blob: feff9c0b6874736b5d3821589f082b9b85c7e9c6
1 | /* |
2 | * HQX DSP routines |
3 | * |
4 | * This file is part of FFmpeg. |
5 | * |
6 | * FFmpeg is free software; you can redistribute it and/or |
7 | * modify it under the terms of the GNU Lesser General Public |
8 | * License as published by the Free Software Foundation; either |
9 | * version 2.1 of the License, or (at your option) any later version. |
10 | * |
11 | * FFmpeg is distributed in the hope that it will be useful, |
12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
14 | * Lesser General Public License for more details. |
15 | * |
16 | * You should have received a copy of the GNU Lesser General Public |
17 | * License along with FFmpeg; if not, write to the Free Software |
18 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
19 | */ |
20 | |
21 | #include <stdint.h> |
22 | |
23 | #include "libavutil/common.h" |
24 | |
25 | #include "hqxdsp.h" |
26 | |
27 | static inline void idct_col(int16_t *blk, const uint8_t *quant) |
28 | { |
29 | int t0, t1, t2, t3, t4, t5, t6, t7, t8, t9, tA, tB, tC, tD, tE, tF; |
30 | int t10, t11, t12, t13; |
31 | int s0, s1, s2, s3, s4, s5, s6, s7; |
32 | |
33 | s0 = (int) blk[0 * 8] * quant[0 * 8]; |
34 | s1 = (int) blk[1 * 8] * quant[1 * 8]; |
35 | s2 = (int) blk[2 * 8] * quant[2 * 8]; |
36 | s3 = (int) blk[3 * 8] * quant[3 * 8]; |
37 | s4 = (int) blk[4 * 8] * quant[4 * 8]; |
38 | s5 = (int) blk[5 * 8] * quant[5 * 8]; |
39 | s6 = (int) blk[6 * 8] * quant[6 * 8]; |
40 | s7 = (int) blk[7 * 8] * quant[7 * 8]; |
41 | |
42 | t0 = (s3 * 19266 + s5 * 12873) >> 15; |
43 | t1 = (s5 * 19266 - s3 * 12873) >> 15; |
44 | t2 = ((s7 * 4520 + s1 * 22725) >> 15) - t0; |
45 | t3 = ((s1 * 4520 - s7 * 22725) >> 15) - t1; |
46 | t4 = t0 * 2 + t2; |
47 | t5 = t1 * 2 + t3; |
48 | t6 = t2 - t3; |
49 | t7 = t3 * 2 + t6; |
50 | t8 = (t6 * 11585) >> 14; |
51 | t9 = (t7 * 11585) >> 14; |
52 | tA = (s2 * 8867 - s6 * 21407) >> 14; |
53 | tB = (s6 * 8867 + s2 * 21407) >> 14; |
54 | tC = (s0 >> 1) - (s4 >> 1); |
55 | tD = (s4 >> 1) * 2 + tC; |
56 | tE = tC - (tA >> 1); |
57 | tF = tD - (tB >> 1); |
58 | t10 = tF - t5; |
59 | t11 = tE - t8; |
60 | t12 = tE + (tA >> 1) * 2 - t9; |
61 | t13 = tF + (tB >> 1) * 2 - t4; |
62 | |
63 | blk[0 * 8] = t13 + t4 * 2; |
64 | blk[1 * 8] = t12 + t9 * 2; |
65 | blk[2 * 8] = t11 + t8 * 2; |
66 | blk[3 * 8] = t10 + t5 * 2; |
67 | blk[4 * 8] = t10; |
68 | blk[5 * 8] = t11; |
69 | blk[6 * 8] = t12; |
70 | blk[7 * 8] = t13; |
71 | } |
72 | |
73 | static inline void idct_row(int16_t *blk) |
74 | { |
75 | int t0, t1, t2, t3, t4, t5, t6, t7, t8, t9, tA, tB, tC, tD, tE, tF; |
76 | int t10, t11, t12, t13; |
77 | |
78 | t0 = (blk[3] * 19266 + blk[5] * 12873) >> 14; |
79 | t1 = (blk[5] * 19266 - blk[3] * 12873) >> 14; |
80 | t2 = ((blk[7] * 4520 + blk[1] * 22725) >> 14) - t0; |
81 | t3 = ((blk[1] * 4520 - blk[7] * 22725) >> 14) - t1; |
82 | t4 = t0 * 2 + t2; |
83 | t5 = t1 * 2 + t3; |
84 | t6 = t2 - t3; |
85 | t7 = t3 * 2 + t6; |
86 | t8 = (t6 * 11585) >> 14; |
87 | t9 = (t7 * 11585) >> 14; |
88 | tA = (blk[2] * 8867 - blk[6] * 21407) >> 14; |
89 | tB = (blk[6] * 8867 + blk[2] * 21407) >> 14; |
90 | tC = blk[0] - blk[4]; |
91 | tD = blk[4] * 2 + tC; |
92 | tE = tC - tA; |
93 | tF = tD - tB; |
94 | t10 = tF - t5; |
95 | t11 = tE - t8; |
96 | t12 = tE + tA * 2 - t9; |
97 | t13 = tF + tB * 2 - t4; |
98 | |
99 | blk[0] = (t13 + t4 * 2 + 4) >> 3; |
100 | blk[1] = (t12 + t9 * 2 + 4) >> 3; |
101 | blk[2] = (t11 + t8 * 2 + 4) >> 3; |
102 | blk[3] = (t10 + t5 * 2 + 4) >> 3; |
103 | blk[4] = (t10 + 4) >> 3; |
104 | blk[5] = (t11 + 4) >> 3; |
105 | blk[6] = (t12 + 4) >> 3; |
106 | blk[7] = (t13 + 4) >> 3; |
107 | } |
108 | |
109 | static void hqx_idct_put(uint16_t *dst, ptrdiff_t stride, |
110 | int16_t *block, const uint8_t *quant) |
111 | { |
112 | int i, j; |
113 | |
114 | for (i = 0; i < 8; i++) |
115 | idct_col(block + i, quant + i); |
116 | for (i = 0; i < 8; i++) |
117 | idct_row(block + i * 8); |
118 | |
119 | for (i = 0; i < 8; i++) { |
120 | for (j = 0; j < 8; j++) { |
121 | int v = av_clip_uintp2(block[j + i * 8] + 0x800, 12); |
122 | dst[j] = (v << 4) | (v >> 8); |
123 | } |
124 | dst += stride >> 1; |
125 | } |
126 | } |
127 | |
128 | av_cold void ff_hqxdsp_init(HQXDSPContext *c) |
129 | { |
130 | c->idct_put = hqx_idct_put; |
131 | } |
132 |