Difference between revisions of "Sinc function"
Jump to navigation
Jump to search
Line 5: | Line 5: | ||
<pre> |
<pre> |
||
epsilon :: RealFloat a => a |
epsilon :: RealFloat a => a |
||
− | epsilon = |
+ | epsilon = encodeFloat 1 (fromIntegral $ 1-floatDigits epsilon) |
{- Boosted from Boost http://www.boost.org/boost/math/special_functions/sinc.hpp -} |
{- Boosted from Boost http://www.boost.org/boost/math/special_functions/sinc.hpp -} |
Revision as of 17:56, 25 February 2006
Sinc function
The sinc function is a useful function that is a little tricky to use because it becomes 0/0 as x tends to 0. Here is an implementation taken from the Boost library.
epsilon :: RealFloat a => a epsilon = encodeFloat 1 (fromIntegral $ 1-floatDigits epsilon) {- Boosted from Boost http://www.boost.org/boost/math/special_functions/sinc.hpp -} sinc :: (RealFloat a) => a -> a sinc x | (abs x) >= taylor_n_bound = (sin x)/x | otherwise = 1 - (x^2/6) + (x^4/120) where taylor_n_bound = sqrt $ sqrt epsilon