cst_utterance *utt_init(cst_utterance *u, cst_voice *vox)
{
feat_copy_into(vox->features,u->features);
feat_copy_into(vox->ffunctions,u->ffunctions);
if (vox->utt_init)
vox->utt_init(u, vox);
return u;
}
static void add_raw_data(cst_utterance *u, const char *raw_data,
cst_features *attributes)
{
/* Add all tokens in raw _data to u */
cst_tokenstream *ts;
cst_relation *r;
cst_item *t;
const char *token;
r = utt_relation_create(u,"Token");
ts =
ts_open_string(raw_data,
get_param_string(u->features,"text_whitespace",NULL),
get_param_string(u->features,"text_singlecharsymbols",NULL),
get_param_string(u->features,"text_prepunctuation",NULL),
get_param_string(u->features,"text_pospunctuation",NULL));
while (!(ts_eof(ts)))
{
t = relation_append(r,NULL);
feat_copy_into(item_feats(t),attributes);
token = ts_get(ts);
if (cst_strlen(token) > 0)
{
t = relation_append(r,NULL);
item_set_string(t,"name",token);
item_set_string(t,"whitespace",ts->whitespace);
item_set_string(t,"prepunctuation",ts->prepunctuation);
item_set_string(t,"punc",ts->postpunctuation);
}
}
}
cst_utterance *default_textanalysis(cst_utterance *u)
{
cst_item *t,*word;
cst_relation *word_rel;
cst_val *words;
const cst_val *w;
const cst_val *ttwv;
word_rel = utt_relation_create(u,"Word");
ttwv = feat_val(u->features, "tokentowords_func");
for (t=relation_head(utt_relation(u,"Token")); t; t=item_next(t))
{
if (ttwv)
words = (cst_val *)(*val_itemfunc(ttwv))(t);
else
words = default_tokentowords(t);
for (w=words; w; w=val_cdr(w))
{
word = item_add_daughter(t,NULL);
if (cst_val_consp(val_car(w)))
{ /* Has extra features */
item_set_string(word,"name",val_string(val_car(val_car(w))));
feat_copy_into(val_features(val_cdr(val_car(w))),
item_feats(word));
}
else
item_set_string(word,"name",val_string(val_car(w)));
relation_append(word_rel,word);
}
delete_val(words);
}
return u;
}
int main(int argc, char **argv)
{
struct timeval tv;
cst_voice *v;
const char *filename;
const char *outtype;
int i;
float durs;
double time_start, time_end;
int flite_verbose, flite_loop, flite_bench;
int explicit_filename, explicit_text, explicit_phones;
#define ITER_MAX 3
int bench_iter = 0;
cst_features *extra_feats;
filename = 0;
outtype = "play"; /* default is to play */
flite_verbose = FALSE;
flite_loop = FALSE;
flite_bench = FALSE;
explicit_text = explicit_filename = explicit_phones = FALSE;
extra_feats = new_features();
flite_init();
for (i=1; i<argc; i++)
{
if (cst_streq(argv[i],"--version"))
{
flite_version();
return 1;
}
else if (cst_streq(argv[i],"-h") ||
cst_streq(argv[i],"--help") ||
cst_streq(argv[i],"-?"))
flite_usage();
else if (cst_streq(argv[i],"-v"))
flite_verbose = TRUE;
else if (cst_streq(argv[i],"-l"))
flite_loop = TRUE;
else if (cst_streq(argv[i],"-b"))
{
flite_bench = TRUE;
break; /* ignore other arguments */
}
else if ((cst_streq(argv[i],"-o")) && (i+1 < argc))
{
outtype = argv[i+1];
i++;
}
else if (cst_streq(argv[i],"-f") && (i+1 < argc))
{
filename = argv[i+1];
explicit_filename = TRUE;
i++;
}
else if (cst_streq(argv[i],"-pw"))
{
feat_set_string(extra_feats,"print_info_relation","Word");
feat_set(extra_feats,"post_synth_hook_func",
uttfunc_val(&print_info));
}
else if (cst_streq(argv[i],"-ps"))
{
feat_set_string(extra_feats,"print_info_relation","Segment");
feat_set(extra_feats,"post_synth_hook_func",
uttfunc_val(&print_info));
}
else if (cst_streq(argv[i],"-pr") && (i+1 < argc))
{
feat_set_string(extra_feats,"print_info_relation",argv[i+1]);
feat_set(extra_feats,"post_synth_hook_func",
uttfunc_val(&print_info));
i++;
}
else if ((cst_streq(argv[i],"-set") || cst_streq(argv[i],"-s"))
&& (i+1 < argc))
{
ef_set(extra_feats,argv[i+1],0);
i++;
}
else if (cst_streq(argv[i],"--seti") && (i+1 < argc))
{
ef_set(extra_feats,argv[i+1],"int");
i++;
}
else if (cst_streq(argv[i],"--setf") && (i+1 < argc))
{
ef_set(extra_feats,argv[i+1],"float");
i++;
}
else if (cst_streq(argv[i],"--sets") && (i+1 < argc))
{
ef_set(extra_feats,argv[i+1],"string");
i++;
}
else if (cst_streq(argv[i],"-p") && (i+1 < argc))
{
filename = argv[i+1];
explicit_phones = TRUE;
i++;
}
else if (cst_streq(argv[i],"-t") && (i+1 < argc))
{
filename = argv[i+1];
explicit_text = TRUE;
i++;
}
else if (filename)
outtype = argv[i];
else
filename = argv[i];
}
if (filename == NULL) filename = "-"; /* stdin */
v = REGISTER_VOX(NULL);
feat_copy_into(extra_feats,v->features);
durs = 0.0;
if (flite_bench)
{
outtype = "none";
filename = "A whole joy was reaping, but they've gone south, you should fetch azure mike.";
explicit_text = TRUE;
}
loop:
gettimeofday(&tv,NULL);
time_start = (double)(tv.tv_sec)+(((double)tv.tv_usec)/1000000.0);
if (explicit_phones)
durs = flite_phones_to_speech(filename,v,outtype);
else if ((strchr(filename,' ') && !explicit_filename) || explicit_text)
durs = flite_text_to_speech(filename,v,outtype);
else
durs = flite_file_to_speech(filename,v,outtype);
gettimeofday(&tv,NULL);
time_end = ((double)(tv.tv_sec))+((double)tv.tv_usec/1000000.0);
if (flite_verbose || (flite_bench && bench_iter == ITER_MAX))
printf("times faster than real-time: %f\n(%f seconds of speech synthesized in %f)\n",
durs/(float)(time_end-time_start),
durs,
(float)(time_end-time_start));
if (flite_loop || (flite_bench && bench_iter++ < ITER_MAX))
goto loop;
delete_features(extra_feats);
UNREGISTER_VOX(v);
return 0;
}
int main(int argc, char **argv)
{
cst_voice *v;
const char *filename;
const char *outtype;
int i;
float durs;
int flite_verbose, flite_loop, flite_bench;
int explicit_filename, explicit_text, explicit_phones;
#define ITER_MAX 3
int bench_iter = 0;
cst_features *extra_feats;
InitThreads(HT_NOMONITOR); // enable msg driven monitoring
if(InitShell(argc,argv,version)<SUCCESS) return FAIL;
InitMem(); InitLabel();
InitMath(); InitSigP(); InitUtil();
InitWave(); InitAudio();
filename = 0;
outtype = "play"; /* default is to play */
flite_verbose = FALSE;
flite_loop = FALSE;
flite_bench = FALSE;
explicit_text = explicit_filename = explicit_phones = FALSE;
extra_feats = new_features();
flite_init();
for (i=1; i<argc; i++)
{
if (cst_streq(argv[i],"--version"))
{
flite_version();
return 1;
}
else if (cst_streq(argv[i],"-h") ||
cst_streq(argv[i],"--help") ||
cst_streq(argv[i],"-?"))
flite_usage();
else if (cst_streq(argv[i],"-v"))
flite_verbose = TRUE;
else if (cst_streq(argv[i],"-l"))
flite_loop = TRUE;
else if (cst_streq(argv[i],"-b"))
{
flite_bench = TRUE;
break; /* ignore other arguments */
}
else if ((cst_streq(argv[i],"-o")) && (i+1 < argc))
{
outtype = argv[i+1];
i++;
}
else if (cst_streq(argv[i],"-f") && (i+1 < argc))
{
filename = argv[i+1];
explicit_filename = TRUE;
i++;
}
else if ((cst_streq(argv[i],"-set") || cst_streq(argv[i],"-s"))
&& (i+1 < argc))
{
ef_set(extra_feats,argv[i+1],0);
i++;
}
else if (cst_streq(argv[i],"--seti") && (i+1 < argc))
{
ef_set(extra_feats,argv[i+1],"int");
i++;
}
else if (cst_streq(argv[i],"--setf") && (i+1 < argc))
{
ef_set(extra_feats,argv[i+1],"float");
i++;
}
else if (cst_streq(argv[i],"--sets") && (i+1 < argc))
{
ef_set(extra_feats,argv[i+1],"string");
i++;
}
else if (cst_streq(argv[i],"-p") && (i+1 < argc))
{
filename = argv[i+1];
explicit_phones = TRUE;
i++;
}
else if (cst_streq(argv[i],"-t") && (i+1 < argc))
{
filename = argv[i+1];
explicit_text = TRUE;
i++;
}
else if (filename)
outtype = argv[i];
else
filename = argv[i];
}
if (filename == NULL) filename = "-"; /* stdin */
v = register_cmu_us_kal16(NULL);
feat_copy_into(extra_feats,v->features);
durs = 0.0;
if (flite_bench)
{
outtype = "none";
filename = "A whole joy was reaping, but they've gone south, you should fetch azure mike.";
explicit_text = TRUE;
}
loop:
if (explicit_phones)
durs = flite_phones_to_speech(filename,v,outtype);
else if ((strchr(filename,' ') && !explicit_filename) || explicit_text)
durs = flite_text_to_speech(filename,v,outtype);
else
durs = flite_file_to_speech(filename,v,outtype);
if (flite_loop || (flite_bench && bench_iter++ < ITER_MAX))
goto loop;
delete_features(extra_feats);
unregister_cmu_us_kal16(v);
return 0;
}
int main(int argc, char **argv)
{
struct timeval tv;
cst_voice *v;
const char *filename;
const char *outtype;
cst_voice *desired_voice = 0;
const char *voicedir = NULL;
int i;
float durs;
double time_start, time_end;
int flite_verbose, flite_loop, flite_bench;
int explicit_filename, explicit_text, explicit_phones, ssml_mode;
#define ITER_MAX 3
int bench_iter = 0;
cst_features *extra_feats;
const char *lex_addenda_file = NULL;
const char *voicedumpfile = NULL;
cst_audio_streaming_info *asi;
filename = 0;
outtype = "play"; /* default is to play */
flite_verbose = FALSE;
flite_loop = FALSE;
flite_bench = FALSE;
explicit_text = explicit_filename = explicit_phones = FALSE;
ssml_mode = FALSE;
extra_feats = new_features();
flite_init();
flite_add_lang("eng",usenglish_init,cmu_lex_init);
for (i=1; i<argc; i++)
{
if (cst_streq(argv[i],"--version"))
{
flite_version();
return 1;
}
else if (cst_streq(argv[i],"-h") ||
cst_streq(argv[i],"--help") ||
cst_streq(argv[i],"-?"))
flite_usage();
else if (cst_streq(argv[i],"-v"))
flite_verbose = TRUE;
else if (cst_streq(argv[i],"-lv"))
{
if (flite_voice_list == NULL)
flite_set_voice_list(voicedir);
flite_voice_list_print();
exit(0);
}
else if (cst_streq(argv[i],"-l"))
flite_loop = TRUE;
else if (cst_streq(argv[i],"-b"))
{
flite_bench = TRUE;
break; /* ignore other arguments */
}
else if ((cst_streq(argv[i],"-o")) && (i+1 < argc))
{
outtype = argv[i+1];
i++;
}
else if ((cst_streq(argv[i],"-voice")) && (i+1 < argc))
{
if (flite_voice_list == NULL)
flite_set_voice_list(voicedir);
desired_voice = flite_voice_select(argv[i+1]);
i++;
}
else if ((cst_streq(argv[i],"-voicedir")) && (i+1 < argc))
{
voicedir = argv[i+1];
if (flite_voice_list == NULL)
flite_set_voice_list(voicedir);
i++;
}
else if ((cst_streq(argv[i],"-add_lex")) && (i+1 < argc))
{
lex_addenda_file = argv[i+1];
i++;
}
else if (cst_streq(argv[i],"-f") && (i+1 < argc))
{
filename = argv[i+1];
explicit_filename = TRUE;
i++;
}
else if (cst_streq(argv[i],"-pw"))
{
feat_set_string(extra_feats,"print_info_relation","Word");
feat_set(extra_feats,"post_synth_hook_func",
uttfunc_val(&print_info));
}
else if (cst_streq(argv[i],"-ps"))
{
feat_set_string(extra_feats,"print_info_relation","Segment");
feat_set(extra_feats,"post_synth_hook_func",
uttfunc_val(&print_info));
}
else if (cst_streq(argv[i],"-ssml"))
{
ssml_mode = TRUE;
}
else if (cst_streq(argv[i],"-pr") && (i+1 < argc))
{
feat_set_string(extra_feats,"print_info_relation",argv[i+1]);
feat_set(extra_feats,"post_synth_hook_func",
uttfunc_val(&print_info));
i++;
}
else if (cst_streq(argv[i],"-voicedump") && (i+1 < argc))
{
voicedumpfile = argv[i+1];
i++;
}
else if ((cst_streq(argv[i],"-set") || cst_streq(argv[i],"-s"))
&& (i+1 < argc))
{
ef_set(extra_feats,argv[i+1],0);
i++;
}
else if (cst_streq(argv[i],"--seti") && (i+1 < argc))
{
ef_set(extra_feats,argv[i+1],"int");
i++;
}
else if (cst_streq(argv[i],"--setf") && (i+1 < argc))
{
ef_set(extra_feats,argv[i+1],"float");
i++;
}
else if (cst_streq(argv[i],"--sets") && (i+1 < argc))
{
ef_set(extra_feats,argv[i+1],"string");
i++;
}
else if (cst_streq(argv[i],"-p") && (i+1 < argc))
{
filename = argv[i+1];
explicit_phones = TRUE;
i++;
}
else if (cst_streq(argv[i],"-t") && (i+1 < argc))
{
filename = argv[i+1];
explicit_text = TRUE;
i++;
}
else if (filename)
outtype = argv[i];
else
filename = argv[i];
}
if (filename == NULL) filename = "-"; /* stdin */
if (flite_voice_list == NULL)
flite_set_voice_list(voicedir);
if (desired_voice == 0)
desired_voice = flite_voice_select(NULL);
v = desired_voice;
feat_copy_into(extra_feats,v->features);
durs = 0.0;
if (voicedumpfile != NULL)
{
flite_voice_dump(v,voicedumpfile);
exit(0);
}
if (lex_addenda_file)
flite_voice_add_lex_addenda(v,lex_addenda_file);
if (cst_streq("stream",outtype))
{
asi = new_audio_streaming_info();
asi->asc = audio_stream_chunk;
feat_set(v->features,"streaming_info",audio_streaming_info_val(asi));
}
if (flite_bench)
{
outtype = "none";
filename = "A whole joy was reaping, but they've gone south, you should fetch azure mike.";
explicit_text = TRUE;
}
loop:
gettimeofday(&tv,NULL);
time_start = (double)(tv.tv_sec)+(((double)tv.tv_usec)/1000000.0);
if (explicit_phones)
durs = flite_phones_to_speech(filename,v,outtype);
else if ((strchr(filename,' ') && !explicit_filename) || explicit_text)
{
if (ssml_mode)
durs = flite_ssml_text_to_speech(filename,v,outtype);
else
durs = flite_text_to_speech(filename,v,outtype);
}
else
{
if (ssml_mode)
durs = flite_ssml_file_to_speech(filename,v,outtype);
else
durs = flite_file_to_speech(filename,v,outtype);
}
gettimeofday(&tv,NULL);
time_end = ((double)(tv.tv_sec))+((double)tv.tv_usec/1000000.0);
if (flite_verbose || (flite_bench && bench_iter == ITER_MAX))
printf("times faster than real-time: %f\n(%f seconds of speech synthesized in %f)\n",
durs/(float)(time_end-time_start),
durs,
(float)(time_end-time_start));
if (flite_loop || (flite_bench && bench_iter++ < ITER_MAX))
goto loop;
delete_features(extra_feats);
delete_val(flite_voice_list); flite_voice_list=0;
/* cst_alloc_debug_summary(); */
return 0;
}
