1 @cindex discrete Hankel transforms
2 @cindex Hankel transforms, discrete
3 @cindex transforms, Hankel
4 This chapter describes functions for performing Discrete Hankel
5 Transforms (DHTs). The functions are declared in the header file
9 * Discrete Hankel Transform Definition::
10 * Discrete Hankel Transform Functions::
11 * Discrete Hankel Transform References::
14 @node Discrete Hankel Transform Definition
17 The discrete Hankel transform acts on a vector of sampled data, where
18 the samples are assumed to have been taken at points related to the
19 zeroes of a Bessel function of fixed order; compare this to the case of
20 the discrete Fourier transform, where samples are taken at points
21 related to the zeroes of the sine or cosine function.
23 Specifically, let @math{f(t)} be a function on the unit interval.
24 Then the finite @math{\nu}-Hankel transform of @math{f(t)} is defined
25 to be the set of numbers @math{g_m} given by,
29 g_m = \int_0^1 t dt\, J_\nu(j_{\nu,m}t) f(t),
35 g_m = \int_0^1 t dt J_\nu(j_(\nu,m)t) f(t),
44 f(t) = \sum_{m=1}^\infty {{2 J_\nu(j_{\nu,m}x)}\over{J_{\nu+1}(j_{\nu,m})^2}} g_m.
50 f(t) = \sum_@{m=1@}^\infty (2 J_\nu(j_(\nu,m)x) / J_(\nu+1)(j_(\nu,m))^2) g_m.
55 Suppose that @math{f} is band-limited in the sense that
56 @math{g_m=0} for @math{m > M}. Then we have the following
57 fundamental sampling theorem.
61 g_m = {{2}\over{j_{\nu,M}^2}}
62 \sum_{k=1}^{M-1} f\left({{j_{\nu,k}}\over{j_{\nu,M}}}\right)
63 {{J_\nu(j_{\nu,m} j_{\nu,k} / j_{\nu,M})}\over{J_{\nu+1}(j_{\nu,k})^2}}.
69 g_m = (2 / j_(\nu,M)^2)
70 \sum_@{k=1@}^@{M-1@} f(j_(\nu,k)/j_(\nu,M))
71 (J_\nu(j_(\nu,m) j_(\nu,k) / j_(\nu,M)) / J_(\nu+1)(j_(\nu,k))^2).
76 It is this discrete expression which defines the discrete Hankel
77 transform. The kernel in the summation above defines the matrix of the
78 @math{\nu}-Hankel transform of size @math{M-1}. The coefficients of
79 this matrix, being dependent on @math{\nu} and @math{M}, must be
80 precomputed and stored; the @code{gsl_dht} object encapsulates this
81 data. The allocation function @code{gsl_dht_alloc} returns a
82 @code{gsl_dht} object which must be properly initialized with
83 @code{gsl_dht_init} before it can be used to perform transforms on data
84 sample vectors, for fixed @math{\nu} and @math{M}, using the
85 @code{gsl_dht_apply} function. The implementation allows a scaling of
86 the fundamental interval, for convenience, so that one can assume the
87 function is defined on the interval @math{[0,X]}, rather than the unit
90 Notice that by assumption @math{f(t)} vanishes at the endpoints
91 of the interval, consistent with the inversion formula
92 and the sampling formula given above. Therefore, this transform
93 corresponds to an orthogonal expansion in eigenfunctions
94 of the Dirichlet problem for the Bessel differential equation.
97 @node Discrete Hankel Transform Functions
100 @deftypefun {gsl_dht *} gsl_dht_alloc (size_t @var{size})
101 This function allocates a Discrete Hankel transform object of size
105 @deftypefun int gsl_dht_init (gsl_dht * @var{t}, double @var{nu}, double @var{xmax})
106 This function initializes the transform @var{t} for the given values of
107 @var{nu} and @var{x}.
110 @deftypefun {gsl_dht *} gsl_dht_new (size_t @var{size}, double @var{nu}, double @var{xmax})
111 This function allocates a Discrete Hankel transform object of size
112 @var{size} and initializes it for the given values of @var{nu} and
116 @deftypefun void gsl_dht_free (gsl_dht * @var{t})
117 This function frees the transform @var{t}.
120 @deftypefun int gsl_dht_apply (const gsl_dht * @var{t}, double * @var{f_in}, double * @var{f_out})
121 This function applies the transform @var{t} to the array @var{f_in}
122 whose size is equal to the size of the transform. The result is stored
123 in the array @var{f_out} which must be of the same length.
126 @deftypefun double gsl_dht_x_sample (const gsl_dht * @var{t}, int @var{n})
127 This function returns the value of the @var{n}-th sample point in the unit interval,
128 @c{${({j_{\nu,n+1}} / {j_{\nu,M}}}) X$}
129 @math{(j_@{\nu,n+1@}/j_@{\nu,M@}) X}. These are the
130 points where the function @math{f(t)} is assumed to be sampled.
133 @deftypefun double gsl_dht_k_sample (const gsl_dht * @var{t}, int @var{n})
134 This function returns the value of the @var{n}-th sample point in ``k-space'',
135 @c{${{j_{\nu,n+1}} / X}$}
136 @math{j_@{\nu,n+1@}/X}.
139 @node Discrete Hankel Transform References
140 @section References and Further Reading
142 The algorithms used by these functions are described in the following papers,
146 H. Fisk Johnson, Comp.@: Phys.@: Comm.@: 43, 181 (1987).
151 D. Lemoine, J. Chem.@: Phys.@: 101, 3936 (1994).