1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
|
/*
* YV12.c
*
* Copyright (C) Charles 'Buck' Krasic - April 2000
* Copyright (C) Erik Walthinsen - April 2000
*
* This file is part of libdv, a free DV (IEC 61834/SMPTE 314M)
* codec.
*
* libdv is free software; you can redistribute it and/or modify it
* under the terms of the GNU Lesser Public License as published by
* the Free Software Foundation; either version 2.1, or (at your
* option) any later version.
*
* libdv is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser Public License for more details.
*
* You should have received a copy of the GNU Lesser Public License
* along with libdv; see the file COPYING. If not, write to
* the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA.
*
* The libdv homepage is http://libdv.sourceforge.net/.
*/
#if HAVE_CONFIG_H
# include <config.h>
#endif
#include <stdlib.h>
#include "YV12.h"
#if ARCH_X86 || ARCH_X86_64
#include "mmx.h"
#endif // ARCH_X86 || ARCH_X86_64
/* Lookup tables for mapping signed to unsigned, and clamping */
static unsigned char real_uvlut[256], *uvlut;
static unsigned char real_ylut[768], *ylut;
#if ARCH_X86 || ARCH_X86_64
/* Define some constants used in MMX range mapping and clamping logic */
static mmx_t mmx_0x10s = (mmx_t) 0x1010101010101010LL,
mmx_0x0080s = (mmx_t) 0x0080008000800080LL,
mmx_0x7f24s = (mmx_t) 0x7f247f247f247f24LL,
mmx_0x7f94s = (mmx_t) 0x7f947f947f947f94LL,
mmx_0x7f7fs = (mmx_t) 0x7f7f7f7f7f7f7f7fLL,
mmx_0x7f00s = (mmx_t) 0x7f007f007f007f00LL;
#endif // ARCH_X86 || ARCH_X86_64
void
dv_YV12_init(int clamp_luma, int clamp_chroma) {
int i;
int value;
uvlut = real_uvlut + 128; // index from -128 .. 127
for(i=-128;
i<128;
++i) {
value = i + 128;
if (clamp_chroma == TRUE) value = CLAMP(value, 16, 240);
uvlut[i] = value;
} /* for */
ylut = real_ylut + 256; // index from -256 .. 511
for(i=-256;
i<512;
++i) {
value = i + 128;
if (clamp_luma == TRUE) value = CLAMP(value, 16, 235);
ylut[i] = value;
} /* for */
} /* dv_YV12_init */
void
dv_mb420_YV12(dv_macroblock_t *mb, uint8_t **pixels, uint16_t *pitches) {
dv_coeff_t *Y[4], *UV[2], *Ytmp, *UVtmp;
unsigned char *py, *pwy, *puv, *pwuv;
int i, j, row, col;
Y [0] = mb->b[0].coeffs;
Y [1] = mb->b[1].coeffs;
Y [2] = mb->b[2].coeffs;
Y [3] = mb->b[3].coeffs;
UV[0] = mb->b[4].coeffs;
UV[1] = mb->b[5].coeffs;
py = pixels[0] + mb->x + (mb->y * pitches[0]);
for(i=0; i<4; i+=2) { // two rows of Y blocks
for(row=0;
row < 8;
row++) {
pwy = py;
for(j=0; j<2; j++) { // two columns of Y blocks
Ytmp = Y[i+j];
for(col = 0; col < 8; col++) {
*pwy++ = ylut[CLAMP(*Ytmp, -256, 511)];
Ytmp++;
}
Y[i+j] = Ytmp;
} /* for j */
py += pitches[0];
} /* for row */
} /* for i */
for(i=1;
i<3;
i++) { // two Chroma blocks
puv = pixels[i] + (mb->x/2) + ((mb->y/2) * pitches[i]);
UVtmp = UV[i-1];
for(row=0;
row < 8;
row++, puv += pitches[i]) {
pwuv = puv;
for(col=0;
col<8;
col++) {
*pwuv++ = uvlut[CLAMP(*UVtmp, -128, 127)];
UVtmp++;
} /* for col */
} /* for row */
} // for i
} /* dv_mb420_YV12 */
#if ARCH_X86 || ARCH_X86_64
/* TODO: clamping */
void
dv_mb420_YV12_mmx(dv_macroblock_t *mb, uint8_t **pixels, uint16_t *pitches,
int clamp_luma, int clamp_chroma) {
dv_coeff_t *Y[4], *UV[2], *Ytmp, *UVtmp;
unsigned char *py, *pwy, *puv;
int i, j, row;
Y [0] = mb->b[0].coeffs;
Y [1] = mb->b[1].coeffs;
Y [2] = mb->b[2].coeffs;
Y [3] = mb->b[3].coeffs;
UV[0] = mb->b[4].coeffs;
UV[1] = mb->b[5].coeffs;
py = pixels[0] + mb->x + (mb->y * pitches[0]);
movq_m2r(mmx_0x0080s,mm0);
movq_m2r(mmx_0x10s,mm1);
if (clamp_luma == TRUE) {
movq_m2r(mmx_0x7f94s,mm2); /* hi clamp */
movq_m2r(mmx_0x7f24s,mm3); /* lo clamp */
} else {
movq_m2r(mmx_0x7f7fs,mm2); /* no clamp */
movq_m2r(mmx_0x7f00s,mm3); /* no clamp */
}
for(i=0; i<4; i+=2) { // two rows of Y blocks
for(row=0;
row < 8;
row++) {
pwy = py;
for(j=0; j<2; j++) { // two columns of Y blocks
Ytmp = Y[i+j];
movq_m2r(*(Ytmp ),mm4);
movq_m2r(*(Ytmp+4),mm5);
paddw_r2r(mm0,mm4); // +128
paddw_r2r(mm0,mm5); // +128
packuswb_r2r(mm5,mm4); // 16 -> 8 bit
movq_r2m(mm4, *pwy);
Y[i+j] = Ytmp + 8;
pwy += 8;
} /* for j */
py += pitches[0];
} /* for row */
} /* for i */
for(i=1;
i<3;
i++) { // two Chroma blocks
puv = pixels[i] + (mb->x/2) + ((mb->y/2) * pitches[i]);
UVtmp = UV[i-1];
for(row=0;
row < 8;
row++) {
movq_m2r(*(UVtmp ),mm4);
movq_m2r(*(UVtmp+4),mm5);
paddw_r2r(mm0,mm4); // +128
paddw_r2r(mm0,mm5); // +128
packuswb_r2r(mm5,mm4); // 16 -> 8 bit
movq_r2m(mm4, *puv);
UVtmp += 8;
puv += pitches[i];
} /* for row */
} // for i
emms();
} /* dv_mb420_YV12_mmx */
#endif // ARCH_X86 || ARCH_X86_64
|
