Update to faad2 cvs 20040915+MPlayer fixes

Patch by me and Emanuele Giaquinta


git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@18142 b3059339-0415-0410-9bf9-f77b7e298cf2

89 files changed, 2143 insertions, 1427 deletions

diff --git a/libfaad2/Makefile b/libfaad2/Makefile
index 7e3be6a33a..ca911d054d 100644
--- a/libfaad2/Makefile
+++ b/libfaad2/Makefile
@@ -8,6 +8,7 @@ SRCS    = bits.c \
           common.c \
           decoder.c \
           drc.c \
+          drm_dec.c \
           error.c \
           filtbank.c \
           hcr.c \
diff --git a/libfaad2/README b/libfaad2/README
index bd36b51fc2..5fb717062e 100644
--- a/libfaad2/README
+++ b/libfaad2/README
@@ -1,3 +1,3 @@
-This is a FAAD2 2.1 beta CVS snapshot from 2004-07-12.
+This is a FAAD2 CVS snapshot from 2004-09-15.
 Only the content of the libfaad/ and include/ subdirectories is present here.
 MPlayer-specific modifications can be found in local_changes.diff.
diff --git a/libfaad2/analysis.h b/libfaad2/analysis.h
index 0a562eaaf1..25dd9d90ce 100644
--- a/libfaad2/analysis.h
+++ b/libfaad2/analysis.h
@@ -22,7 +22,7 @@
 ** Commercial non-GPL licensing of this software is possible.
 ** For more info contact Ahead Software through Mpeg4AAClicense@nero.com.
 **
-** $Id: analysis.h,v 1.12 2004/02/26 09:29:26 menno Exp $
+** $Id: analysis.h,v 1.14 2004/09/04 14:56:27 menno Exp $
 **/
 
 #ifndef __ANALYSIS_H__
diff --git a/libfaad2/bits.c b/libfaad2/bits.c
index 3c4cb80a24..0801e0e4ff 100644
--- a/libfaad2/bits.c
+++ b/libfaad2/bits.c
@@ -22,7 +22,7 @@
 ** Commercial non-GPL licensing of this software is possible.
 ** For more info contact Ahead Software through Mpeg4AAClicense@nero.com.
 **
-** $Id: bits.c,v 1.37 2004/02/26 09:29:26 menno Exp $
+** $Id: bits.c,v 1.39 2004/09/04 14:56:27 menno Exp $
 **/
 
 #include "common.h"
diff --git a/libfaad2/cfft.c b/libfaad2/cfft.c
index 6a1b005fc3..3b81665595 100644
--- a/libfaad2/cfft.c
+++ b/libfaad2/cfft.c
@@ -22,7 +22,7 @@
 ** Commercial non-GPL licensing of this software is possible.
 ** For more info contact Ahead Software through Mpeg4AAClicense@nero.com.
 **
-** $Id: cfft.c,v 1.27 2004/06/30 12:45:55 menno Exp $
+** $Id: cfft.c,v 1.30 2004/09/08 09:43:11 gcp Exp $
 **/
 
 /*
@@ -44,14 +44,6 @@
 
 
 /* static function declarations */
-#ifdef USE_SSE
-static void passf2pos_sse(const uint16_t l1, const complex_t *cc,
-                          complex_t *ch, const complex_t *wa);
-static void passf2pos_sse_ido(const uint16_t ido, const uint16_t l1, const complex_t *cc,
-                              complex_t *ch, const complex_t *wa);
-static void passf4pos_sse_ido(const uint16_t ido, const uint16_t l1, const complex_t *cc, complex_t *ch,
-                              const complex_t *wa1, const complex_t *wa2, const complex_t *wa3);
-#endif
 static void passf2pos(const uint16_t ido, const uint16_t l1, const complex_t *cc,
                       complex_t *ch, const complex_t *wa);
 static void passf2neg(const uint16_t ido, const uint16_t l1, const complex_t *cc,
@@ -74,89 +66,6 @@ static void cffti1(uint16_t n, complex_t *wa, uint16_t *ifac);
    passf2, passf3, passf4, passf5. Complex FFT passes fwd and bwd.
   ----------------------------------------------------------------------*/
 
-#if 0 //def USE_SSE
-static void passf2pos_sse(const uint16_t l1, const complex_t *cc,
-                          complex_t *ch, const complex_t *wa)
-{
-    uint16_t k, ah, ac;
-
-    for (k = 0; k < l1; k++)
-    {
-        ah = 2*k;
-        ac = 4*k;
-
-        RE(ch[ah])    = RE(cc[ac]) + RE(cc[ac+1]);
-        IM(ch[ah])    = IM(cc[ac]) + IM(cc[ac+1]);
-
-        RE(ch[ah+l1]) = RE(cc[ac]) - RE(cc[ac+1]);
-        IM(ch[ah+l1]) = IM(cc[ac]) - IM(cc[ac+1]);
-    }
-}
-
-static void passf2pos_sse_ido(const uint16_t ido, const uint16_t l1, const complex_t *cc,
-                              complex_t *ch, const complex_t *wa)
-{
-    uint16_t i, k, ah, ac;
-
-    for (k = 0; k < l1; k++)
-    {
-        ah = k*ido;
-        ac = 2*k*ido;
-
-        for (i = 0; i < ido; i+=4)
-        {
-            __m128 m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14;
-            __m128 m15, m16, m17, m18, m19, m20, m21, m22, m23, m24;
-            __m128 w1, w2, w3, w4;
-
-            m1 = _mm_load_ps(&RE(cc[ac+i]));
-            m2 = _mm_load_ps(&RE(cc[ac+ido+i]));
-            m5 = _mm_load_ps(&RE(cc[ac+i+2]));
-            m6 = _mm_load_ps(&RE(cc[ac+ido+i+2]));
-            w1 = _mm_load_ps(&RE(wa[i]));
-            w3 = _mm_load_ps(&RE(wa[i+2]));
-
-            m3 = _mm_add_ps(m1, m2);
-            m15 = _mm_add_ps(m5, m6);
-
-            m4 = _mm_sub_ps(m1, m2);
-            m16 = _mm_sub_ps(m5, m6);
-
-            _mm_store_ps(&RE(ch[ah+i]), m3);
-            _mm_store_ps(&RE(ch[ah+i+2]), m15);
-
-
-            w2 = _mm_shuffle_ps(w1, w1, _MM_SHUFFLE(2, 3, 0, 1));
-            w4 = _mm_shuffle_ps(w3, w3, _MM_SHUFFLE(2, 3, 0, 1));
-
-            m7 = _mm_mul_ps(m4, w1);
-            m17 = _mm_mul_ps(m16, w3);
-            m8 = _mm_mul_ps(m4, w2);
-            m18 = _mm_mul_ps(m16, w4);
-
-            m9  = _mm_shuffle_ps(m7, m8, _MM_SHUFFLE(2, 0, 2, 0));
-            m19 = _mm_shuffle_ps(m17, m18, _MM_SHUFFLE(2, 0, 2, 0));
-            m10 = _mm_shuffle_ps(m7, m8, _MM_SHUFFLE(3, 1, 3, 1));
-            m20 = _mm_shuffle_ps(m17, m18, _MM_SHUFFLE(3, 1, 3, 1));
-
-            m11 = _mm_add_ps(m9, m10);
-            m21 = _mm_add_ps(m19, m20);
-            m12 = _mm_sub_ps(m9, m10);
-            m22 = _mm_sub_ps(m19, m20);
-
-            m13 = _mm_shuffle_ps(m11, m11, _MM_SHUFFLE(0, 0, 3, 2));
-            m23 = _mm_shuffle_ps(m21, m21, _MM_SHUFFLE(0, 0, 3, 2));
-
-            m14 = _mm_unpacklo_ps(m12, m13);
-            m24 = _mm_unpacklo_ps(m22, m23);
-
-            _mm_store_ps(&RE(ch[ah+i+l1*ido]), m14);
-            _mm_store_ps(&RE(ch[ah+i+2+l1*ido]), m24);
-        }
-    }
-}
-#endif
-
 static void passf2pos(const uint16_t ido, const uint16_t l1, const complex_t *cc,
                       complex_t *ch, const complex_t *wa)
 {
@@ -385,218 +294,6 @@ static void passf3(const uint16_t ido, const uint16_t l1, const complex_t *cc,
     }
 }
 
-#ifdef USE_SSE
-ALIGN static const int32_t negate[4] = { 0x0, 0x0, 0x0, 0x80000000 };
-
-__declspec(naked) static void passf4pos_sse(const uint16_t l1, const complex_t *cc,
-                                     complex_t *ch, const complex_t *wa1, const complex_t *wa2,
-                                     const complex_t *wa3)
-{
-    __asm {
-        push      ebx
-        mov       ebx, esp
-        and       esp, -16
-        push      edi
-        push      esi
-        sub       esp, 8
-        movzx     edi, WORD PTR [ebx+8]
-
-        movaps    xmm1, XMMWORD PTR negate
-
-        test      edi, edi
-        jle       l1_is_zero
-
-        lea       esi, DWORD PTR [edi+edi]
-        add       esi, esi
-        sub       esi, edi
-        add       esi, esi
-        add       esi, esi
-        add       esi, esi
-        mov       eax, DWORD PTR [ebx+16]
-        add       esi, eax
-        lea       ecx, DWORD PTR [edi+edi]
-        add       ecx, ecx
-        add       ecx, ecx
-        add       ecx, ecx
-        add       ecx, eax
-        lea       edx, DWORD PTR [edi+edi]
-        add       edx, edx
-        add       edx, edx
-        add       edx, eax
-        xor       eax, eax
-        mov       DWORD PTR [esp], ebp
-        mov       ebp, DWORD PTR [ebx+12]
-
-fftloop:
-        lea       edi, DWORD PTR [eax+eax]
-        add       edi, edi
-        movaps    xmm2, XMMWORD PTR [ebp+edi*8]
-        movaps    xmm0, XMMWORD PTR [ebp+edi*8+16]
-        movaps    xmm7, XMMWORD PTR [ebp+edi*8+32]
-        movaps    xmm5, XMMWORD PTR [ebp+edi*8+48]
-        movaps    xmm6, xmm2
-        addps     xmm6, xmm0
-        movaps    xmm4, xmm1
-        xorps     xmm4, xmm7
-        movaps    xmm3, xmm1
-        xorps     xmm3, xmm5
-        xorps     xmm2, xmm1
-        xorps     xmm0, xmm1
-        addps     xmm7, xmm5
-        subps     xmm2, xmm0
-        movaps    xmm0, xmm6
-        shufps    xmm0, xmm7, 68
-        subps     xmm4, xmm3
-        shufps    xmm6, xmm7, 238
-        movaps    xmm5, xmm2
-        shufps    xmm5, xmm4, 68
-        movaps    xmm3, xmm0
-        addps     xmm3, xmm6
-        shufps    xmm2, xmm4, 187
-        subps     xmm0, xmm6
-        movaps    xmm4, xmm5
-        addps     xmm4, xmm2
-        mov       edi, DWORD PTR [ebx+16]
-        movaps    XMMWORD PTR [edi+eax*8], xmm3
-        subps     xmm5, xmm2
-        movaps    XMMWORD PTR [edx+eax*8], xmm4
-        movaps    XMMWORD PTR [ecx+eax*8], xmm0
-        movaps    XMMWORD PTR [esi+eax*8], xmm5
-        add       eax, 2
-        movzx     eax, ax
-        movzx     edi, WORD PTR [ebx+8]
-        cmp       eax, edi
-        jl        fftloop
-
-        mov       ebp, DWORD PTR [esp]
-
-l1_is_zero:
-
-        add       esp, 8
-        pop       esi
-        pop       edi
-        mov       esp, ebx
-        pop       ebx
-        ret
-    }
-}
-#endif
-
-#if 0
-static void passf4pos_sse_ido(const uint16_t ido, const uint16_t l1, const complex_t *cc,
-                              complex_t *ch, const complex_t *wa1, const complex_t *wa2,
-                              const complex_t *wa3)
-{
-    uint16_t i, k, ac, ah;
-
-    for (k = 0; k < l1; k++)
-    {
-        ac = 4*k*ido;
-        ah = k*ido;
-
-        for (i = 0; i < ido; i+=2)
-        {
-            __m128 m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14, m15, m16;
-            __m128 n1, n2, n3, n4, n5, n6, n7, n8, n9, m17, m18, m19, m20, m21, m22, m23;
-            __m128 w1, w2, w3, w4, w5, w6, m24, m25, m26, m27, m28, m29, m30;
-            __m128 neg1 = _mm_set_ps(-1.0, 1.0, -1.0, 1.0);
-
-            m1 = _mm_load_ps(&RE(cc[ac+i]));
-            m2 = _mm_load_ps(&RE(cc[ac+i+2*ido]));
-            m3 = _mm_add_ps(m1, m2);
-            m4 = _mm_sub_ps(m1, m2);
-
-            n1 = _mm_load_ps(&RE(cc[ac+i+ido]));
-            n2 = _mm_load_ps(&RE(cc[ac+i+3*ido]));
-            n3 = _mm_add_ps(n1, n2);
-
-            n4 = _mm_mul_ps(neg1, n1);
-            n5 = _mm_mul_ps(neg1, n2);
-            n6 = _mm_sub_ps(n4, n5);
-
-            m5 = _mm_add_ps(m3, n3);
-
-            n7 = _mm_shuffle_ps(n6, n6, _MM_SHUFFLE(2, 3, 0, 1));
-            n8 = _mm_add_ps(m4, n7);
-
-            m6 = _mm_sub_ps(m3, n3);
-            n9 = _mm_sub_ps(m4, n7);
-
-            _mm_store_ps(&RE(ch[ah+i]), m5);
-
-#if 0
-            static INLINE void ComplexMult(real_t *y1, real_t *y2,
-                real_t x1, real_t x2, real_t c1, real_t c2)
-            {
-                *y1 = MUL_F(x1, c1) + MUL_F(x2, c2);
-                *y2 = MUL_F(x2, c1) - MUL_F(x1, c2);
-            }
-
-            m7.0 = RE(c2)*RE(wa1[i])
-            m7.1 = IM(c2)*IM(wa1[i])
-            m7.2 = RE(c6)*RE(wa1[i+1])
-            m7.3 = IM(c6)*IM(wa1[i+1])
-
-            m8.0 = RE(c2)*IM(wa1[i])
-            m8.1 = IM(c2)*RE(wa1[i])
-            m8.2 = RE(c6)*IM(wa1[i+1])
-            m8.3 = IM(c6)*RE(wa1[i+1])
-
-            RE(0) = m7.0 - m7.1
-            IM(0) = m8.0 + m8.1
-            RE(1) = m7.2 - m7.3
-            IM(1) = m8.2 + m8.3
-
-            ////
-            RE(0) = RE(c2)*RE(wa1[i])   - IM(c2)*IM(wa1[i])
-            IM(0) = RE(c2)*IM(wa1[i])   + IM(c2)*RE(wa1[i])
-            RE(1) = RE(c6)*RE(wa1[i+1]) - IM(c6)*IM(wa1[i+1])
-            IM(1) = RE(c6)*IM(wa1[i+1]) + IM(c6)*RE(wa1[i+1])
-#endif
-
-            w1 = _mm_load_ps(&RE(wa1[i]));
-            w3 = _mm_load_ps(&RE(wa2[i]));
-            w5 = _mm_load_ps(&RE(wa3[i]));
-
-            w2 = _mm_shuffle_ps(w1, w1, _MM_SHUFFLE(2, 3, 0, 1));
-            w4 = _mm_shuffle_ps(w3, w3, _MM_SHUFFLE(2, 3, 0, 1));
-            w6 = _mm_shuffle_ps(w5, w5, _MM_SHUFFLE(2, 3, 0, 1));
-
-            m7 = _mm_mul_ps(n8, w1);
-            m15 = _mm_mul_ps(m6, w3);
-            m23 = _mm_mul_ps(n9, w5);
-            m8 = _mm_mul_ps(n8, w2);
-            m16 = _mm_mul_ps(m6, w4);
-            m24 = _mm_mul_ps(n9, w6);
-
-            m9  = _mm_shuffle_ps(m7, m8, _MM_SHUFFLE(2, 0, 2, 0));
-            m17 = _mm_shuffle_ps(m15, m16, _MM_SHUFFLE(2, 0, 2, 0));
-            m25 = _mm_shuffle_ps(m23, m24, _MM_SHUFFLE(2, 0, 2, 0));
-            m10 = _mm_shuffle_ps(m7, m8, _MM_SHUFFLE(3, 1, 3, 1));
-            m18 = _mm_shuffle_ps(m15, m16, _MM_SHUFFLE(3, 1, 3, 1));
-            m26 = _mm_shuffle_ps(m23, m24, _MM_SHUFFLE(3, 1, 3, 1));
-
-            m11 = _mm_add_ps(m9, m10);
-            m19 = _mm_add_ps(m17, m18);
-            m27 = _mm_add_ps(m25, m26);
-            m12 = _mm_sub_ps(m9, m10);
-            m20 = _mm_sub_ps(m17, m18);
-            m28 = _mm_sub_ps(m25, m26);
-
-            m13 = _mm_shuffle_ps(m11, m11, _MM_SHUFFLE(0, 0, 3, 2));
-            m21 = _mm_shuffle_ps(m19, m19, _MM_SHUFFLE(0, 0, 3, 2));
-            m29 = _mm_shuffle_ps(m27, m27, _MM_SHUFFLE(0, 0, 3, 2));
-            m14 = _mm_unpacklo_ps(m12, m13);
-            m22 = _mm_unpacklo_ps(m20, m21);
-            m30 = _mm_unpacklo_ps(m28, m29);
-
-            _mm_store_ps(&RE(ch[ah+i+l1*ido]), m14);
-            _mm_store_ps(&RE(ch[ah+i+2*l1*ido]), m22);
-            _mm_store_ps(&RE(ch[ah+i+3*l1*ido]), m30);
-        }
-    }
-}
-#endif
 
 static void passf4pos(const uint16_t ido, const uint16_t l1, const complex_t *cc,
                       complex_t *ch, const complex_t *wa1, const complex_t *wa2,
@@ -992,101 +689,6 @@ static void passf5(const uint16_t ido, const uint16_t l1, const complex_t *cc,
    cfftf1, cfftf, cfftb, cffti1, cffti. Complex FFTs.
   ----------------------------------------------------------------------*/
 
-#ifdef USE_SSE