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
|
/*
* Copyright (c) 2001-2016, Alliance for Open Media. All rights reserved
*
* This source code is subject to the terms of the BSD 2 Clause License and
* the Alliance for Open Media Patent License 1.0. If the BSD 2 Clause License
* was not distributed with this source code in the LICENSE file, you can
* obtain it at www.aomedia.org/license/software. If the Alliance for Open
* Media Patent License 1.0 was not distributed with this source code in the
* PATENTS file, you can obtain it at www.aomedia.org/license/patent.
*/
/* clang-format off */
#ifdef HAVE_CONFIG_H
# include "config.h"
#endif
#include <stdio.h>
#include "aom_dsp/bitwriter.h"
#include "av1/common/generic_code.h"
#include "av1/common/odintrin.h"
#include "pvq_encoder.h"
/** Encodes a value from 0 to N-1 (with N up to 16) based on a cdf and adapts
* the cdf accordingly.
*
* @param [in,out] w multi-symbol entropy encoder
* @param [in] val variable being encoded
* @param [in,out] cdf CDF of the variable (Q15)
* @param [in] n number of values possible
* @param [in,out] count number of symbols encoded with that cdf so far
* @param [in] rate adaptation rate shift (smaller is faster)
*/
void aom_encode_cdf_adapt_q15(aom_writer *w, int val, uint16_t *cdf, int n,
int *count, int rate) {
int i;
if (*count == 0) {
/* On the first call, we normalize the cdf to (32768 - n). This should
eventually be moved to the state init, but for now it makes it much
easier to experiment and convert symbols to the Q15 adaptation.*/
int ft;
ft = cdf[n - 1];
for (i = 0; i < n; i++) {
cdf[i] = AOM_ICDF(cdf[i]*32768/ft);
}
}
aom_write_cdf(w, val, cdf, n);
aom_cdf_adapt_q15(val, cdf, n, count, rate);
}
/** Encodes a random variable using a "generic" model, assuming that the
* distribution is one-sided (zero and up), has a single mode, and decays
* exponentially past the model.
*
* @param [in,out] w multi-symbol entropy encoder
* @param [in,out] model generic probability model
* @param [in] x variable being encoded
* @param [in,out] ExQ16 expectation of x (adapted)
* @param [in] integration integration period of ExQ16 (leaky average over
* 1<<integration samples)
*/
void generic_encode(aom_writer *w, generic_encoder *model, int x,
int *ex_q16, int integration) {
int lg_q1;
int shift;
int id;
uint16_t *cdf;
int xs;
lg_q1 = log_ex(*ex_q16);
OD_LOG((OD_LOG_ENTROPY_CODER, OD_LOG_DEBUG,
"%d %d", *ex_q16, lg_q1));
/* If expectation is too large, shift x to ensure that
all we have past xs=15 is the exponentially decaying tail
of the distribution */
shift = OD_MAXI(0, (lg_q1 - 5) >> 1);
/* Choose the cdf to use: we have two per "octave" of ExQ16 */
id = OD_MINI(GENERIC_TABLES - 1, lg_q1);
cdf = model->cdf[id];
xs = (x + (1 << shift >> 1)) >> shift;
aom_write_symbol_pvq(w, OD_MINI(15, xs), cdf, 16);
if (xs >= 15) {
int e;
unsigned decay;
/* Estimate decay based on the assumption that the distribution is close
to Laplacian for large values. We should probably have an adaptive
estimate instead. Note: The 2* is a kludge that's not fully understood
yet. */
OD_ASSERT(*ex_q16 < INT_MAX >> 1);
e = ((2**ex_q16 >> 8) + (1 << shift >> 1)) >> shift;
decay = OD_MAXI(2, OD_MINI(254, 256*e/(e + 256)));
/* Encode the tail of the distribution assuming exponential decay. */
aom_laplace_encode_special(w, xs - 15, decay);
}
if (shift != 0) {
int special;
/* Because of the rounding, there's only half the number of possibilities
for xs=0. */
special = xs == 0;
if (shift - special > 0) {
aom_write_literal(w, x - (xs << shift) + (!special << (shift - 1)),
shift - special);
}
}
generic_model_update(ex_q16, x, integration);
OD_LOG((OD_LOG_ENTROPY_CODER, OD_LOG_DEBUG,
"enc: %d %d %d %d %d %x", *ex_q16, x, shift, id, xs, enc->rng));
}
/** Estimates the cost of encoding a value with generic_encode().
*
* @param [in,out] model generic probability model
* @param [in] x variable being encoded
* @param [in,out] ExQ16 expectation of x (adapted)
* @return number of bits (approximation)
*/
double generic_encode_cost(generic_encoder *model, int x, int *ex_q16) {
int lg_q1;
int shift;
int id;
uint16_t *cdf;
int xs;
int extra;
lg_q1 = log_ex(*ex_q16);
/* If expectation is too large, shift x to ensure that
all we have past xs=15 is the exponentially decaying tail
of the distribution */
shift = OD_MAXI(0, (lg_q1 - 5) >> 1);
/* Choose the cdf to use: we have two per "octave" of ExQ16 */
id = OD_MINI(GENERIC_TABLES - 1, lg_q1);
cdf = model->cdf[id];
xs = (x + (1 << shift >> 1)) >> shift;
extra = 0;
if (shift) extra = shift - (xs == 0);
xs = OD_MINI(15, xs);
/* Shortcut: assume it's going to cost 2 bits for the Laplace coder. */
if (xs == 15) extra += 2;
return
extra - OD_LOG2((double)(cdf[xs] - (xs == 0 ? 0 : cdf[xs - 1]))/cdf[15]);
}
/*Estimates the cost of encoding a value with a given CDF.*/
double od_encode_cdf_cost(int val, uint16_t *cdf, int n) {
int total_prob;
int prev_prob;
double val_prob;
OD_ASSERT(n > 0);
total_prob = cdf[n - 1];
if (val == 0) {
prev_prob = 0;
}
else {
prev_prob = cdf[val - 1];
}
val_prob = (cdf[val] - prev_prob) / (double)total_prob;
return -OD_LOG2(val_prob);
}
|
