so3_test_csv.c File Reference

#include <stdio.h>
#include <stdlib.h>
#include <complex.h>
#include <time.h>
#include <so3.h>

Include dependency graph for so3_test_csv.c:

Macros
#define	NREPEAT   10
#define	MIN(a, b)   ((a < b) ? (a) : (b))
#define	MAX(a, b)   ((a > b) ? (a) : (b))

Functions
double	get_max_error (complex double *expected, complex double *actual, int n)
double	ran2_dp (int idum)
void	so3_test_gen_flmn_complex (complex double *flmn, const so3_parameters_t *parameters, int seed)
void	so3_test_gen_flmn_real (complex double *flmn, const so3_parameters_t *parameters, int seed)
int	main (int argc, char **argv)

Detailed Description

Applies SO3 algorithms to perform inverse and forward Wigner transforms (respectively) to check that the original signal is reconstructed exactly (to numerical precision). Test is performed on a random signal with harmonic coefficients uniformly sampled from (-1,1), once for a real and once for a complex signal. This test is run for multiple values of L and outputs the results in CSV format (to stdout). If no value for N is given in the arguments, each test will run with N = L, otherwise all tests will be run with the given N (and tests with L < N will be skipped). L will take all powers of 2 less or equal than the given Lmax. If L0 is given, all powers of 2 less than or equal to L0 will be skipped (and L0 will be used as the lower band-limit for all transforms).

Usage

so3_test_csv [Lmax [L0 [N]]]

e.g.

so3_test 128 0 4

Defaults: L = 64, L0 = 0, N = L
To save it to a file, just use something like

so3_test_csv 64 7 4 > test.csv

Author

Martin Büttner

Jason McEwen

Macro Definition Documentation

MAX

#define MAX	(		a,
			b
	)		((a > b) ? (a) : (b))

MIN

#define MIN	(		a,
			b
	)		((a < b) ? (a) : (b))

NREPEAT

#define NREPEAT   10

Function Documentation

get_max_error()

double get_max_error	(	complex double *	expected,
		complex double *	actual,
		int	n
	)

main()

int main	(	int	argc,
		char **	argv
	)

ran2_dp()

double ran2_dp

(

int

idum

)

Generate uniform deviate in range [0,1) given seed. (Using double precision.)

Note

Uniform deviate (Num rec 1992, chap 7.1), original routine said to be 'perfect'.

Parameters

[in]

idum

Seed.

Return values

ran_dp

Generated uniform deviate.

Author

Martin Büttner

Jason McEwen

so3_test_gen_flmn_complex()

void so3_test_gen_flmn_complex	(	complex double *	flmn,
		const so3_parameters_t *	parameters,
		int	seed
	)

Generate random Wigner coefficients of a complex signal.

Parameters

[out]	flmn	Random spherical harmonic coefficients generated. Provide enough memory for fully padded storage, i.e. (2*N-1)*L*L elements. Unused trailing elements will be set to zero.
[in]	parameters	A parameters object with (at least) the following fields: L0, L, N, storage, n_mode The reality flag is ignored. Use so3_test_gen_flmn_real instead for real signals.
[in]	seed	Integer seed required for random number generator.

Return values

none

Author

Martin Büttner

Jason McEwen

so3_test_gen_flmn_real()

void so3_test_gen_flmn_real	(	complex double *	flmn,
		const so3_parameters_t *	parameters,
		int	seed
	)

Generate random Wigner coefficients of a real signal. We only generate coefficients for n >= 0, and for n = 0, we need flm0* = (-1)^(m)*fl-m0, so that fl00 has to be real.

Parameters

[out]	flmn	Random spherical harmonic coefficients generated. Provide enough memory for fully padded, complex (!) storage, i.e. (2*N-1)*L*L elements. Unused trailing elements will be set to zero.
[in]	parameters	A parameters object with (at least) the following fields: L0, L, N, storage, n_mode The reality flag is ignored. Use so3_test_gen_flmn_complex instead for complex signals.
[in]	seed	Integer seed required for random number generator.

Return values

none

Author

Martin Büttner

Jason McEwen