int main(int argc, char **argv)
{
cst_wave *nw, *all;
cst_val *files;
const cst_val *w;
cst_val *wavelist;
cst_features *args;
int i,j;
float ntime;
int stime;
const char *nwfile;
args = new_features();
files =
cst_args(argv,argc,
"usage: combine_waves OPTIONS\n"
"Combine waves into single waveform\n"
"-o <string> Output waveform\n"
"-f <int> Input sample rate (for raw input)\n"
"-itype <string> Input type, raw or headered\n"
"-wavelist <string> File containing times and wave filenames\n",
args);
wavelist = get_wavelist(get_param_string(args,"-wavelist","-"));
if (wavelist == 0)
return -1;
all = new_wave();
for (w = wavelist; w; w = val_cdr(w))
{
ntime = decode_time(val_string(val_car(w)));
nwfile = val_string(val_car(val_cdr(w)));
nw = new_wave();
if (cst_wave_load_riff(nw,nwfile) != CST_OK_FORMAT)
{
fprintf(stderr,
"combine_waves: can't read file or wrong format \"%s\"\n",
nwfile);
continue;
}
stime = ntime * nw->sample_rate;
cst_wave_resize(all,stime+nw->num_samples,1);
for (i=0,j=stime; i<nw->num_samples; i++,j++)
{
/* this will cause overflows */
all->samples[j] += nw->samples[i];
}
delete_wave(nw);
}
cst_wave_save_riff(all,get_param_string(args,"-o","-"));
return 0;
}
cst_val *cst_args(char **argv, int argc,
const char *description,
cst_features *args)
{
/* parses the given arguments wrt the description */
cst_features *op_types = new_features();
cst_val *files = NULL;
int i;
const char *type;
parse_description(description,op_types);
for (i=1; i<argc; i++)
{
if (argv[i][0] == '-')
{
if ((!feat_present(op_types,argv[i])) ||
(cst_streq("-h",argv[i])) ||
(cst_streq("-?",argv[i])) ||
(cst_streq("--help",argv[i])) ||
(cst_streq("-help",argv[i])))
parse_usage(argv[0],"","",description);
else
{
type = feat_string(op_types,argv[i]);
if (cst_streq("<binary>",type))
feat_set_string(args,argv[i],"true");
else
{
if (i+1 == argc)
parse_usage(argv[0],
"missing argument for ",argv[i],
description);
if (cst_streq("<int>",type))
feat_set_int(args,argv[i],atoi(argv[i+1]));
else if (cst_streq("<float>",type))
feat_set_float(args,argv[i],atof(argv[i+1]));
else if (cst_streq("<string>",type))
feat_set_string(args,argv[i],argv[i+1]);
else
parse_usage(argv[0],
"unknown arg type ",type,
description);
i++;
}
}
}
else
files = cons_val(string_val(argv[i]),files);
}
delete_features(op_types);
return val_reverse(files);
}
const cst_val *val_string_x(const char *n)
{
const cst_val *v;
/* *BUG* This will have to be fixed soon */
if (val_string_consts == NULL)
val_string_consts = new_features();
v = feat_val(val_string_consts,n);
if (v)
return v;
else
{
feat_set_string(val_string_consts,n,n);
return feat_val(val_string_consts,n);
}
}
static cst_features *ssml_get_attributes(cst_tokenstream *ts)
{
cst_features *a = new_features();
const char* name, *val;
set_charclasses(ts,
ts->p_whitespacesymbols,
ssml_singlecharsymbols_inattr,
ts->p_prepunctuationsymbols,
ts->p_postpunctuationsymbols);
name = ts_get(ts);
while (!cst_streq(">",name))
{
if (cst_streq(name,"/"))
feat_set_string(a,"_type","startend");
else
{
feat_set_string(a,"_type","start");
feat_set_string(a,"_name0",name);
if (cst_streq("=",ts_get(ts)))
{
val = ts_get_quoted_remainder(ts);
feat_set_string(a,"_val0",val);
}
}
if (ts_eof(ts))
{
fprintf(stderr,"ssml: unexpected EOF\n");
delete_features(a);
return 0;
}
name = ts_get(ts);
}
set_charclasses(ts,
ts->p_whitespacesymbols,
ssml_singlecharsymbols_general,
ts->p_prepunctuationsymbols,
ts->p_postpunctuationsymbols);
return a;
}
int main(int argc, char **argv)
{
cst_wave *in, *out;
cst_val *files;
cst_features *args;
int i,j;
int w, a, t;
args = new_features();
files =
cst_args(argv,argc,
"usage: dcoffset_wave OPTIONS\n"
"Subtract window average from waveform\n"
"-i <string> Input waveform\n"
"-o <string> Output waveform\n"
"-w <int> Window size (in samples)\n",
args);
w = flite_get_param_int(args,"-w",20);
in = new_wave();
cst_wave_load_riff(in,flite_get_param_string(args,"-i","-"));
out = copy_wave(in);
for (i=0; i<=out->num_samples; i++)
{
for (t=a=0,j=i-w/2; j < i+w/2; j++)
{
if ((j > 0) && (j < out->num_samples))
{
t += 1;
a+=in->samples[j];
}
}
/* printf("%d %d %d %d %d\n",i,out->samples[i],a/t,t,out->samples[i]-a/t); */
out->samples[i] -= a/t;
}
cst_wave_save_riff(out,flite_get_param_string(args,"-o","-"));
return 0;
}
string getPhonemes( const char* sText )
{
string sRet;
cst_features* args = new_features();
cst_voice* v;
cst_utterance* u;
cst_item* s;
const char* name;
//const cst_val* d;
flite_init();
v = register_cmu_us_no_wave( NULL );
u = flite_synth_text( sText, v );
for ( s = relation_head( utt_relation( u, "Segment" ) ); s; s = item_next( s ) )
{
sRet += item_feat_string( s, "name" );
float test = item_feat_float( s, "end" );
//d = segment_duration( s );
/* If its a vowel and is stressed output stress value */
if ( ( cst_streq( "+", ffeature_string( s, "ph_vc" ) ) ) &&
( cst_streq( "1", ffeature_string( s, "R:SylStructure.parent.stress" ) ) ) )
{
sRet += "1";
}
sRet += " ";
}
delete_utterance( u );
delete_features( args );
return sRet;
}
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;
}
static float flite_ssml_to_speech_ts(cst_tokenstream *ts,
cst_voice *voice,
const char *outtype)
{
cst_features *ssml_feats, *ssml_word_feats;
cst_features *attributes;
const char *token;
char *tag;
cst_utterance *utt;
cst_relation *tokrel;
int num_tokens;
cst_breakfunc breakfunc = default_utt_break;
cst_uttfunc utt_user_callback = 0;
float durs = 0.0;
cst_item *t;
ssml_feats = new_features();
ssml_word_feats = new_features();
set_charclasses(ts,
" \t\n\r",
ssml_singlecharsymbols_general,
get_param_string(voice->features,"text_prepunctuation",""),
get_param_string(voice->features,"text_postpunctuation","")
);
if (feat_present(voice->features,"utt_break"))
breakfunc = val_breakfunc(feat_val(voice->features,"utt_break"));
if (feat_present(voice->features,"utt_user_callback"))
utt_user_callback = val_uttfunc(feat_val(voice->features,"utt_user_callback"));
num_tokens = 0;
utt = new_utterance();
tokrel = utt_relation_create(utt, "Token");
while (!ts_eof(ts) || num_tokens > 0)
{
token = ts_get(ts);
if (cst_streq("<",token))
{ /* A tag */
tag = cst_upcase(ts_get(ts));
if (cst_streq("/",tag)) /* an end tag */
{
tag = cst_upcase(ts_get(ts));
attributes = ssml_get_attributes(ts);
feat_set_string(attributes,"_type","end");
}
else
attributes = ssml_get_attributes(ts);
utt = ssml_apply_tag(tag,attributes,utt,ssml_word_feats);
cst_free(tag);
}
else if (cst_streq("&",token))
{ /* an escape sequence */
/* skip to ; and insert value in rawdata */
}
else
{
if ((cst_strlen(token) == 0) ||
(num_tokens > 500) || /* need an upper bound */
(relation_head(tokrel) &&
breakfunc(ts,token,tokrel)))
{
/* An end of utt, so synthesize it */
if (utt_user_callback)
utt = (utt_user_callback)(utt);
if (utt)
{
utt = flite_do_synth(utt,voice,utt_synth_tokens);
durs += flite_process_output(utt,outtype,TRUE);
delete_utterance(utt); utt = NULL;
}
else
break;
if (ts_eof(ts)) break;
utt = new_utterance();
tokrel = utt_relation_create(utt, "Token");
num_tokens = 0;
}
num_tokens++;
t = relation_append(tokrel, 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);
/* Mark it at the beginning of the token */
item_set_int(t,"file_pos",
ts->file_pos-(1+ /* as we are already on the next char */
cst_strlen(token)+
cst_strlen(ts->prepunctuation)+
cst_strlen(ts->postpunctuation)));
item_set_int(t,"line_number",ts->line_number);
}
}
delete_utterance(utt);
return durs;
}
/*************************************************************************/
/* */
/* Voice definition */
/* */
/*************************************************************************/
#include "cst_alloc.h"
#include "cst_voice.h"
#include "flite.h"
CST_VAL_REGISTER_TYPE(voice,cst_voice)
cst_voice *new_voice()
{
cst_voice *v = cst_alloc(struct cst_voice_struct,1);
v->features = new_features();
v->ffunctions = new_features();
return v;
}
void delete_voice(cst_voice *v)
{
if (v)
{
if (feat_present(v->features,"voxdata"))
{
if (feat_present(v->features,"clunit_db"))
flite_munmap_clunit_voxdata(v);
}
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)
{
cst_track *t1;
cst_track *me_filters = NULL;
cst_wave *w1, *w2, *res = NULL;
cst_val *files;
cst_features *args;
int i, j;
int order, o, s;
int frame_length;
float *lpcs, *residual;
float m;
const char *f1, *f2;
const char *resfn = NULL;
int last_peak = 0, next_peak;
int period;
float power;
int rfc = 0;
int str = 0;
int fn, fo, ss;
float xpulse, xnoise;
float fxpulse, fxnoise;
float x, me;
float *hpulse = NULL, *hnoise = NULL;
float *xpulsesig = NULL, *xnoisesig = NULL;
int q = 0;
int position;
int lpc_start = 0;
args = new_features();
files =
cst_args(argv, argc,
"usage: lpc_resynth OPTIONS INTRACK OUTWAVE\n"
"Resynth an lpc track\n"
"-res <string> residual (as waveform)\n"
"-save_res Save the generated residual\n"
"-lpc_start <int> start of lpc params in lpc track {1}\n"
"-order <int> LPC order {16}\n"
"-str mixed excitation strengths\n"
"-me_filters <string> mixed excitation filters\n"
"-rfc Coefficents are reflection coefficients\n", args);
f1 = val_string(val_car(files));
f2 = val_string(val_car(val_cdr(files)));
t1 = new_track();
lpc_start = mimic_get_param_int(args, "-lpc_start", 1);
if (feat_present(args, "-rfc"))
rfc = 1;
if (feat_present(args, "-str"))
str = 1;
if (feat_present(args, "-me_filters"))
{
me_filters = new_track();
if (cst_track_load_est
(me_filters,
mimic_get_param_string(args, "-me_filters",
"me_filters.track")) != CST_OK_FORMAT)
{
fprintf(stderr,
"lpc_resynth: can't read file or wrong format \"%s\"\n",
f1);
return -1;
}
hpulse = cst_alloc(float, me_filters->num_channels);
hnoise = cst_alloc(float, me_filters->num_channels);
xpulsesig = cst_alloc(float, me_filters->num_channels);
xnoisesig = cst_alloc(float, me_filters->num_channels);
}
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;
}
int main(int argc, char **argv)
{
bard_reader *br;
cst_features *args;
cst_features *config;
float idle_time;
int spos;
args = new_features();
cst_args(argv,argc,
"usage: bard OPTIONS\n"
"Bard Storyteller\n"
"-text <string> Text filename to read\n"
"-text_pos <int> Starting position in text file\n"
"-config_file <string> Override default config file ($HOME/.bard_config)\n"
"-font <string> Pathname to ttf font file\n"
"-font_size <int> Initial font size\n"
"-audio_method <string> SDL or flite\n"
"-voice <string> Voice name (or pathname to dumped voice)\n"
"-audio_stream_buffer_factor <int> Bigger for slower processors\n"
"-gain <float> Volume factor (1.0 is default)\n"
"-speed <float> Duration stretch (inverse speed)\n"
"-battery_script <string> Script to get battery charge state\n"
"-screen_height <int> \n"
"-screen_width <int> \n"
"-screen_blank_idle_time <float> In secs (0 means no blanking)\n"
"-scroll_delay <int> In ms\n"
"-voices_dir <string> Directory containing *.flitevox voices\n"
"-debug <int> Print debug messages\n"
"",
args);
/* If text specified and no position given, start from 0 */
if (feat_present(args,"-text") && !feat_present(args,"-text_pos"))
feat_set_int(args,"-text_pos",0);
if (!feat_present(args,"-font"))
feat_set_string(args,"-font",BARD_DEFAULT_FONT);
bard_debug = get_param_int(args,"-debug",0);
config = bard_read_config(args);
cst_feat_print(stdout,config);
br = bard_open(config);
if (br == NULL) return -1; /* something went really wrong */
/* Make the windows */
br->text = bard_make_text_window(br);
br->file_select = bard_make_file_select_window(br);
br->info = bard_make_info_window(br); /* general info/params */
br->help = bard_make_help_window(br);
br->recent = bard_make_recent_window(br);
br->menu = bard_make_menu_window(br);
if (feat_present(br->config,"-voices_dir"))
br->voice_select = bard_make_voice_select_window(br);
/* Select focus */
br->display->current = br->text;
bard_display_clear(br->display);
bard_display_update(br->display);
bard_screen_on(NULL); /* just in case it is currently off */
while (br->quit == 0)
{
if (input_process_events(br) == 0)
{
if (br->scroll)
{
SDL_Delay(br->scroll_delay); /* no key press but scrolling */
br->quiet_time = SDL_GetTicks(); /* no blank while scrolling */
}
else
SDL_Delay(br->no_key_delay); /* no key press, so pause a bit */
}
if (br->speak == 1)
{
/* Within the audio callback function, we also call */
/* input_process_events so we can stop if requested */
br->pause = 0;
if (br->display->current->current_token)
spos = br->display->current->current_token->file_pos;
else
spos = br->display->current->sop_pos;
/* We need spos-1 here to make it display well */
bard_speak_text(br->speech,br->display->current->textfile,spos-1);
br->speak = 0;
br->quiet_time = SDL_GetTicks();
}
else if (br->scroll)
{ /* Speaking and smooth scrolling are mutually exclusive */
if (bard_text_scroll(br->text) == 0)
br->scroll = 0; /* Reached end of file so stop scroll */
bard_display_update(br->display); /* Update screen */
}
else
br->text->scroll_offset = 0;
/* Check if we should black/hide cursor */
idle_time = (SDL_GetTicks()-br->quiet_time)/1000.0;
if ((br->screen_blank_idle_time > 0) &&
(idle_time > br->screen_blank_idle_time) &&
(br->blank == 0))
{
br->blank = 1;
bard_screen_off(NULL);
}
idle_time = (SDL_GetTicks()-br->mouse_quiet_time)/1000.0;
if (idle_time > 5.0)
SDL_ShowCursor(SDL_DISABLE);
}
/* Save context */
bard_write_config(br);
bard_close(br);
delete_features(args);
return 0;
}
