X-Git-Url: https://git.sesse.net/?a=blobdiff_plain;f=libavcodec%2Ffaandct.c;h=7888633d7cc5139a37ba575bd20f6158a72f9703;hb=851960f6f8cf1f946fe42fa36cf6598fac68072c;hp=b16d46b48b1a62a38c855df0369db2d32657b303;hpb=bad5537e2c2caeb5deb1ff9d771ea01058b8010c;p=ffmpeg

diff --git a/libavcodec/faandct.c b/libavcodec/faandct.c
index b16d46b48b1..7888633d7cc 100644
--- a/libavcodec/faandct.c
+++ b/libavcodec/faandct.c
@@ -19,21 +19,17 @@
  */
 
 /**
- * @file libavcodec/faandct.c
+ * @file
  * @brief
  *     Floating point AAN DCT
  * @author Michael Niedermayer <michaelni@gmx.at>
  */
 
-#include "dsputil.h"
 #include "faandct.h"
+#include "libavutil/internal.h"
+#include "libavutil/libm.h"
 
 #define FLOAT float
-#ifdef FAAN_POSTSCALE
-#    define SCALE(x) postscale[x]
-#else
-#    define SCALE(x) 1
-#endif
 
 //numbers generated by simple c code (not as accurate as they could be)
 /*
@@ -67,12 +63,11 @@ B6*B0, B6*B1, B6*B2, B6*B3, B6*B4, B6*B5, B6*B6, B6*B7,
 B7*B0, B7*B1, B7*B2, B7*B3, B7*B4, B7*B5, B7*B6, B7*B7,
 };
 
-static av_always_inline void row_fdct(FLOAT temp[64], DCTELEM * data)
+static av_always_inline void row_fdct(FLOAT temp[64], int16_t *data)
 {
     FLOAT tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
     FLOAT tmp10, tmp11, tmp12, tmp13;
     FLOAT z2, z4, z11, z13;
-    FLOAT av_unused z5;
     int i;
 
     for (i=0; i<8*8; i+=8) {
@@ -102,14 +97,9 @@ static av_always_inline void row_fdct(FLOAT temp[64], DCTELEM * data)
         tmp5 += tmp6;
         tmp6 += tmp7;
 
-#if 0
-        z5= (tmp4 - tmp6) * A5;
-        z2= tmp4*A2 + z5;
-        z4= tmp6*A4 + z5;
-#else
         z2= tmp4*(A2+A5) - tmp6*A5;
         z4= tmp6*(A4-A5) + tmp4*A5;
-#endif
+
         tmp5*=A1;
 
         z11= tmp7 + tmp5;
@@ -122,12 +112,11 @@ static av_always_inline void row_fdct(FLOAT temp[64], DCTELEM * data)
     }
 }
 
-void ff_faandct(DCTELEM * data)
+void ff_faandct(int16_t *data)
 {
     FLOAT tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
     FLOAT tmp10, tmp11, tmp12, tmp13;
     FLOAT z2, z4, z11, z13;
-    FLOAT av_unused z5;
     FLOAT temp[64];
     int i;
 
@@ -150,39 +139,34 @@ void ff_faandct(DCTELEM * data)
         tmp11= tmp1 + tmp2;
         tmp12= tmp1 - tmp2;
 
-        data[8*0 + i]= lrintf(SCALE(8*0 + i) * (tmp10 + tmp11));
-        data[8*4 + i]= lrintf(SCALE(8*4 + i) * (tmp10 - tmp11));
+        data[8*0 + i]= lrintf(postscale[8*0 + i] * (tmp10 + tmp11));
+        data[8*4 + i]= lrintf(postscale[8*4 + i] * (tmp10 - tmp11));
 
         tmp12 += tmp13;
         tmp12 *= A1;
-        data[8*2 + i]= lrintf(SCALE(8*2 + i) * (tmp13 + tmp12));
-        data[8*6 + i]= lrintf(SCALE(8*6 + i) * (tmp13 - tmp12));
+        data[8*2 + i]= lrintf(postscale[8*2 + i] * (tmp13 + tmp12));
+        data[8*6 + i]= lrintf(postscale[8*6 + i] * (tmp13 - tmp12));
 
         tmp4 += tmp5;
         tmp5 += tmp6;
         tmp6 += tmp7;
 
-#if 0
-        z5= (tmp4 - tmp6) * A5;
-        z2= tmp4*A2 + z5;
-        z4= tmp6*A4 + z5;
-#else
         z2= tmp4*(A2+A5) - tmp6*A5;
         z4= tmp6*(A4-A5) + tmp4*A5;
-#endif
+
         tmp5*=A1;
 
         z11= tmp7 + tmp5;
         z13= tmp7 - tmp5;
 
-        data[8*5 + i]= lrintf(SCALE(8*5 + i) * (z13 + z2));
-        data[8*3 + i]= lrintf(SCALE(8*3 + i) * (z13 - z2));
-        data[8*1 + i]= lrintf(SCALE(8*1 + i) * (z11 + z4));
-        data[8*7 + i]= lrintf(SCALE(8*7 + i) * (z11 - z4));
+        data[8*5 + i]= lrintf(postscale[8*5 + i] * (z13 + z2));
+        data[8*3 + i]= lrintf(postscale[8*3 + i] * (z13 - z2));
+        data[8*1 + i]= lrintf(postscale[8*1 + i] * (z11 + z4));
+        data[8*7 + i]= lrintf(postscale[8*7 + i] * (z11 - z4));
     }
 }
 
-void ff_faandct248(DCTELEM * data)
+void ff_faandct248(int16_t *data)
 {
     FLOAT tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
     FLOAT tmp10, tmp11, tmp12, tmp13;
@@ -208,25 +192,25 @@ void ff_faandct248(DCTELEM * data)
         tmp12 = tmp1 - tmp2;
         tmp13 = tmp0 - tmp3;
 
-        data[8*0 + i] = lrintf(SCALE(8*0 + i) * (tmp10 + tmp11));
-        data[8*4 + i] = lrintf(SCALE(8*4 + i) * (tmp10 - tmp11));
+        data[8*0 + i] = lrintf(postscale[8*0 + i] * (tmp10 + tmp11));
+        data[8*4 + i] = lrintf(postscale[8*4 + i] * (tmp10 - tmp11));
 
         tmp12 += tmp13;
         tmp12 *= A1;
-        data[8*2 + i] = lrintf(SCALE(8*2 + i) * (tmp13 + tmp12));
-        data[8*6 + i] = lrintf(SCALE(8*6 + i) * (tmp13 - tmp12));
+        data[8*2 + i] = lrintf(postscale[8*2 + i] * (tmp13 + tmp12));
+        data[8*6 + i] = lrintf(postscale[8*6 + i] * (tmp13 - tmp12));
 
         tmp10 = tmp4 + tmp7;
         tmp11 = tmp5 + tmp6;
         tmp12 = tmp5 - tmp6;
         tmp13 = tmp4 - tmp7;
 
-        data[8*1 + i] = lrintf(SCALE(8*0 + i) * (tmp10 + tmp11));
-        data[8*5 + i] = lrintf(SCALE(8*4 + i) * (tmp10 - tmp11));
+        data[8*1 + i] = lrintf(postscale[8*0 + i] * (tmp10 + tmp11));
+        data[8*5 + i] = lrintf(postscale[8*4 + i] * (tmp10 - tmp11));
 
         tmp12 += tmp13;
         tmp12 *= A1;
-        data[8*3 + i] = lrintf(SCALE(8*2 + i) * (tmp13 + tmp12));
-        data[8*7 + i] = lrintf(SCALE(8*6 + i) * (tmp13 - tmp12));
+        data[8*3 + i] = lrintf(postscale[8*2 + i] * (tmp13 + tmp12));
+        data[8*7 + i] = lrintf(postscale[8*6 + i] * (tmp13 - tmp12));
     }
 }