FFmpeg
 All Data Structures Files Functions Variables Typedefs Enumerations Enumerator Macros Groups Pages
faandct.c
Go to the documentation of this file.
1 /*
2  * Floating point AAN DCT
3  * this implementation is based upon the IJG integer AAN DCT (see jfdctfst.c)
4  *
5  * Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at>
6  * Copyright (c) 2003 Roman Shaposhnik
7  *
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.
11  *
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.
19  */
20 
21 /**
22  * @file
23  * @brief
24  * Floating point AAN DCT
25  * @author Michael Niedermayer <michaelni@gmx.at>
26  */
27 
28 #include "dsputil.h"
29 #include "faandct.h"
30 #include "libavutil/internal.h"
31 #include "libavutil/libm.h"
32 
33 #define FLOAT float
34 
35 //numbers generated by simple c code (not as accurate as they could be)
36 /*
37 for(i=0; i<8; i++){
38  printf("#define B%d %1.20llf\n", i, (long double)1.0/(cosl(i*acosl(-1.0)/(long double)16.0)*sqrtl(2)));
39 }
40 */
41 #define B0 1.00000000000000000000
42 #define B1 0.72095982200694791383 // (cos(pi*1/16)sqrt(2))^-1
43 #define B2 0.76536686473017954350 // (cos(pi*2/16)sqrt(2))^-1
44 #define B3 0.85043009476725644878 // (cos(pi*3/16)sqrt(2))^-1
45 #define B4 1.00000000000000000000 // (cos(pi*4/16)sqrt(2))^-1
46 #define B5 1.27275858057283393842 // (cos(pi*5/16)sqrt(2))^-1
47 #define B6 1.84775906502257351242 // (cos(pi*6/16)sqrt(2))^-1
48 #define B7 3.62450978541155137218 // (cos(pi*7/16)sqrt(2))^-1
49 
50 
51 #define A1 0.70710678118654752438 // cos(pi*4/16)
52 #define A2 0.54119610014619698435 // cos(pi*6/16)sqrt(2)
53 #define A5 0.38268343236508977170 // cos(pi*6/16)
54 #define A4 1.30656296487637652774 // cos(pi*2/16)sqrt(2)
55 
56 static const FLOAT postscale[64]={
57 B0*B0, B0*B1, B0*B2, B0*B3, B0*B4, B0*B5, B0*B6, B0*B7,
58 B1*B0, B1*B1, B1*B2, B1*B3, B1*B4, B1*B5, B1*B6, B1*B7,
59 B2*B0, B2*B1, B2*B2, B2*B3, B2*B4, B2*B5, B2*B6, B2*B7,
60 B3*B0, B3*B1, B3*B2, B3*B3, B3*B4, B3*B5, B3*B6, B3*B7,
61 B4*B0, B4*B1, B4*B2, B4*B3, B4*B4, B4*B5, B4*B6, B4*B7,
62 B5*B0, B5*B1, B5*B2, B5*B3, B5*B4, B5*B5, B5*B6, B5*B7,
63 B6*B0, B6*B1, B6*B2, B6*B3, B6*B4, B6*B5, B6*B6, B6*B7,
64 B7*B0, B7*B1, B7*B2, B7*B3, B7*B4, B7*B5, B7*B6, B7*B7,
65 };
66 
68 {
69  FLOAT tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
70  FLOAT tmp10, tmp11, tmp12, tmp13;
71  FLOAT z2, z4, z11, z13;
72  FLOAT av_unused z5;
73  int i;
74 
75  for (i=0; i<8*8; i+=8) {
76  tmp0= data[0 + i] + data[7 + i];
77  tmp7= data[0 + i] - data[7 + i];
78  tmp1= data[1 + i] + data[6 + i];
79  tmp6= data[1 + i] - data[6 + i];
80  tmp2= data[2 + i] + data[5 + i];
81  tmp5= data[2 + i] - data[5 + i];
82  tmp3= data[3 + i] + data[4 + i];
83  tmp4= data[3 + i] - data[4 + i];
84 
85  tmp10= tmp0 + tmp3;
86  tmp13= tmp0 - tmp3;
87  tmp11= tmp1 + tmp2;
88  tmp12= tmp1 - tmp2;
89 
90  temp[0 + i]= tmp10 + tmp11;
91  temp[4 + i]= tmp10 - tmp11;
92 
93  tmp12 += tmp13;
94  tmp12 *= A1;
95  temp[2 + i]= tmp13 + tmp12;
96  temp[6 + i]= tmp13 - tmp12;
97 
98  tmp4 += tmp5;
99  tmp5 += tmp6;
100  tmp6 += tmp7;
101 
102 #if 0
103  z5= (tmp4 - tmp6) * A5;
104  z2= tmp4*A2 + z5;
105  z4= tmp6*A4 + z5;
106 #else
107  z2= tmp4*(A2+A5) - tmp6*A5;
108  z4= tmp6*(A4-A5) + tmp4*A5;
109 #endif
110  tmp5*=A1;
111 
112  z11= tmp7 + tmp5;
113  z13= tmp7 - tmp5;
114 
115  temp[5 + i]= z13 + z2;
116  temp[3 + i]= z13 - z2;
117  temp[1 + i]= z11 + z4;
118  temp[7 + i]= z11 - z4;
119  }
120 }
121 
123 {
124  FLOAT tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
125  FLOAT tmp10, tmp11, tmp12, tmp13;
126  FLOAT z2, z4, z11, z13;
127  FLOAT av_unused z5;
128  FLOAT temp[64];
129  int i;
130 
131  emms_c();
132 
133  row_fdct(temp, data);
134 
135  for (i=0; i<8; i++) {
136  tmp0= temp[8*0 + i] + temp[8*7 + i];
137  tmp7= temp[8*0 + i] - temp[8*7 + i];
138  tmp1= temp[8*1 + i] + temp[8*6 + i];
139  tmp6= temp[8*1 + i] - temp[8*6 + i];
140  tmp2= temp[8*2 + i] + temp[8*5 + i];
141  tmp5= temp[8*2 + i] - temp[8*5 + i];
142  tmp3= temp[8*3 + i] + temp[8*4 + i];
143  tmp4= temp[8*3 + i] - temp[8*4 + i];
144 
145  tmp10= tmp0 + tmp3;
146  tmp13= tmp0 - tmp3;
147  tmp11= tmp1 + tmp2;
148  tmp12= tmp1 - tmp2;
149 
150  data[8*0 + i]= lrintf(postscale[8*0 + i] * (tmp10 + tmp11));
151  data[8*4 + i]= lrintf(postscale[8*4 + i] * (tmp10 - tmp11));
152 
153  tmp12 += tmp13;
154  tmp12 *= A1;
155  data[8*2 + i]= lrintf(postscale[8*2 + i] * (tmp13 + tmp12));
156  data[8*6 + i]= lrintf(postscale[8*6 + i] * (tmp13 - tmp12));
157 
158  tmp4 += tmp5;
159  tmp5 += tmp6;
160  tmp6 += tmp7;
161 
162 #if 0
163  z5= (tmp4 - tmp6) * A5;
164  z2= tmp4*A2 + z5;
165  z4= tmp6*A4 + z5;
166 #else
167  z2= tmp4*(A2+A5) - tmp6*A5;
168  z4= tmp6*(A4-A5) + tmp4*A5;
169 #endif
170  tmp5*=A1;
171 
172  z11= tmp7 + tmp5;
173  z13= tmp7 - tmp5;
174 
175  data[8*5 + i]= lrintf(postscale[8*5 + i] * (z13 + z2));
176  data[8*3 + i]= lrintf(postscale[8*3 + i] * (z13 - z2));
177  data[8*1 + i]= lrintf(postscale[8*1 + i] * (z11 + z4));
178  data[8*7 + i]= lrintf(postscale[8*7 + i] * (z11 - z4));
179  }
180 }
181 
183 {
184  FLOAT tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
185  FLOAT tmp10, tmp11, tmp12, tmp13;
186  FLOAT temp[64];
187  int i;
188 
189  emms_c();
190 
191  row_fdct(temp, data);
192 
193  for (i=0; i<8; i++) {
194  tmp0 = temp[8*0 + i] + temp[8*1 + i];
195  tmp1 = temp[8*2 + i] + temp[8*3 + i];
196  tmp2 = temp[8*4 + i] + temp[8*5 + i];
197  tmp3 = temp[8*6 + i] + temp[8*7 + i];
198  tmp4 = temp[8*0 + i] - temp[8*1 + i];
199  tmp5 = temp[8*2 + i] - temp[8*3 + i];
200  tmp6 = temp[8*4 + i] - temp[8*5 + i];
201  tmp7 = temp[8*6 + i] - temp[8*7 + i];
202 
203  tmp10 = tmp0 + tmp3;
204  tmp11 = tmp1 + tmp2;
205  tmp12 = tmp1 - tmp2;
206  tmp13 = tmp0 - tmp3;
207 
208  data[8*0 + i] = lrintf(postscale[8*0 + i] * (tmp10 + tmp11));
209  data[8*4 + i] = lrintf(postscale[8*4 + i] * (tmp10 - tmp11));
210 
211  tmp12 += tmp13;
212  tmp12 *= A1;
213  data[8*2 + i] = lrintf(postscale[8*2 + i] * (tmp13 + tmp12));
214  data[8*6 + i] = lrintf(postscale[8*6 + i] * (tmp13 - tmp12));
215 
216  tmp10 = tmp4 + tmp7;
217  tmp11 = tmp5 + tmp6;
218  tmp12 = tmp5 - tmp6;
219  tmp13 = tmp4 - tmp7;
220 
221  data[8*1 + i] = lrintf(postscale[8*0 + i] * (tmp10 + tmp11));
222  data[8*5 + i] = lrintf(postscale[8*4 + i] * (tmp10 - tmp11));
223 
224  tmp12 += tmp13;
225  tmp12 *= A1;
226  data[8*3 + i] = lrintf(postscale[8*2 + i] * (tmp13 + tmp12));
227  data[8*7 + i] = lrintf(postscale[8*6 + i] * (tmp13 - tmp12));
228  }
229 }