diff options
Diffstat (limited to 'valverect~.c')
| -rw-r--r-- | valverect~.c | 32 |
1 files changed, 22 insertions, 10 deletions
diff --git a/valverect~.c b/valverect~.c index 246878c..37ac536 100644 --- a/valverect~.c +++ b/valverect~.c @@ -7,6 +7,7 @@ #include "m_pd.h"
#include <math.h>
#include <string.h>
+#include <stdint.h>
#ifdef _MSC_VER
#pragma warning( disable : 4244 )
#pragma warning( disable : 4305 )
@@ -65,19 +66,30 @@ static void *valverect_tilde_new(t_floatarg sag, t_floatarg dist) }
/* Andrew Simper's pow(2, x) aproximation from the music-dsp list */
-static float f_pow2(float x)
+
+/* 32 bit "pointer cast" union */
+typedef union {
+ float f;
+ int32_t i;
+} ls_pcast32;
+
+/* union version */
+static inline float f_pow2(float x)
{
- long *px = (long*)(&x); // store address of float as long pointer
- const float tx = (x-0.5f) + (3<<22); // temporary value for truncation
- const long lx = *((long*)&tx) - 0x4b400000; // integer power of 2
- const float dx = x-(float)(lx); // float remainder of power of 2
+ ls_pcast32 *px, tx, lx;
+ float dx;
+
+ px = (ls_pcast32 *)&x; // store address of float as long pointer
+ tx.f = (x-0.5f) + (3<<22); // temporary value for truncation
+ lx.i = tx.i - 0x4b400000; // integer power of 2
+ dx = x - (float)lx.i; // float remainder of power of 2
- x = 1.0f + dx*(0.6960656421638072f + // cubic apporoximation of 2^x
- dx*(0.224494337302845f + // for x in the range [0, 1]
- dx*(0.07944023841053369f)));
- *px += (lx<<23); // add integer power of 2 to exponent
+ x = 1.0f + dx * (0.6960656421638072f + // cubic apporoximation of 2^x
+ dx * (0.224494337302845f + // for x in the range [0, 1]
+ dx * (0.07944023841053369f)));
+ (*px).i += (lx.i << 23); // add integer power of 2 to exponent
- return x;
+ return (*px).f;
}
static t_int *valverect_tilde_perform(t_int *w)
|