aacdec: use float_dsp in apply_independent_coupling

Could not be used in apply_dependent_coupling because of alignment issues.

libavcodec/aacdec.c

@@ -247,14 +247,12 @@ static void apply_independent_coupling(AACContext *ac,
                                        SingleChannelElement *target,
                                        ChannelElement *cce, int index)
 {
-    int i;
     const float gain = cce->coup.gain[index][0];
     const float *src = cce->ch[0].ret;
     float *dest = target->ret;
     const int len = 1024 << (ac->oc[1].m4ac.sbr == 1);
 
-    for (i = 0; i < len; i++)
-        dest[i] += gain * src[i];
+    ac->fdsp->vector_fmac_scalar(dest, src, gain, len);
 }
 
 #include "aacdec_template.c"