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
|
/*
* Copyright (c) 2019 Paul B Mahol
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg 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 General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include <pocketsphinx/pocketsphinx.h>
#include "libavutil/avstring.h"
#include "libavutil/channel_layout.h"
#include "libavutil/opt.h"
#include "audio.h"
#include "avfilter.h"
#include "internal.h"
typedef struct ASRContext {
const AVClass *class;
int rate;
char *hmm;
char *dict;
char *lm;
char *lmctl;
char *lmname;
char *logfn;
ps_decoder_t *ps;
cmd_ln_t *config;
int utt_started;
} ASRContext;
#define OFFSET(x) offsetof(ASRContext, x)
#define FLAGS AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_FILTERING_PARAM
static const AVOption asr_options[] = {
{ "rate", "set sampling rate", OFFSET(rate), AV_OPT_TYPE_INT, {.i64=16000}, 0, INT_MAX, .flags = FLAGS },
{ "hmm", "set directory containing acoustic model files", OFFSET(hmm), AV_OPT_TYPE_STRING, {.str=NULL}, .flags = FLAGS },
{ "dict", "set pronunciation dictionary", OFFSET(dict), AV_OPT_TYPE_STRING, {.str=NULL}, .flags = FLAGS },
{ "lm", "set language model file", OFFSET(lm), AV_OPT_TYPE_STRING, {.str=NULL}, .flags = FLAGS },
{ "lmctl", "set language model set", OFFSET(lmctl), AV_OPT_TYPE_STRING, {.str=NULL}, .flags = FLAGS },
{ "lmname","set which language model to use", OFFSET(lmname), AV_OPT_TYPE_STRING, {.str=NULL}, .flags = FLAGS },
{ "logfn", "set output for log messages", OFFSET(logfn), AV_OPT_TYPE_STRING, {.str="/dev/null"}, .flags = FLAGS },
{ NULL }
};
AVFILTER_DEFINE_CLASS(asr);
static int filter_frame(AVFilterLink *inlink, AVFrame *in)
{
AVFilterContext *ctx = inlink->dst;
AVDictionary **metadata = &in->metadata;
ASRContext *s = ctx->priv;
int have_speech;
const char *speech;
ps_process_raw(s->ps, (const int16_t *)in->data[0], in->nb_samples, 0, 0);
have_speech = ps_get_in_speech(s->ps);
if (have_speech && !s->utt_started)
s->utt_started = 1;
if (!have_speech && s->utt_started) {
ps_end_utt(s->ps);
speech = ps_get_hyp(s->ps, NULL);
if (speech != NULL)
av_dict_set(metadata, "lavfi.asr.text", speech, 0);
ps_start_utt(s->ps);
s->utt_started = 0;
}
return ff_filter_frame(ctx->outputs[0], in);
}
static int config_input(AVFilterLink *inlink)
{
AVFilterContext *ctx = inlink->dst;
ASRContext *s = ctx->priv;
ps_start_utt(s->ps);
return 0;
}
static av_cold int asr_init(AVFilterContext *ctx)
{
ASRContext *s = ctx->priv;
const float frate = s->rate;
char *rate = av_asprintf("%f", frate);
const char *argv[] = { "-logfn", s->logfn,
"-hmm", s->hmm,
"-lm", s->lm,
"-lmctl", s->lmctl,
"-lmname", s->lmname,
"-dict", s->dict,
"-samprate", rate,
NULL };
s->config = cmd_ln_parse_r(NULL, ps_args(), 14, (char **)argv, 0);
av_free(rate);
if (!s->config)
return AVERROR(ENOMEM);
ps_default_search_args(s->config);
s->ps = ps_init(s->config);
if (!s->ps)
return AVERROR(ENOMEM);
return 0;
}
static int query_formats(AVFilterContext *ctx)
{
ASRContext *s = ctx->priv;
int sample_rates[] = { s->rate, -1 };
int ret;
AVFilterFormats *formats = NULL;
AVFilterChannelLayouts *layout = NULL;
if ((ret = ff_add_format (&formats, AV_SAMPLE_FMT_S16 )) < 0 ||
(ret = ff_set_common_formats (ctx , formats )) < 0 ||
(ret = ff_add_channel_layout (&layout , AV_CH_LAYOUT_MONO )) < 0 ||
(ret = ff_set_common_channel_layouts (ctx , layout )) < 0 ||
(ret = ff_set_common_samplerates (ctx , ff_make_format_list(sample_rates) )) < 0)
return ret;
return 0;
}
static av_cold void asr_uninit(AVFilterContext *ctx)
{
ASRContext *s = ctx->priv;
ps_free(s->ps);
s->ps = NULL;
cmd_ln_free_r(s->config);
s->config = NULL;
}
static const AVFilterPad asr_inputs[] = {
{
.name = "default",
.type = AVMEDIA_TYPE_AUDIO,
.filter_frame = filter_frame,
.config_props = config_input,
},
{ NULL }
};
static const AVFilterPad asr_outputs[] = {
{
.name = "default",
.type = AVMEDIA_TYPE_AUDIO,
},
{ NULL }
};
const AVFilter ff_af_asr = {
.name = "asr",
.description = NULL_IF_CONFIG_SMALL("Automatic Speech Recognition."),
.priv_size = sizeof(ASRContext),
.priv_class = &asr_class,
.init = asr_init,
.uninit = asr_uninit,
.query_formats = query_formats,
.inputs = asr_inputs,
.outputs = asr_outputs,
};
|
