diff --git a/libavcodec/aacsbr.c b/libavcodec/aacsbr.c
index 4a3377d812..3cc927a779 100644
--- a/libavcodec/aacsbr.c
+++ b/libavcodec/aacsbr.c
@@ -1184,21 +1184,22 @@ static void sbr_qmf_synthesis(DSPContext *dsp, FFTContext *mdct,
             *v_off -= 128 >> div;
         }
         v = v0 + *v_off;
+        if (div) {
+            for (n = 0; n < 32; n++) {
+                X[0][i][   n] = -X[0][i][n];
+                X[0][i][32+n] =  X[1][i][31-n];
+            }
+            ff_imdct_half(mdct, mdct_buf[0], X[0][i]);
+            for (n = 0; n < 32; n++) {
+                v[     n] =  mdct_buf[0][63 - 2*n];
+                v[63 - n] = -mdct_buf[0][62 - 2*n];
+            }
+        } else {
         for (n = 1; n < 64 >> div; n+=2) {
             X[1][i][n] = -X[1][i][n];
         }
-        if (div) {
-            memset(X[0][i]+32, 0, 32*sizeof(float));
-            memset(X[1][i]+32, 0, 32*sizeof(float));
-        }
         ff_imdct_half(mdct, mdct_buf[0], X[0][i]);
         ff_imdct_half(mdct, mdct_buf[1], X[1][i]);
-        if (div) {
-            for (n = 0; n < 32; n++) {
-                v[      n] = -mdct_buf[0][63 - 2*n] + mdct_buf[1][2*n    ];
-                v[ 63 - n] =  mdct_buf[0][62 - 2*n] + mdct_buf[1][2*n + 1];
-            }
-        } else {
             for (n = 0; n < 64; n++) {
                 v[      n] = -mdct_buf[0][63 -   n] + mdct_buf[1][  n    ];
                 v[127 - n] =  mdct_buf[0][63 -   n] + mdct_buf[1][  n    ];