* adding integer/floating point AAN implementations for DCT 2-4-8

Originally committed as revision 2430 to svn://svn.ffmpeg.org/ffmpeg/trunk

libavcodec/dsputil.c

@@ -2884,11 +2884,11 @@ void dsputil_init(DSPContext* c, AVCodecContext *avctx)
 #ifdef CONFIG_ENCODERS
     if(avctx->dct_algo==FF_DCT_FASTINT) {
         c->fdct = fdct_ifast;
-	c->fdct248 = ff_fdct248_islow; // FIXME: need an optimized version
+	c->fdct248 = fdct_ifast248;
     } 
     else if(avctx->dct_algo==FF_DCT_FAAN) {
         c->fdct = ff_faandct;
-	c->fdct248 = ff_fdct248_islow; // FIXME: need an optimized version
+	c->fdct248 = ff_faandct248; 
     } 
     else {
         c->fdct = ff_jpeg_fdct_islow; //slow/accurate/default

libavcodec/dsputil.h

@@ -36,6 +36,7 @@
 typedef short DCTELEM;
 
 void fdct_ifast (DCTELEM *data);
+void fdct_ifast248 (DCTELEM *data);
 void ff_jpeg_fdct_islow (DCTELEM *data);
 void ff_fdct248_islow (DCTELEM *data);
 

libavcodec/faandct.c

@@ -160,3 +160,89 @@ void ff_faandct(DCTELEM * data)
         data[8*7 + i]= lrintf(SCALE(8*7 + i) * (z11 - z4));
     }
 }
+
+void ff_faandct248(DCTELEM * data)
+{
+    FLOAT tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
+    FLOAT tmp10, tmp11, tmp12, tmp13;
+    FLOAT z1, z2, z3, z4, z5, z11, z13;
+    FLOAT temp[64];
+    int i;
+
+    emms_c();
+
+    for (i=0; i<8*8; i+=8) {
+        tmp0= data[0 + i] + data[7 + i];
+        tmp7= data[0 + i] - data[7 + i];
+        tmp1= data[1 + i] + data[6 + i];
+        tmp6= data[1 + i] - data[6 + i];
+        tmp2= data[2 + i] + data[5 + i];
+        tmp5= data[2 + i] - data[5 + i];
+        tmp3= data[3 + i] + data[4 + i];
+        tmp4= data[3 + i] - data[4 + i];
+        
+        tmp10= tmp0 + tmp3;
+        tmp13= tmp0 - tmp3;
+        tmp11= tmp1 + tmp2;
+        tmp12= tmp1 - tmp2;
+        
+        temp[0 + i]= tmp10 + tmp11;
+        temp[4 + i]= tmp10 - tmp11;
+        
+        z1= (tmp12 + tmp13)*A1;
+        temp[2 + i]= tmp13 + z1;
+        temp[6 + i]= tmp13 - z1;
+        
+        tmp10= tmp4 + tmp5;
+        tmp11= tmp5 + tmp6;
+        tmp12= tmp6 + tmp7;
+
+        z5= (tmp10 - tmp12) * A5;
+        z2= tmp10*A2 + z5;
+        z4= tmp12*A4 + z5;
+        z3= tmp11*A1;
+
+        z11= tmp7 + z3;
+        z13= tmp7 - z3;
+
+        temp[5 + i]= z13 + z2;
+        temp[3 + i]= z13 - z2;
+        temp[1 + i]= z11 + z4;
+        temp[7 + i]= z11 - z4;
+    }
+
+    for (i=0; i<8; i++) {
+        tmp0 = temp[8*0 + i] + temp[8*1 + i];
+        tmp1 = temp[8*2 + i] + temp[8*3 + i];
+        tmp2 = temp[8*4 + i] + temp[8*5 + i];
+        tmp3 = temp[8*6 + i] + temp[8*7 + i];
+        tmp4 = temp[8*0 + i] - temp[8*1 + i];
+        tmp5 = temp[8*2 + i] - temp[8*3 + i];
+        tmp6 = temp[8*4 + i] - temp[8*5 + i];
+        tmp7 = temp[8*6 + i] - temp[8*7 + i];
+        
+        tmp10 = tmp0 + tmp3;
+        tmp11 = tmp1 + tmp2;
+        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));
+        
+        z1 = (tmp12 + tmp13)* A1;
+        data[8*2 + i] = lrintf(SCALE(8*2 + i) * (tmp13 + z1));
+        data[8*6 + i] = lrintf(SCALE(8*6 + i) * (tmp13 - z1));
+        
+        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));
+
+	z1 = (tmp12 + tmp13)* A1;
+	data[8*3 + i] = lrintf(SCALE(8*2 + i) * (tmp13 + z1));
+	data[8*7 + i] = lrintf(SCALE(8*6 + i) * (tmp13 - z1));
+    }
+}

libavcodec/faandct.h

@@ -28,3 +28,4 @@
 #define FAAN_POSTSCALE
  
 void ff_faandct(DCTELEM * data);
+void ff_faandct248(DCTELEM * data);

libavcodec/jfdctfst.c

@@ -228,6 +228,113 @@ fdct_ifast (DCTELEM * data)
   }
 }
 
+/*
+ * Perform the forward 2-4-8 DCT on one block of samples.
+ */
+
+GLOBAL(void)
+fdct_ifast248 (DCTELEM * data)
+{
+  DCTELEM tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
+  DCTELEM tmp10, tmp11, tmp12, tmp13;
+  DCTELEM z1, z2, z3, z4, z5, z11, z13;
+  DCTELEM *dataptr;
+  int ctr;
+  SHIFT_TEMPS
+
+  /* Pass 1: process rows. */
+
+  dataptr = data;
+  for (ctr = DCTSIZE-1; ctr >= 0; ctr--) {
+    tmp0 = dataptr[0] + dataptr[7];
+    tmp7 = dataptr[0] - dataptr[7];
+    tmp1 = dataptr[1] + dataptr[6];
+    tmp6 = dataptr[1] - dataptr[6];
+    tmp2 = dataptr[2] + dataptr[5];
+    tmp5 = dataptr[2] - dataptr[5];
+    tmp3 = dataptr[3] + dataptr[4];
+    tmp4 = dataptr[3] - dataptr[4];
+    
+    /* Even part */
+    
+    tmp10 = tmp0 + tmp3;	/* phase 2 */
+    tmp13 = tmp0 - tmp3;
+    tmp11 = tmp1 + tmp2;
+    tmp12 = tmp1 - tmp2;
+    
+    dataptr[0] = tmp10 + tmp11; /* phase 3 */
+    dataptr[4] = tmp10 - tmp11;
+    
+    z1 = MULTIPLY(tmp12 + tmp13, FIX_0_707106781); /* c4 */
+    dataptr[2] = tmp13 + z1;	/* phase 5 */
+    dataptr[6] = tmp13 - z1;
+    
+    /* Odd part */
+
+    tmp10 = tmp4 + tmp5;	/* phase 2 */
+    tmp11 = tmp5 + tmp6;
+    tmp12 = tmp6 + tmp7;
+
+    /* The rotator is modified from fig 4-8 to avoid extra negations. */
+    z5 = MULTIPLY(tmp10 - tmp12, FIX_0_382683433); /* c6 */
+    z2 = MULTIPLY(tmp10, FIX_0_541196100) + z5; /* c2-c6 */
+    z4 = MULTIPLY(tmp12, FIX_1_306562965) + z5; /* c2+c6 */
+    z3 = MULTIPLY(tmp11, FIX_0_707106781); /* c4 */
+
+    z11 = tmp7 + z3;		/* phase 5 */
+    z13 = tmp7 - z3;
+
+    dataptr[5] = z13 + z2;	/* phase 6 */
+    dataptr[3] = z13 - z2;
+    dataptr[1] = z11 + z4;
+    dataptr[7] = z11 - z4;
+
+    dataptr += DCTSIZE;		/* advance pointer to next row */
+  }
+
+  /* Pass 2: process columns. */
+
+  dataptr = data;
+  for (ctr = DCTSIZE-1; ctr >= 0; ctr--) {
+    tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*1];
+    tmp1 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*3];
+    tmp2 = dataptr[DCTSIZE*4] + dataptr[DCTSIZE*5];
+    tmp3 = dataptr[DCTSIZE*6] + dataptr[DCTSIZE*7];
+    tmp4 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*1];
+    tmp5 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*3];
+    tmp6 = dataptr[DCTSIZE*4] - dataptr[DCTSIZE*5];
+    tmp7 = dataptr[DCTSIZE*6] - dataptr[DCTSIZE*7];
+
+    /* Even part */
+    
+    tmp10 = tmp0 + tmp3;
+    tmp11 = tmp1 + tmp2;
+    tmp12 = tmp1 - tmp2;
+    tmp13 = tmp0 - tmp3;
+    
+    dataptr[DCTSIZE*0] = tmp10 + tmp11;
+    dataptr[DCTSIZE*4] = tmp10 - tmp11;
+    
+    z1 = MULTIPLY(tmp12 + tmp13, FIX_0_707106781);
+    dataptr[DCTSIZE*2] = tmp13 + z1;
+    dataptr[DCTSIZE*6] = tmp13 - z1;
+
+    tmp10 = tmp4 + tmp7;
+    tmp11 = tmp5 + tmp6;
+    tmp12 = tmp5 - tmp6;
+    tmp13 = tmp4 - tmp7;
+    
+    dataptr[DCTSIZE*1] = tmp10 + tmp11;
+    dataptr[DCTSIZE*5] = tmp10 - tmp11;
+    
+    z1 = MULTIPLY(tmp12 + tmp13, FIX_0_707106781);
+    dataptr[DCTSIZE*3] = tmp13 + z1;
+    dataptr[DCTSIZE*7] = tmp13 - z1;
+    
+    dataptr++;			/* advance pointer to next column */
+  }
+}
+
 
 #undef GLOBAL
 #undef CONST_BITS