of periodic functions with period L.
recognising odd and even functions
integrating an odd function gives zero
(one uses a trig identity)
fourier coefficients
a0, ar and br
depend on what specific function f(x) is being considered
how to obtain the cosone fourier coefficients
- using basis vector cos(2pipx/L) take the inner product with fourier series expansion
- consider each term and rearrange
the sine and cosine terms in a fourier series span
a vector space of functions
the fourier series for any function f(x) in the space, converges to
f(x) wherever f(x) is continuous
the fourier series of any function f(x) satisfies
the dirichlet conditions
the dirichlet conditions
- f(x) must be periodic
- f(x) must be single-valued and either continuous or with a finite number of finite discontinuities
- f(x) must be of bounded variation
- the integral over |f(x)| must converge
what does the dirichlet condition ‘must be of bounded variation’ mean
essentially the distance travelled by walking along the curve for one period is finite
example of a function with discontinuities
square wave
the fourier series of a function with a discontinuity at x=xd converges to
thehalf-way value
(between left hand limit and right hand limit)
Gibb’s phenomenon
overshooting near each discontinuity remaining the same size despite more and more terms being added
what other functions can be sued as an orthogonal basis
e^irkx and e^-ipkx where k=2pi/L
complex fourier series
f(x)= sum to infinity of c_r e^irkx
for real functions,c_-r=
cr*
how to prove the relationship between real trigonometric and complex fourier series
splitting the definition of the complex fourier series expansion into real and imaginary parts
fourier series of a non-periodic function
need to continue the functions outside the given range to make it periodic (eg use sawtooth function or triangle wave)
Parseval’s theorem
for a function f(x) with complex fourier series coefficients cr and period L
can calculate the inner product of a function with itself
frequency or power spectrum
distribution of the numbers |cr|^2