FFmpeg
asrc_flite.c
Go to the documentation of this file.
1 /*
2  * Copyright (c) 2012 Stefano Sabatini
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 /**
22  * @file
23  * flite voice synth source
24  */
25 
26 #include <flite/flite.h>
27 #include "libavutil/channel_layout.h"
28 #include "libavutil/file.h"
29 #include "libavutil/opt.h"
30 #include "avfilter.h"
31 #include "audio.h"
32 #include "formats.h"
33 #include "internal.h"
34 
35 typedef struct FliteContext {
36  const AVClass *class;
37  char *voice_str;
38  char *textfile;
39  char *text;
40  cst_wave *wave;
41  int16_t *wave_samples;
42  int wave_nb_samples;
43  int list_voices;
44  cst_voice *voice;
45  struct voice_entry *voice_entry;
46  int64_t pts;
47  int frame_nb_samples; ///< number of samples per frame
48 } FliteContext;
49 
50 #define OFFSET(x) offsetof(FliteContext, x)
51 #define FLAGS AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_FILTERING_PARAM
52 
53 static const AVOption flite_options[] = {
54  { "list_voices", "list voices and exit", OFFSET(list_voices), AV_OPT_TYPE_BOOL, {.i64=0}, 0, 1, FLAGS },
55  { "nb_samples", "set number of samples per frame", OFFSET(frame_nb_samples), AV_OPT_TYPE_INT, {.i64=512}, 0, INT_MAX, FLAGS },
56  { "n", "set number of samples per frame", OFFSET(frame_nb_samples), AV_OPT_TYPE_INT, {.i64=512}, 0, INT_MAX, FLAGS },
57  { "text", "set text to speak", OFFSET(text), AV_OPT_TYPE_STRING, {.str=NULL}, 0, 0, FLAGS },
58  { "textfile", "set filename of the text to speak", OFFSET(textfile), AV_OPT_TYPE_STRING, {.str=NULL}, 0, 0, FLAGS },
59  { "v", "set voice", OFFSET(voice_str), AV_OPT_TYPE_STRING, {.str="kal"}, 0, 0, FLAGS },
60  { "voice", "set voice", OFFSET(voice_str), AV_OPT_TYPE_STRING, {.str="kal"}, 0, 0, FLAGS },
61  { NULL }
62 };
63 
64 AVFILTER_DEFINE_CLASS(flite);
65 
66 static volatile int flite_inited = 0;
67 
68 /* declare functions for all the supported voices */
69 #define DECLARE_REGISTER_VOICE_FN(name) \
70  cst_voice *register_cmu_us_## name(const char *); \
71  void unregister_cmu_us_## name(cst_voice *);
72 DECLARE_REGISTER_VOICE_FN(awb);
73 DECLARE_REGISTER_VOICE_FN(kal);
74 DECLARE_REGISTER_VOICE_FN(kal16);
75 DECLARE_REGISTER_VOICE_FN(rms);
76 DECLARE_REGISTER_VOICE_FN(slt);
77 
78 struct voice_entry {
79  const char *name;
80  cst_voice * (*register_fn)(const char *);
81  void (*unregister_fn)(cst_voice *);
82  cst_voice *voice;
83  unsigned usage_count;
84 } voice_entry;
85 
86 #define MAKE_VOICE_STRUCTURE(voice_name) { \
87  .name = #voice_name, \
88  .register_fn = register_cmu_us_ ## voice_name, \
89  .unregister_fn = unregister_cmu_us_ ## voice_name, \
90 }
91 static struct voice_entry voice_entries[] = {
92  MAKE_VOICE_STRUCTURE(awb),
93  MAKE_VOICE_STRUCTURE(kal),
94  MAKE_VOICE_STRUCTURE(kal16),
95  MAKE_VOICE_STRUCTURE(rms),
96  MAKE_VOICE_STRUCTURE(slt),
97 };
98 
99 static void list_voices(void *log_ctx, const char *sep)
100 {
101  int i, n = FF_ARRAY_ELEMS(voice_entries);
102  for (i = 0; i < n; i++)
103  av_log(log_ctx, AV_LOG_INFO, "%s%s",
104  voice_entries[i].name, i < (n-1) ? sep : "\n");
105 }
106 
107 static int select_voice(struct voice_entry **entry_ret, const char *voice_name, void *log_ctx)
108 {
109  int i;
110 
111  for (i = 0; i < FF_ARRAY_ELEMS(voice_entries); i++) {
112  struct voice_entry *entry = &voice_entries[i];
113  if (!strcmp(entry->name, voice_name)) {
114  if (!entry->voice)
115  entry->voice = entry->register_fn(NULL);
116  if (!entry->voice) {
117  av_log(log_ctx, AV_LOG_ERROR,
118  "Could not register voice '%s'\n", voice_name);
119  return AVERROR_UNKNOWN;
120  }
121  entry->usage_count++;
122  *entry_ret = entry;
123  return 0;
124  }
125  }
126 
127  av_log(log_ctx, AV_LOG_ERROR, "Could not find voice '%s'\n", voice_name);
128  av_log(log_ctx, AV_LOG_INFO, "Choose between the voices: ");
129  list_voices(log_ctx, ", ");
130 
131  return AVERROR(EINVAL);
132 }
133 
134 static av_cold int init(AVFilterContext *ctx)
135 {
136  FliteContext *flite = ctx->priv;
137  int ret = 0;
138 
139  if (flite->list_voices) {
140  list_voices(ctx, "\n");
141  return AVERROR_EXIT;
142  }
143 
144  if (!flite_inited) {
145  if (flite_init() < 0) {
146  av_log(ctx, AV_LOG_ERROR, "flite initialization failed\n");
147  return AVERROR_UNKNOWN;
148  }
149  flite_inited++;
150  }
151 
152  if ((ret = select_voice(&flite->voice_entry, flite->voice_str, ctx)) < 0)
153  return ret;
154  flite->voice = flite->voice_entry->voice;
155 
156  if (flite->textfile && flite->text) {
157  av_log(ctx, AV_LOG_ERROR,
158  "Both text and textfile options set: only one must be specified\n");
159  return AVERROR(EINVAL);
160  }
161 
162  if (flite->textfile) {
163  uint8_t *textbuf;
164  size_t textbuf_size;
165 
166  if ((ret = av_file_map(flite->textfile, &textbuf, &textbuf_size, 0, ctx)) < 0) {
167  av_log(ctx, AV_LOG_ERROR,
168  "The text file '%s' could not be read: %s\n",
169  flite->textfile, av_err2str(ret));
170  return ret;
171  }
172 
173  if (!(flite->text = av_malloc(textbuf_size+1))) {
174  av_file_unmap(textbuf, textbuf_size);
175  return AVERROR(ENOMEM);
176  }
177  memcpy(flite->text, textbuf, textbuf_size);
178  flite->text[textbuf_size] = 0;
179  av_file_unmap(textbuf, textbuf_size);
180  }
181 
182  if (!flite->text) {
183  av_log(ctx, AV_LOG_ERROR,
184  "No speech text specified, specify the 'text' or 'textfile' option\n");
185  return AVERROR(EINVAL);
186  }
187 
188  /* synth all the file data in block */
189  flite->wave = flite_text_to_wave(flite->text, flite->voice);
190  flite->wave_samples = flite->wave->samples;
191  flite->wave_nb_samples = flite->wave->num_samples;
192  return 0;
193 }
194 
195 static av_cold void uninit(AVFilterContext *ctx)
196 {
197  FliteContext *flite = ctx->priv;
198 
199  if (flite->voice_entry) {
200  if (!--flite->voice_entry->usage_count) {
201  flite->voice_entry->unregister_fn(flite->voice);
202  flite->voice_entry->voice = NULL;
203  }
204  }
205  delete_wave(flite->wave);
206  flite->wave = NULL;
207 }
208 
209 static int query_formats(AVFilterContext *ctx)
210 {
211  FliteContext *flite = ctx->priv;
212  int ret;
213 
214  AVFilterChannelLayouts *chlayouts = NULL;
215  int64_t chlayout = av_get_default_channel_layout(flite->wave->num_channels);
216  AVFilterFormats *sample_formats = NULL;
217  AVFilterFormats *sample_rates = NULL;
218 
219  if ((ret = ff_add_channel_layout (&chlayouts , chlayout )) < 0 ||
220  (ret = ff_set_common_channel_layouts (ctx , chlayouts )) < 0 ||
221  (ret = ff_add_format (&sample_formats, AV_SAMPLE_FMT_S16 )) < 0 ||
222  (ret = ff_set_common_formats (ctx , sample_formats )) < 0 ||
223  (ret = ff_add_format (&sample_rates , flite->wave->sample_rate)) < 0 ||
224  (ret = ff_set_common_samplerates (ctx , sample_rates )) < 0)
225  return ret;
226 
227  return 0;
228 }
229 
230 static int config_props(AVFilterLink *outlink)
231 {
232  AVFilterContext *ctx = outlink->src;
233  FliteContext *flite = ctx->priv;
234 
235  outlink->sample_rate = flite->wave->sample_rate;
236  outlink->time_base = (AVRational){1, flite->wave->sample_rate};
237 
238  av_log(ctx, AV_LOG_VERBOSE, "voice:%s fmt:%s sample_rate:%d\n",
239  flite->voice_str,
240  av_get_sample_fmt_name(outlink->format), outlink->sample_rate);
241  return 0;
242 }
243 
244 static int request_frame(AVFilterLink *outlink)
245 {
246  AVFrame *samplesref;
247  FliteContext *flite = outlink->src->priv;
248  int nb_samples = FFMIN(flite->wave_nb_samples, flite->frame_nb_samples);
249 
250  if (!nb_samples)
251  return AVERROR_EOF;
252 
253  samplesref = ff_get_audio_buffer(outlink, nb_samples);
254  if (!samplesref)
255  return AVERROR(ENOMEM);
256 
257  memcpy(samplesref->data[0], flite->wave_samples,
258  nb_samples * flite->wave->num_channels * 2);
259  samplesref->pts = flite->pts;
260  samplesref->pkt_pos = -1;
261  samplesref->sample_rate = flite->wave->sample_rate;
262  flite->pts += nb_samples;
263  flite->wave_samples += nb_samples * flite->wave->num_channels;
264  flite->wave_nb_samples -= nb_samples;
265 
266  return ff_filter_frame(outlink, samplesref);
267 }
268 
269 static const AVFilterPad flite_outputs[] = {
270  {
271  .name = "default",
272  .type = AVMEDIA_TYPE_AUDIO,
273  .config_props = config_props,
274  .request_frame = request_frame,
275  },
276  { NULL }
277 };
278 
279 AVFilter ff_asrc_flite = {
280  .name = "flite",
281  .description = NULL_IF_CONFIG_SMALL("Synthesize voice from text using libflite."),
282  .query_formats = query_formats,
283  .init = init,
284  .uninit = uninit,
285  .priv_size = sizeof(FliteContext),
286  .inputs = NULL,
287  .outputs = flite_outputs,
288  .priv_class = &flite_class,
289 };
OFFSET
#define OFFSET(x)
Definition: asrc_flite.c:50
ff_get_audio_buffer
AVFrame * ff_get_audio_buffer(AVFilterLink *link, int nb_samples)
Request an audio samples buffer with a specific set of permissions.
Definition: audio.c:86
AVFilterChannelLayouts
A list of supported channel layouts.
Definition: formats.h:86
name
it s the only field you need to keep assuming you have a context There is some magic you don t need to care about around this just let it vf default minimum maximum flags name is the option name
Definition: writing_filters.txt:88
AVERROR
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format for each input and each output the list of supported formats For video that means pixel format For audio that means channel sample they are references to shared objects When the negotiation mechanism computes the intersection of the formats supported at each end of a all references to both lists are replaced with a reference to the intersection And when a single format is eventually chosen for a link amongst the remaining all references to the list are updated That means that if a filter requires that its input and output have the same format amongst a supported all it has to do is use a reference to the same list of formats query_formats can leave some formats unset and return AVERROR(EAGAIN) to cause the negotiation mechanism toagain later. That can be used by filters with complex requirements to use the format negotiated on one link to set the formats supported on another. Frame references ownership and permissions
flite_options
static const AVOption flite_options[]
Definition: asrc_flite.c:53
opt.h
ff_filter_frame
int ff_filter_frame(AVFilterLink *link, AVFrame *frame)
Send a frame of data to the next filter.
Definition: avfilter.c:1096
AVERROR_EOF
#define AVERROR_EOF
End of file.
Definition: error.h:55
FliteContext::textfile
char * textfile
Definition: asrc_flite.c:38
voice_entry::voice
cst_voice * voice
Definition: asrc_flite.c:82
AVFrame
This structure describes decoded (raw) audio or video data.
Definition: frame.h:318
AVFrame::pts
int64_t pts
Presentation timestamp in time_base units (time when frame should be shown to user).
Definition: frame.h:411
AVOption
AVOption.
Definition: opt.h:248
flite_inited
static volatile int flite_inited
Definition: asrc_flite.c:66
voice_entry::usage_count
unsigned usage_count
Definition: asrc_flite.c:83
AV_LOG_VERBOSE
#define AV_LOG_VERBOSE
Detailed information.
Definition: log.h:210
AVFilter::name
const char * name
Filter name.
Definition: avfilter.h:149
AVERROR_UNKNOWN
#define AVERROR_UNKNOWN
Unknown error, typically from an external library.
Definition: error.h:71
FliteContext::voice_entry
struct voice_entry * voice_entry
Definition: asrc_flite.c:45
FliteContext
Definition: asrc_flite.c:35
AVFILTER_DEFINE_CLASS
AVFILTER_DEFINE_CLASS(flite)
ff_asrc_flite
AVFilter ff_asrc_flite
Definition: asrc_flite.c:279
voice_entry::register_fn
cst_voice *(* register_fn)(const char *)
Definition: asrc_flite.c:80
AVFrame::data
uint8_t * data[AV_NUM_DATA_POINTERS]
pointer to the picture/channel planes.
Definition: frame.h:332
av_malloc
#define av_malloc(s)
Definition: tableprint_vlc.h:31
MAKE_VOICE_STRUCTURE
#define MAKE_VOICE_STRUCTURE(voice_name)
Definition: asrc_flite.c:86
AVFilterFormats
A list of supported formats for one end of a filter link.
Definition: formats.h:65
formats.h
av_file_map
int av_file_map(const char *filename, uint8_t **bufptr, size_t *size, int log_offset, void *log_ctx)
Read the file with name filename, and put its content in a newly allocated buffer or map it with mmap...
Definition: file.c:53
AVFilterContext::priv
void * priv
private data for use by the filter
Definition: avfilter.h:356
FliteContext::wave_samples
int16_t * wave_samples
Definition: asrc_flite.c:41
AVFilterPad
A filter pad used for either input or output.
Definition: internal.h:54
FliteContext::list_voices
int list_voices
Definition: asrc_flite.c:43
AV_LOG_ERROR
#define AV_LOG_ERROR
Something went wrong and cannot losslessly be recovered.
Definition: log.h:194
FF_ARRAY_ELEMS
#define FF_ARRAY_ELEMS(a)
Definition: sinewin_tablegen.c:29
av_cold
#define av_cold
Definition: attributes.h:90
ff_set_common_formats
int ff_set_common_formats(AVFilterContext *ctx, AVFilterFormats *formats)
A helper for query_formats() which sets all links to the same list of formats.
Definition: formats.c:587
ff_add_channel_layout
int ff_add_channel_layout(AVFilterChannelLayouts **l, uint64_t channel_layout)
Definition: formats.c:338
AVFrame::pkt_pos
int64_t pkt_pos
reordered pos from the last AVPacket that has been input into the decoder
Definition: frame.h:589
AVMEDIA_TYPE_AUDIO
@ AVMEDIA_TYPE_AUDIO
Definition: avutil.h:202
outputs
static const AVFilterPad outputs[]
Definition: af_acontrast.c:203
list_voices
static void list_voices(void *log_ctx, const char *sep)
Definition: asrc_flite.c:99
ctx
AVFormatContext * ctx
Definition: movenc.c:48
av_get_sample_fmt_name
const char * av_get_sample_fmt_name(enum AVSampleFormat sample_fmt)
Return the name of sample_fmt, or NULL if sample_fmt is not recognized.
Definition: samplefmt.c:49
av_file_unmap
void av_file_unmap(uint8_t *bufptr, size_t size)
Unmap or free the buffer bufptr created by av_file_map().
Definition: file.c:144
AVClass
Describe the class of an AVClass context structure.
Definition: log.h:67
NULL
#define NULL
Definition: coverity.c:32
flite_outputs
static const AVFilterPad flite_outputs[]
Definition: asrc_flite.c:269
voice_entries
static struct voice_entry voice_entries[]
Definition: asrc_flite.c:91
query_formats
static int query_formats(AVFilterContext *ctx)
Definition: asrc_flite.c:209
AVRational
Rational number (pair of numerator and denominator).
Definition: rational.h:58
ff_add_format
int ff_add_format(AVFilterFormats **avff, int64_t fmt)
Add fmt to the list of media formats contained in *avff.
Definition: formats.c:332
inputs
these buffered frames must be flushed immediately if a new input produces new the filter must not call request_frame to get more It must just process the frame or queue it The task of requesting more frames is left to the filter s request_frame method or the application If a filter has several inputs
Definition: filter_design.txt:243
voice_entry::unregister_fn
void(* unregister_fn)(cst_voice *)
Definition: asrc_flite.c:81
FliteContext::voice_str
char * voice_str
Definition: asrc_flite.c:37
NULL_IF_CONFIG_SMALL
#define NULL_IF_CONFIG_SMALL(x)
Return NULL if CONFIG_SMALL is true, otherwise the argument without modification.
Definition: internal.h:117
av_err2str
#define av_err2str(errnum)
Convenience macro, the return value should be used only directly in function arguments but never stan...
Definition: error.h:119
AVFrame::sample_rate
int sample_rate
Sample rate of the audio data.
Definition: frame.h:490
FFMIN
#define FFMIN(a, b)
Definition: common.h:105
config_props
static int config_props(AVFilterLink *outlink)
Definition: asrc_flite.c:230
AV_LOG_INFO
#define AV_LOG_INFO
Standard information.
Definition: log.h:205
sample_rates
sample_rates
Definition: ffmpeg_filter.c:170
internal.h
FliteContext::pts
int64_t pts
Definition: asrc_flite.c:46
i
int i
Definition: input.c:407
uint8_t
uint8_t
Definition: audio_convert.c:194
AV_SAMPLE_FMT_S16
@ AV_SAMPLE_FMT_S16
signed 16 bits
Definition: samplefmt.h:61
FliteContext::text
char * text
Definition: asrc_flite.c:39
AVFilterPad::name
const char * name
Pad name.
Definition: internal.h:60
AVFilter
Filter definition.
Definition: avfilter.h:145
ret
ret
Definition: filter_design.txt:187
uninit
static av_cold void uninit(AVFilterContext *ctx)
Definition: asrc_flite.c:195
channel_layout.h
AV_OPT_TYPE_INT
@ AV_OPT_TYPE_INT
Definition: opt.h:225
avfilter.h
init
static av_cold int init(AVFilterContext *ctx)
Definition: asrc_flite.c:134
voice_entry
struct voice_entry voice_entry
file.h
AVFilterContext
An instance of a filter.
Definition: avfilter.h:341
FLAGS
#define FLAGS
Definition: asrc_flite.c:51
DECLARE_REGISTER_VOICE_FN
#define DECLARE_REGISTER_VOICE_FN(name)
Definition: asrc_flite.c:69
audio.h
FliteContext::wave_nb_samples
int wave_nb_samples
Definition: asrc_flite.c:42
av_get_default_channel_layout
int64_t av_get_default_channel_layout(int nb_channels)
Return default channel layout for a given number of channels.
Definition: channel_layout.c:231
AV_OPT_TYPE_BOOL
@ AV_OPT_TYPE_BOOL
Definition: opt.h:242
av_log
#define av_log(a,...)
Definition: tableprint_vlc.h:28
select_voice
static int select_voice(struct voice_entry **entry_ret, const char *voice_name, void *log_ctx)
Definition: asrc_flite.c:107
ff_set_common_samplerates
int ff_set_common_samplerates(AVFilterContext *ctx, AVFilterFormats *samplerates)
Definition: formats.c:575
AVERROR_EXIT
#define AVERROR_EXIT
Immediate exit was requested; the called function should not be restarted.
Definition: error.h:56
AV_OPT_TYPE_STRING
@ AV_OPT_TYPE_STRING
Definition: opt.h:229
FliteContext::frame_nb_samples
int frame_nb_samples
number of samples per frame
Definition: asrc_flite.c:47
FliteContext::voice
cst_voice * voice
Definition: asrc_flite.c:44
voice_entry
Definition: asrc_flite.c:78
voice_entry::name
const char * name
Definition: asrc_flite.c:79
request_frame
static int request_frame(AVFilterLink *outlink)
Definition: asrc_flite.c:244
ff_set_common_channel_layouts
int ff_set_common_channel_layouts(AVFilterContext *ctx, AVFilterChannelLayouts *channel_layouts)
A helper for query_formats() which sets all links to the same list of channel layouts/sample rates.
Definition: formats.c:568
FliteContext::wave
cst_wave * wave
Definition: asrc_flite.c:40
Generated on Wed Aug 24 2022 21:34:27 for FFmpeg by   doxygen 1.8.17