Make ff_acelp_lp_synthesis_filter() receives a pointer to the actual filter coefficients and not the pointer minus one

Originally committed as revision 14031 to svn://svn.ffmpeg.org/ffmpeg/trunk

libavcodec/acelp_filters.c

@@ -121,6 +121,10 @@ int ff_acelp_lp_synthesis_filter(
 {
     int i,n;
 
+    // These two lines are two avoid a -1 subtraction in the main loop
+    filter_length++;
+    filter_coeffs--;
+
     for(n=0; n<buffer_length; n++)
     {
         int sum = rounder;

libavcodec/acelp_filters.h

@@ -125,7 +125,7 @@ void ff_acelp_convolve_circ(
  * \param filter_coeffs filter coefficients (-0x8000 <= (3.12) < 0x8000)
  * \param in input signal
  * \param buffer_length amount of data to process
- * \param filter_length filter length (11 for 10th order LP filter)
+ * \param filter_length filter length (10 for 10th order LP filter)
  * \param stop_on_overflow   1 - return immediately if overflow occurs
  *                           0 - ignore overflows
  * \param rounder the amount to add for rounding (usually 0x800 or 0xfff)

libavcodec/ra144.c

@@ -209,9 +209,9 @@ static void do_output_subblock(RA144Context *ractx, const uint16_t  *lpc_coefs,
            BLOCKSIZE*sizeof(*ractx->curr_sblock));
 
     if (ff_acelp_lp_synthesis_filter(
-                                     ractx->curr_sblock + 10, lpc_coefs -1,
+                                     ractx->curr_sblock + 10, lpc_coefs,
                                      ractx->curr_sblock + 10, BLOCKSIZE,
-                                     11, 1, 0xfff)
+                                     10, 1, 0xfff)
         )
         memset(ractx->curr_sblock, 0, 50*sizeof(*ractx->curr_sblock));
 }