[1.30.01] espeakChar() & say-as tts:char, speak the unicode number for unknown characters.

lang=ta: change stress rule to first syllable.


git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@131 d46cf337-b52f-0410-862d-fd96e6ae7743

src/Makefile_pulseaudio

 LIBDIR=/usr/lib
 DATADIR=/usr/share/espeak-data
 
-RELEASE = 1.30
+RELEASE = 1.31
 BIN2_NAME = espeak
 LIB_NAME = libespeak.so
 STATIC_LIB_NAME = libespeak.a
 LN_SF = /bin/ln -sf
 MKDIR = mkdir -p
 
-#AUDIO = portaudio
-AUDIO = pulseaudio
+#BIN_NAME = speak
 
-BIN_NAME = speak
-
-ifeq ($(AUDIO),pulseaudio)
-WAVE=wave_pulse.cpp
-LIB_AUDIO=pulse
-USE_AUDIO=USE_PULSEAUDIO
-# the speak executable is not yet compliant with PulseAudio, 
-# speak_SOURCES defined for avoiding compilation errors.
-speak_SOURCES = speak.cpp compiledict.cpp dictionary.cpp intonation.cpp \
-	readclause.cpp setlengths.cpp numbers.cpp synth_mbrola.cpp \
-	synthdata.cpp synthesize.cpp translate.cpp tr_english.cpp \
-	tr_languages.cpp voices.cpp wavegen.cpp phonemelist.cpp \
-	espeak_command.cpp event.cpp fifo.cpp $(WAVE) debug.cpp
-else
-#portaudio
-WAVE=wave.cpp
-LIB_AUDIO=portaudio
-USE_AUDIO=USE_PORTAUDIO
-speak_SOURCES = speak.cpp compiledict.cpp dictionary.cpp intonation.cpp \
-	readclause.cpp setlengths.cpp numbers.cpp synth_mbrola.cpp \
-	synthdata.cpp synthesize.cpp translate.cpp tr_english.cpp \
-	tr_languages.cpp voices.cpp wavegen.cpp phonemelist.cpp
-
-endif
-
-speak_SOURCES = speak.cpp compiledict.cpp dictionary.cpp intonation.cpp \
-	readclause.cpp setlengths.cpp numbers.cpp synth_mbrola.cpp \
-	synthdata.cpp synthesize.cpp translate.cpp tr_english.cpp \
-	tr_languages.cpp voices.cpp wavegen.cpp phonemelist.cpp
+# speak_SOURCES = speak.cpp compiledict.cpp dictionary.cpp intonation.cpp \
+# 	readclause.cpp setlengths.cpp numbers.cpp synth_mbrola.cpp \
+# 	synthdata.cpp synthesize.cpp translate.cpp tr_english.cpp \
+# 	tr_languages.cpp voices.cpp wavegen.cpp phonemelist.cpp
 
 libespeak_SOURCES = speak_lib.cpp compiledict.cpp dictionary.cpp intonation.cpp \
 	readclause.cpp setlengths.cpp numbers.cpp synth_mbrola.cpp \
 	synthdata.cpp synthesize.cpp translate.cpp tr_english.cpp \
 	tr_languages.cpp voices.cpp wavegen.cpp phonemelist.cpp \
-	espeak_command.cpp event.cpp fifo.cpp $(WAVE) debug.cpp
+	espeak_command.cpp event.cpp fifo.cpp wave_pulse.cpp debug.cpp
 
 
 SRCS1=$(speak_SOURCES)
 OBJS1=$(patsubst %.cpp,%.o,$(SRCS1))
-LIBS1=-lstdc++ -l$(LIB_AUDIO) -lpthread
+LIBS1=-lstdc++ -lpulse -lpthread
 
 SRCS2=$(libespeak_SOURCES)
 OBJS2=$(patsubst %.cpp,x_%.o,$(SRCS2))
-LIBS2=-lstdc++ -l$(LIB_AUDIO) -lpthread
+LIBS2=-lstdc++ -lpulse -lpthread
 
 SRCS3 = espeak.cpp
 OBJS3=$(patsubst %.cpp,%.o,$(SRCS3))
 LIBS3=-lstdc++ ./libespeak.so
 
-#CXXFLAGS=-O2 -D$(USE_AUDIO)
-CXXFLAGS=-D$(USE_AUDIO) -ggdb -DDEBUG_ENABLED
+CXXFLAGS=-O2 -DUSE_PULSEAUDIO
+#CXXFLAGS=-DUSE_PULSEAUDIO -ggdb -DDEBUG_ENABLED
 
 
-all: $(BIN_NAME) $(LIB_NAME) $(STATIC_LIB_NAME) $(BIN2_NAME)
+#all: $(BIN_NAME) $(LIB_NAME) $(STATIC_LIB_NAME) $(BIN2_NAME)
+all: $(LIB_NAME) $(STATIC_LIB_NAME) $(BIN2_NAME)
 	mv $(LIB_NAME) $(LIB_NAME).$(LIBTAG)
 
 .cpp.o:
 	$(CXX) $(CXXFLAGS) -D PATH_ESPEAK_DATA=\"$(DATADIR)\" -Wall -pedantic -I. -c -fno-exceptions $<
 
-$(BIN_NAME): $(OBJS1)
-	$(CXX) -o $@ $(OBJS1) $(LIBS1)
+# $(BIN_NAME): $(OBJS1)
+# 	$(CXX) -o $@ $(OBJS1) $(LIBS1)
 
 $(BIN2_NAME): $(OBJS3) $(LIB_NAME)
 	$(CXX) -o $@ $(OBJS3) $(LIBS3)
 	rm -f *.o *.a *~
 
 distclean: clean
-	rm -f $(BIN_NAME)
+#	rm -f $(BIN_NAME)
 	rm -f $(BIN2_NAME)
 	rm -f $(LIB_NAME)*
 

src/dictionary.cpp

 							continue;  // start again at the beginning of the word
 						}
 						else
-						if((letter >= 0x3200) && (letter < 0xa700))
+						if((letter >= 0x3200) && (letter < 0xa700) && (end_phonemes != NULL))
 						{
 							// ideograms
 							// outside the range of the accent table, speak the unknown symbol sound

src/numbers.cpp

 
 
 
+void Translator::LookupLetter(int letter, int next_byte, char *ph_buf1)
+{//==================================================================
+	int len;
+	unsigned char *p;
+	static char single_letter[10] = {0,0};
+	char ph_stress[2];
+	char ph_buf3[30];
+
+	len = utf8_out(letter,&single_letter[2]);
+	single_letter[len+2] = ' ';
+
+	if(next_byte != ' ')
+		next_byte = RULE_SPELLING;
+	single_letter[3+len] = next_byte;   // follow by space-space if the end of the word, or space-0x31
+
+	single_letter[1] = '_';
+	if(Lookup(&single_letter[1],ph_buf3) == 0)
+	{
+		single_letter[1] = ' ';
+		if(Lookup(&single_letter[2],ph_buf3) == 0)
+		{
+			TranslateRules(&single_letter[2], ph_buf3, sizeof(ph_buf3), NULL,0,0);
+		}
+	}
+
+	if(ph_buf3[0] == 0)
+	{
+		ph_buf1[0] = 0;
+		return;
+	}
+
+	// at a stress marker at the start of the letter name, unless one is already marked
+	ph_stress[0] = phonSTRESS_P;
+	ph_stress[1] = 0;
+
+	for(p=(unsigned char *)ph_buf3; *p != 0; p++)
+	{
+		if(phoneme_tab[*p]->type == phSTRESS)
+			ph_stress[0] = 0;  // stress is already marked
+	}
+	sprintf(ph_buf1,"%s%s",ph_stress,ph_buf3);
+}
+
+
+
 int Translator::TranslateLetter(char *word, char *phonemes, int control)
 {//=====================================================================
 // get pronunciation for an isolated letter
 	int n_bytes;
 	int letter;
 	int len;
-	int next;
-	unsigned char *p;
-	char ph_stress[2];
+	char *p2;
+	char *pbuf;
 	char capital[20];
-	char ph_buf[30];
+	char ph_buf[50];
 	char ph_buf2[50];
+	char hexbuf[6];
 	static char single_letter[10] = {0,0};
 
 	ph_buf[0] = 0;
 		return(n_bytes);
 	}
 
-	len = utf8_out(letter,&single_letter[2]);
-	single_letter[len+2] = ' ';
-
-	next = RULE_SPELLING;
-	if(word[n_bytes] == ' ')
-		next = ' ';
-	single_letter[3+len] = next;   // follow by space-space if the end of the word, or space-0x31
-
-	single_letter[1] = '_';
-	if(Lookup(&single_letter[1],ph_buf) == 0)
-	{
-		single_letter[1] = ' ';
-		if(Lookup(&single_letter[2],ph_buf) == 0)
-		{
-			TranslateRules(&single_letter[2], ph_buf, sizeof(ph_buf), NULL,0,0);
-		}
-	}
+	LookupLetter(letter, word[n_bytes], ph_buf);
 
 	if(ph_buf[0] == phonSWITCH)
 	{
 
 		if((ph_buf[0]==0) && !iswspace(letter))
 			Lookup("_??",ph_buf);
-	}
-
-	// at a stress marker at the start of the letter name, unless one is already marked
-	ph_stress[0] = phonSTRESS_P;
-	ph_stress[1] = 0;
 
-	for(p=(unsigned char *)ph_buf; *p != 0; p++)
-	{
-		if(phoneme_tab[*p]->type == phSTRESS)
-			ph_stress[0] = 0;  // stress is already marked
+		if((control==4) && (ph_buf[0] != 0))
+		{
+			// speak the hexadecimal number of the character code
+			sprintf(hexbuf,"%x",letter);
+			pbuf = ph_buf;
+			for(p2 = hexbuf; *p2 != 0; p2++)
+			{
+				pbuf += strlen(pbuf);
+				LookupLetter(*p2, 0, pbuf);
+			}
+		}
 	}
 
 	len = strlen(phonemes);
-	sprintf(ph_buf2,"%c%s%s%s",0xff,capital,ph_stress,ph_buf);  // the 0xff marker will be removed or replaced in SetSpellingStress()
+	sprintf(ph_buf2,"%c%s%s",0xff,capital,ph_buf);  // the 0xff marker will be removed or replaced in SetSpellingStress()
 	if((len + strlen(ph_buf2)) < N_WORD_PHONEMES)
 	{
 		strcpy(&phonemes[len],ph_buf2);
 			}
 			else
 			{
-				if((count != n_stress) && (control < 4))
+				if(count != n_stress)
 				{
 					if(((count % 3) != 0) || (count == n_stress-1))
 						c = phonSTRESS_3;   // reduce to secondary stress
 			switch(langopts.numbers & 0x6000)
 			{
 			case 0x4000:
-				if(decimal_count < 4)
+				// French/Polish decimal fraction
+				while(word[n_digits] == '0')
 				{
-					// Polish decimal fraction
-					if ((word[n_digits] != '0'))
-					{
-						LookupNum3(atoi(&word[n_digits]),buf1,0,0,0);
-						strcat(ph_out,buf1);
-						n_digits += decimal_count;
-					}
-					else
-					if ((word[n_digits+1] != '0'))
-					{
-						Lookup("_0",buf1);
-						strcat(ph_out,buf1);
-						LookupNum3(atoi(&word[n_digits]),buf1,0,0,0);
-						strcat(ph_out,buf1);
-						n_digits += decimal_count;
-					}
+					Lookup("_0",buf1);
+					strcat(ph_out,buf1);
+					decimal_count--;
+					n_digits++;
+				}
+				if(decimal_count < 6)
+				{
+					LookupNum3(atoi(&word[n_digits]),buf1,0,0,0);
+					strcat(ph_out,buf1);
+					n_digits += decimal_count;
 				}
 				break;
 

src/synthdata.cpp

 #include "translate.h"
 #include "wave.h"
 
-const char *version_string = "1.31  19.Jan.08";
+const char *version_string = "1.31.01  23.Jan.08";
 const int version_phdata  = 0x013100;
 
 int option_device_number = -1;

src/tr_languages.cpp

 			tr->langopts.stress_flags = 0x0024;  // don't use secondary stress
 			tr->langopts.param[LOPT_IT_LENGTHEN] = 1;    // remove lengthen indicator from unstressed syllables
 
-			tr->langopts.numbers = 0x1509 +  NUM_NOPAUSE | NUM_ROMAN;
+			tr->langopts.numbers = 0x1509 + 0x4000 + NUM_NOPAUSE | NUM_ROMAN;
 			SetLetterVowel(tr,'y');
 		}
 		break;
 
 	case L('t','a'):
 		{
-			static const short stress_lengths_ta[8] = {190, 190,  210, 210,  0, 0,  220, 220};
-			static const unsigned char stress_amps_ta[8] = {17,14, 20,19, 20,24, 24,22 };
+			static const short stress_lengths_ta[8] = {190, 190,  200, 200,  0, 0,  220, 220};
+			static const unsigned char stress_amps_ta[8] = {18,18, 18,18, 20,20, 22,22 };
 
 			tr = new Translator();
 			SetupTranslator(tr,stress_lengths_ta,stress_amps_ta);
 			tr->langopts.length_mods0 = tr->langopts.length_mods;  // don't lengthen vowels in the last syllable
 
-			tr->langopts.stress_rule = 6;      // stress on last heaviest syllable, excluding final syllable
+			tr->langopts.stress_rule = 1;
 			tr->langopts.stress_flags =  0x10004;   // use 'diminished' for unstressed final syllable
-//			tr->langopts.numbers = 0x811;
-//			tr->langopts.numbers2 = 0x100;
 			tr->letter_bits_offset = OFFSET_TAMIL;
 
 			memset(tr->letter_bits,0,sizeof(tr->letter_bits));

src/translate.h

 	const char *LookupDict2(const char *word, const char *word2, char *phonetic, unsigned int *flags, int end_flags, WORD_TAB *wtab);
 	const char *LookupSpecial(const char *string);
 	const char *LookupCharName(int c);
+	void LookupLetter(int letter, int next_byte, char *ph_buf);
 	int LookupNum2(int value, int control, char *ph_out);
 	int LookupNum3(int value, char *ph_out, int suppress_null, int thousandplex, int prev_thousands);
 	int LookupThousands(int value, int thousandplex, char *ph_out);