g723.1: simplify and fix multiplication overflow

In 16-bit arithmetic, x * 0xffffc is simply x * -4 with extra overflows,
(and the constant was probably meant to be 0xfffc).  Combined with the
shift, this simplifies to -x >> 1.  Finally, clearing the low two bits
with a 32-bit mask and switching to a 32-bit type allows more efficient
code on 32-bit machines.

Signed-off-by: Mans Rullgard <mans@mansr.com>

libavcodec/g723_1.c

@@ -956,7 +956,7 @@ static void formant_postfilter(G723_1_Context *p, int16_t *lpc, int16_t *buf)
     signal_ptr = filter_signal + LPC_ORDER;
     for (i = 0; i < SUBFRAMES; i++) {
         int16_t temp_vector[SUBFRAME_LEN];
-        int16_t temp;
+        int temp;
         int auto_corr[2];
         int scale, energy;
 
@@ -975,7 +975,7 @@ static void formant_postfilter(G723_1_Context *p, int16_t *lpc, int16_t *buf)
             temp = (auto_corr[0] >> 2) / temp;
         }
         p->reflection_coef = (3 * p->reflection_coef + temp + 2) >> 2;
-        temp = (p->reflection_coef * 0xffffc >> 3) & 0xfffc;
+        temp = -p->reflection_coef >> 1 & ~3;
 
         /* Compensation filter */
         for (j = 0; j < SUBFRAME_LEN; j++) {