2 * Floating point AAN DCT
3 * this implementation is based upon the IJG integer AAN DCT (see jfdctfst.c)
5 * Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at>
6 * Copyright (c) 2003 Roman Shaposhnik
8 * Permission to use, copy, modify, and/or distribute this software for any
9 * purpose with or without fee is hereby granted, provided that the above
10 * copyright notice and this permission notice appear in all copies.
12 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
13 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
14 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
15 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
16 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
17 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
18 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
24 * Floating point AAN DCT
25 * @author Michael Niedermayer <michaelni@gmx.at>
29 #include "libavutil/internal.h"
30 #include "libavutil/libm.h"
34 //numbers generated by simple c code (not as accurate as they could be)
37 printf("#define B%d %1.20llf\n", i, (long double)1.0/(cosl(i*acosl(-1.0)/(long double)16.0)*sqrtl(2)));
40 #define B0 1.00000000000000000000
41 #define B1 0.72095982200694791383 // (cos(pi*1/16)sqrt(2))^-1
42 #define B2 0.76536686473017954350 // (cos(pi*2/16)sqrt(2))^-1
43 #define B3 0.85043009476725644878 // (cos(pi*3/16)sqrt(2))^-1
44 #define B4 1.00000000000000000000 // (cos(pi*4/16)sqrt(2))^-1
45 #define B5 1.27275858057283393842 // (cos(pi*5/16)sqrt(2))^-1
46 #define B6 1.84775906502257351242 // (cos(pi*6/16)sqrt(2))^-1
47 #define B7 3.62450978541155137218 // (cos(pi*7/16)sqrt(2))^-1
50 #define A1 0.70710678118654752438 // cos(pi*4/16)
51 #define A2 0.54119610014619698435 // cos(pi*6/16)sqrt(2)
52 #define A5 0.38268343236508977170 // cos(pi*6/16)
53 #define A4 1.30656296487637652774 // cos(pi*2/16)sqrt(2)
55 static const FLOAT postscale[64]={
56 B0*B0, B0*B1, B0*B2, B0*B3, B0*B4, B0*B5, B0*B6, B0*B7,
57 B1*B0, B1*B1, B1*B2, B1*B3, B1*B4, B1*B5, B1*B6, B1*B7,
58 B2*B0, B2*B1, B2*B2, B2*B3, B2*B4, B2*B5, B2*B6, B2*B7,
59 B3*B0, B3*B1, B3*B2, B3*B3, B3*B4, B3*B5, B3*B6, B3*B7,
60 B4*B0, B4*B1, B4*B2, B4*B3, B4*B4, B4*B5, B4*B6, B4*B7,
61 B5*B0, B5*B1, B5*B2, B5*B3, B5*B4, B5*B5, B5*B6, B5*B7,
62 B6*B0, B6*B1, B6*B2, B6*B3, B6*B4, B6*B5, B6*B6, B6*B7,
63 B7*B0, B7*B1, B7*B2, B7*B3, B7*B4, B7*B5, B7*B6, B7*B7,
66 static av_always_inline void row_fdct(FLOAT temp[64], int16_t *data)
68 FLOAT tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
69 FLOAT tmp10, tmp11, tmp12, tmp13;
70 FLOAT z2, z4, z11, z13;
74 for (i=0; i<8*8; i+=8) {
75 tmp0= data[0 + i] + data[7 + i];
76 tmp7= data[0 + i] - data[7 + i];
77 tmp1= data[1 + i] + data[6 + i];
78 tmp6= data[1 + i] - data[6 + i];
79 tmp2= data[2 + i] + data[5 + i];
80 tmp5= data[2 + i] - data[5 + i];
81 tmp3= data[3 + i] + data[4 + i];
82 tmp4= data[3 + i] - data[4 + i];
89 temp[0 + i]= tmp10 + tmp11;
90 temp[4 + i]= tmp10 - tmp11;
94 temp[2 + i]= tmp13 + tmp12;
95 temp[6 + i]= tmp13 - tmp12;
102 z5= (tmp4 - tmp6) * A5;
106 z2= tmp4*(A2+A5) - tmp6*A5;
107 z4= tmp6*(A4-A5) + tmp4*A5;
114 temp[5 + i]= z13 + z2;
115 temp[3 + i]= z13 - z2;
116 temp[1 + i]= z11 + z4;
117 temp[7 + i]= z11 - z4;
121 void ff_faandct(int16_t *data)
123 FLOAT tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
124 FLOAT tmp10, tmp11, tmp12, tmp13;
125 FLOAT z2, z4, z11, z13;
132 row_fdct(temp, data);
134 for (i=0; i<8; i++) {
135 tmp0= temp[8*0 + i] + temp[8*7 + i];
136 tmp7= temp[8*0 + i] - temp[8*7 + i];
137 tmp1= temp[8*1 + i] + temp[8*6 + i];
138 tmp6= temp[8*1 + i] - temp[8*6 + i];
139 tmp2= temp[8*2 + i] + temp[8*5 + i];
140 tmp5= temp[8*2 + i] - temp[8*5 + i];
141 tmp3= temp[8*3 + i] + temp[8*4 + i];
142 tmp4= temp[8*3 + i] - temp[8*4 + i];
149 data[8*0 + i]= lrintf(postscale[8*0 + i] * (tmp10 + tmp11));
150 data[8*4 + i]= lrintf(postscale[8*4 + i] * (tmp10 - tmp11));
154 data[8*2 + i]= lrintf(postscale[8*2 + i] * (tmp13 + tmp12));
155 data[8*6 + i]= lrintf(postscale[8*6 + i] * (tmp13 - tmp12));
162 z5= (tmp4 - tmp6) * A5;
166 z2= tmp4*(A2+A5) - tmp6*A5;
167 z4= tmp6*(A4-A5) + tmp4*A5;
174 data[8*5 + i]= lrintf(postscale[8*5 + i] * (z13 + z2));
175 data[8*3 + i]= lrintf(postscale[8*3 + i] * (z13 - z2));
176 data[8*1 + i]= lrintf(postscale[8*1 + i] * (z11 + z4));
177 data[8*7 + i]= lrintf(postscale[8*7 + i] * (z11 - z4));
181 void ff_faandct248(int16_t *data)
183 FLOAT tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
184 FLOAT tmp10, tmp11, tmp12, tmp13;
190 row_fdct(temp, data);
192 for (i=0; i<8; i++) {
193 tmp0 = temp[8*0 + i] + temp[8*1 + i];
194 tmp1 = temp[8*2 + i] + temp[8*3 + i];
195 tmp2 = temp[8*4 + i] + temp[8*5 + i];
196 tmp3 = temp[8*6 + i] + temp[8*7 + i];
197 tmp4 = temp[8*0 + i] - temp[8*1 + i];
198 tmp5 = temp[8*2 + i] - temp[8*3 + i];
199 tmp6 = temp[8*4 + i] - temp[8*5 + i];
200 tmp7 = temp[8*6 + i] - temp[8*7 + i];
207 data[8*0 + i] = lrintf(postscale[8*0 + i] * (tmp10 + tmp11));
208 data[8*4 + i] = lrintf(postscale[8*4 + i] * (tmp10 - tmp11));
212 data[8*2 + i] = lrintf(postscale[8*2 + i] * (tmp13 + tmp12));
213 data[8*6 + i] = lrintf(postscale[8*6 + i] * (tmp13 - tmp12));
220 data[8*1 + i] = lrintf(postscale[8*0 + i] * (tmp10 + tmp11));
221 data[8*5 + i] = lrintf(postscale[8*4 + i] * (tmp10 - tmp11));
225 data[8*3 + i] = lrintf(postscale[8*2 + i] * (tmp13 + tmp12));
226 data[8*7 + i] = lrintf(postscale[8*6 + i] * (tmp13 - tmp12));