Merge commit '4a7af92cc80ced8498626401ed21f25ffe6740c8'

* commit '4a7af92cc80ced8498626401ed21f25ffe6740c8':
  sbrdsp: Unroll and use integer operations
  sbrdsp: Unroll sbr_autocorrelate_c
  x86: sbrdsp: Implement SSE2 qmf_deint_bfly

Conflicts:
	libavcodec/sbrdsp.c
	libavcodec/x86/sbrdsp.asm
	libavcodec/x86/sbrdsp_init.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>

libavcodec/sbrdsp.c

@@ -54,8 +54,7 @@ static void sbr_neg_odd_64_c(float *x)
 {
     union av_intfloat32 *xi = (union av_intfloat32*) x;
     int i;
-    for (i = 1; i < 64; i += 4)
+    for (i = 1; i < 64; i += 4) {
-    {
         xi[i + 0].i ^= 1U << 31;
         xi[i + 2].i ^= 1U << 31;
     }
@@ -73,6 +72,7 @@ static void sbr_qmf_pre_shuffle_c(float *z)
         zi[64 + 2 * k + 2].i = zi[63 - k].i ^ (1U << 31);
         zi[64 + 2 * k + 3].i = zi[ k + 2].i;
     }
+
     zi[64 + 2 * 31 + 0].i = zi[64 - 31].i ^ (1U << 31);
     zi[64 + 2 * 31 + 1].i = zi[31 +  1].i;
 }
@@ -139,16 +139,16 @@ static av_always_inline void autocorrelate(const float x[40][2],
 static void sbr_autocorrelate_c(const float x[40][2], float phi[3][2][2])
 {
 #if 0
-    // This code is slower because it multiplies memory accesses.
+    /* This code is slower because it multiplies memory accesses.
-    // It is left as eucational purpose and because it may offer
+     * It is left for educational purposes and because it may offer
-    // a better reference for writing arch-specific dsp functions.
+     * a better reference for writing arch-specific DSP functions. */
     autocorrelate(x, phi, 0);
     autocorrelate(x, phi, 1);
     autocorrelate(x, phi, 2);
 #else
     float real_sum2 = x[0][0] * x[2][0] + x[0][1] * x[2][1];
     float imag_sum2 = x[0][0] * x[2][1] - x[0][1] * x[2][0];
-    float real_sum1 = 0.f, imag_sum1 = 0.f, real_sum0 = 0.0f;
+    float real_sum1 = 0.0f, imag_sum1 = 0.0f, real_sum0 = 0.0f;
     int   i;
     for (i = 1; i < 38; i++) {
         real_sum0 += x[i][0] * x[i    ][0] + x[i][1] * x[i    ][1];