HOLBERG1D - Compute coefficients of Holberg's 1st derivative filter

holberg1d	comput the coefficients of Holberg's 1st derivative filter

Function Prototype:
void holbergd1 (float e, int n, float d[]);

Input:
e		maximum relative error in group velocity
n		number of coefficients in filter (must be 2, 4, 6, or 8)

Output:
d		array[n] of coefficients

Notes:
Coefficients are output in a form suitable for convolution.  The
derivative is centered halfway between coefficients d[n/2-1] and d[n/2].

Coefficients are computed via the power series method of Kindelan et al.,
1990, On the construction and efficiency of staggered numerical
differentiators for the wave equation:  Geophysics 55, 107-110.
See also, Holberg, 1987, Computational aspects of the choice of
operator and sampling interval for numerical differentiation in
large-scale simulation of wave phenomena:  Geophys. Prosp., 35, 629-655

Reference:
Kindelan et al., 1990, 
On the construction and efficiency of staggered numerical
differentiators for the wave equation:  Geophysics 55, 107-110.

See also, Holberg, 1987, Computational aspects of the choice of
operator and sampling interval for numerical differentiation in
large-scale simulation of wave phenomena:  Geophys. Prosp., 35, 629-655

Author:  Dave Hale, Colorado School of Mines, 06/06/91