[1.33] continued

git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@159 d46cf337-b52f-0410-862d-fd96e6ae7743

dictsource/ku_list

+
+// This file is UTF8 encoded 
+// Spelling-to-phoneme words for Kurdish
+
+
+_0	s'Ifr
+_1	j'Ek
+_2	d'U
+_3	s'e
+_4	dZ'ar
+_5	p'endZ
+_6	S'ES
+_7	h'Eft
+_8	h'EjSt
+_9	n'Eh
+_10	d'Eh
+_11	jazd'E#
+_12	danzd'E#
+_13	sezd'E#
+_14	tSaRd'E#
+_15	panzd'E#
+_16	Sanzd'E#
+_17	hIvd'Eh
+_18	hIZd'Eh
+_19	nozd'Eh
+_2X	b'Ist
+_3X	s'i
+_4X	tS'Il
+_5X	pendZ'i
+_6X	S'est
+_7X	hEft'e
+_8X	heSt'e
+_9X	n'od
+_0C	s'Ed
+_1C	s'Ed
+_0M1	hEz'ar
+_1M1	hEz'ar
+_0M2	mIl'jon
+_1M2	mIl'jon
+
+_0and	_u
+_dpt	_xal
+_??	tipa
+_cap	mezin   // or "ya mezin" after the letter
+
+
+// characters
+// letter names
+_a	a
+b	be
+c	dZe
+ç	tSe
+d	de
+_e	E
+_ê	e
+f	fe
+g	ge
+h	he
+_i	I
+_î	i
+j	Ze
+k	ke
+l	le
+m	me
+n	ne
+_o	o
+p	pe
+q	qe
+r	Re
+s	se
+ş	Se
+t	te
+_u	U
+_û	u
+v	ve
+w	we
+x	xe
+y	je
+z	ze
+
+
+_,	behnok
+_.	xal
+_:	dZotxal
+_;	xalbehnok
+_-	bEndIk
+_?	pIRsniS'an
+_!	banES'an
+%	sEde
+*	steRk
+
+// function words
+
+û	$u+ $pause
+a	$u+
+ê	$u+	
+
+// pronouns
+ez	$u+
+te	$u+
+wê	$u+
+wî	$u+
+me	$u+
+we	$u+
+wan	$u+
+
+bes	$pause
+an	$u $pause
+anji	$u $pause
+ên	$u
+ya	$u
+yê	$u
+yan	$u
+yên	$u
+ji	$u
+bi	$u
+re	$u
+di	$u
+de	$u
+kir	$u
+got	$u
+
+
+// abbreviations
+hwd	hEr||wEki||dIn
+
+// pronunciation exceptions

dictsource/ku_rules

+
+// Kurdish translation rules
+// This file is UTF-8 encoded
+
+.L01 î iy ê ew ey   // front vowels
+
+.group a
+        a        a
+
+     @) a (_S1   %a
+     @) an (_S2  %an
+
+.group b
+        b        b
+
+.group c
+        c        dZ
+
+.group ç
+        ç        tS
+
+.group d
+        d        d
+
+.group e
+        e        E
+        e (_     E#
+        ew       eU
+        ey       eI
+
+.group ê
+        ê        e
+
+     @) ê (_S1   %e
+     @) ên (_S2  %en
+
+
+.group f
+        f        f
+
+.group g
+        g        g
+        g (L01   J
+
+.group h
+        h        h
+
+.group i
+        i        I
+
+.group î
+        î        i
+
+.group j
+        j        Z
+
+.group k
+        k        k
+        k (L01   c
+
+.group l
+        l        l
+
+.group m
+        m        m
+
+.group n
+        n        n
+
+.group o
+        o        o
+
+.group p
+        p        p
+
+.group q
+        q        q
+
+.group r
+        r        R
+
+.group s
+        s        s
+
+.group ş
+        ş        S
+
+.group t
+        t        t
+
+.group u
+        u        U
+
+.group û
+        û        u
+
+.group v
+        v        v
+
+.group w
+        w        w
+
+.group x
+        x        x
+
+.group y
+        y        j
+
+.group z
+        z        z
+
+.group
+        ö        I:
+        ü        y
+        $        dUlar

phsource/compile_report

-49 phoneme tables
+47 phoneme tables
           new total
     base  98   98
    base2  24  117
      grc  12  122
       sv  25  120
       no  28  124
-      is  32  123
+      is  33  124
       vi  42  136
      zhy  40  131
       zh  65  148
       tr  18  122
       ku  17  121
       id  14  119
-      ja   7  103
-      rw  12  122
 
 Data file      Used by
 b/b             [b] base
                 [s] es
                 [s#] pt
                 [z;] ro
+                [s2] is
 ufric/s!        [s] base
                 [s2] fi
                 [s] fr
                 [s2] hu
                 [s] es
                 [s#] pt
+                [s2] is
 ufric/s_continue  [s2] fi
                 [s2] hu
+                [s2] is
 ufric/sh        [S] base
                 [S] fr
                 [s#] pt_pt
                 [p2] fr
                 [ph] hi
                 [ph] zh
-ustop/percus10  [(X1] base
+ustop/percus10  [(X1)] base
 ustop/pl        [p] base
                 [p] base2
                 [p] fi
 ustop/t_short   [t.] hi
                 [t] sk
                 [t] el
-ustop/tsh_pzd   [tS;] zh
+ustop/tsh_pzd   [tS;h] zh
 ustop/tsh_pzd_unasp  [tS;] zh
 ustop/tsh_sr    [tS] sr
 ustop/ts_pzd    [tS;] base
                 [c] hu
 ustop/t_sr      [d] sr
                 [t] sr
-ustop/ts_rfx    [ts.] zh
+ustop/ts_rfx    [ts.h] zh
 ustop/ts_rfx_unasp  [ts.] zh
 ustop/ts_sr     [tS;] sr
                 [dZ;] sr
 vnasal/a#u_n    [&U~] pt
                 [&U~] pt_pt
 vnasal/ee_n     [E~] en
-                [a] en_us
                 [E~] fr_ca
-vnasal/ee_n2    [E~] hi
+vnasal/ee_n2    [a] en_us
+                [E~] hi
                 [E~] pl
                 [&~] pt_pt
 vnasal/e_n      [e~] af
                 [a] it
                 [a] zh
                 [a] id
-                [a] ja
 vowel/a#        [a/] base2
                 [a2] en_sc
                 [a2] en_wi
                 [&] sr
                 [&] pt
                 [&/] pt
-                [a#] rw
-                [a/] rw
 vowel/a_2       [a] base2
                 [a] en_wi
                 [A:] en_wi
                 [a] pt_pt
                 [a] ro
                 [a:] vi
-                [a] rw
 vowel/a#_2      [&] hr
                 [a2] sv
                 [&] is
                 [e] grc
                 [e] vi
                 [e] id
-                [e] rw
-                [e#] rw
 vowel/e#        [I] en_sc
                 [I2] en_sc
 vowel/e_2       [eI] en_sc
                 [E] tr
                 [E] id
                 [E2] id
-                [e] ja
 vowel/e_mid2    [E] af
                 [E] de
                 [E2] de
                 [i] zhy
                 [i] zh
                 [i] sw
-                [i] ja
-                [i] rw
-                [i#] rw
 vowel/i#        [i] en_rp
                 [y:] cy
 vowel/i_2       [i:] en_rp
                 [o] grc
                 [o:] sv
                 [o] id
-                [o] rw
-                [o#] rw
 vowel/o_2       [o:] cy
                 [o:] hi
                 [o:] hu
                 [o] zhy
                 [O] tr
                 [O] id
-                [o] ja
 vowel/oo_1      [O:] en_n
                 [O:] en_rp
                 [O@] en_rp
                 [u] is
                 [u] zhy
                 [u] zh
-                [u] rw
-                [u#] rw
 vowel/u_bck2    [u:] la
 vowel/u_fnt     [u:] en_rp
 vowel/uu        [U] en
                 [U] pt
                 [U] no
                 [u] zhy
-vowel/u_unr     [u] ja
 vowel/V         [3] en_sc
 vowel/V_2       [V] en
                 [a] af
 vwl_ru/u#       [u#] ru
 vwl_ru/u#u      [u"] ru
 vwl_sv/r_sv3    [r] sv
-vwl_zh/aang     [aan] zhy
+vwl_zh/aang     [aang] zhy
 vwl_zh/aau      [Au] zh
 vwl_zh/a_n      [a] zh
 vwl_zh/ang      [ang] zhy
 vwl_zh/ing      [ing] zhy
 vwl_zh/iou      [iou] zh
 vwl_zh/ng       [ng] zhy
-vwl_zh/oeng     [oen] zhy
+vwl_zh/oeng     [oeng] zhy
 vwl_zh/ong      [ong] zhy
 vwl_zh/uaa      [uA] zh
 vwl_zh/uai      [uai] zh

phsource/ph_english_us

   vowel  starttype (a)  endtype (@)
   length 210
   formants vwl_en_us/a
-  before n vnasal/ee_n
-  before N vnasal/ee_n
-  before m vnasal/ee_n
+  before n vnasal/ee_n2
+  before N vnasal/ee_n2
+  before m vnasal/ee_n2
   reduceto a2  0
 endphoneme
 

phsource/ph_icelandic

 endphoneme
 
 
+phoneme s2       // 2nd part of geminated s
+  vls alv frc sibilant
+  vowelin  f1=0  f2=1700 -300 300  f3=-100 80
+  vowelout f1=0  f2=1700 -300 250  f3=-100 80  rms=20
+  lengthmod 3
+  wave ufric/s_continue
+  before _ ufric/s_
+  before p ufric/s!
+  before t ufric/s!
+  before k ufric/s!
+  switchvoicing z
+endphoneme
+

phsource/ph_tamil

 
 phoneme u
   vowel  starttype (u)  endtype (u)
-  length 130
+  length 110
   formants vowel/u
   reduceto U 4
 endphoneme
 
 phoneme U
   vowel  starttype (u)  endtype (u)
-  length 130
+  length 110
   formants vowel/u#_3
 endphoneme
 
 phoneme u:
   vowel  starttype (u)  endtype (u)
-  length 270
+  length 250
   formants vowel/u
 endphoneme
 

phsource/phonemes

 phonemetable id base2
 include ph_indonesian
 
-phonemetable ja base
-include ph_japanese
-
-phonemetable rw base2
-include ph_kinyarwanda
-

platforms/windows/make_espeak.iss

 
 [Setup]
 AppName=eSpeak
-AppVerName=eSpeak version 1.32
+AppVerName=eSpeak version 1.33
 AppCopyright=Licensed under GNU General Public License version 3.   (See file License.txt for details).
 WindowVisible=yes
 

platforms/windows/make_espeakedit.iss

 
 [Setup]
 AppName=eSpeakEdit
-AppVerName=eSpeakEdit version 1.32
+AppVerName=eSpeakEdit version 1.33
 DefaultDirName={pf}\eSpeak
 DefaultGroupName=eSpeak
 OutputBaseFilename=setup_espeakedit

src/Makefile

-# removed -fno-exceptions  because it caused problems with stdio library on some platforms
+# remove -fno-exceptions  because it caused problems with stdio library on some platforms (ARM)
 
 PREFIX=/
 BINDIR=$(PREFIX)/usr/bin
 LIBDIR=$(PREFIX)/usr/lib
 DATADIR=$(PREFIX)/usr/share/espeak-data
 
-RELEASE = 1.31
+RELEASE = 1.33
 BIN_NAME = speak
 BIN2_NAME = espeak
 LIB_NAME = libespeak.so
 	mv $(LIB_NAME) $(LIB_NAME).$(LIBTAG)
 
 .cpp.o:
-	$(CXX) $(CXXFLAGS) -D PATH_ESPEAK_DATA=\"$(DATADIR)\" -Wall -pedantic -I. -c  $<
+	$(CXX) $(CXXFLAGS) -D PATH_ESPEAK_DATA=\"$(DATADIR)\" -Wall -pedantic -I. -c -fno-exceptions  $<
 
 $(BIN_NAME): $(OBJS1)
 	$(CXX) -o $@ $(OBJS1) $(LIBS1)
 
 x_%.o: %.cpp
 	$(CXX) $(CXXFLAGS) -Wall -fpic -pedantic \
-	-I. -D LIBRARY -c  $<  -o x_$*.o
+	-I. -D LIBRARY -c -fno-exceptions  $<  -o x_$*.o
 
 $(LIB_NAME): $(OBJS2)
 	$(CXX) -shared $(SONAME_OPT)$(LIB_NAME).$(LIB_VERSION) -o $@ $(OBJS2) $(LIBS2)

src/compiledata.cpp

 #include "voice.h"
 #include "spect.h"
 
+#ifdef PLATFORM_POSIX
+#include <unistd.h>
+#endif
+
+
 
 #define tNUMBER   1
 #define tSTRING   2
 	int resample_wav = 0;
 	char fname_temp[100];
 	int scale_factor=0;
+	int fd_temp;
 	char command[sizeof(path_source)+200];
 
 	fseek(f,24,SEEK_SET);
 			Error("Not mono: ",fname);
 		}
 #else
-		sprintf(fname_temp,"%s.wav",tmpnam(NULL));
-		sprintf(command,"sox \"%s%s.wav\" -r %d -c 1 -w  %s polyphase\n",path_source,fname,samplerate,fname_temp);
-		if(system(command) < 0)
+		strcpy(fname_temp,"/tmp/espeakXXXXXX");
+		if((fd_temp = mkstemp(fname_temp)) >= 0)
 		{
-			Error("Failed to resample: ",command);
-			return(0);
+			close(fd_temp);
+			sprintf(command,"sox \"%s%s.wav\" -r %d -c 1 -w  %s polyphase\n",path_source,fname,samplerate,fname_temp);
+			if(system(command) < 0)
+			{
+				Error("Failed to resample: ",command);
+				return(0);
+			}
 		}
 
 		f = fopen(fname_temp,"rb");

src/compiledict.cpp

 	}
 
 	value = N_HASH_DICT;
-	fwrite(&value,4,1,f_out);
-	fwrite(&offset_rules,4,1,f_out);
+	Write4Bytes(f_out,value);
+	Write4Bytes(f_out,offset_rules);
 
 	compile_dictlist_start();
 

src/dictionary.cpp

 	to[size-1] = 0;
 }
 
+
+// use the following macros for translation from little endian to host's endianness
+
 #ifdef ARCH_BIG
-int reverse_word_bytes(int word)
-{//=============================
-	// reverse the order of bytes from little-endian to big-endian
-	int ix;
-	int word2 = 0;
+#define FIX_ENDIAN_USHORT(val)    ((unsigned short) ( \
+     (unsigned short) ((unsigned short) (val) >> 8) |  \
+     (unsigned short) ((unsigned short) (val) << 8)))
 
-	for(ix=24; ix>=0; ix -= 8)
-	{
-		word2 |= (word >> ix) & 0xff;
-		word2 = word2 << 8;
-	}
-	return(word2);
-}
+#define FIX_ENDIAN_UINT(val)    ((unsigned int) ( \
+     (((unsigned int) (val) & (unsigned int) 0x000000ffU) << 24) | \
+     (((unsigned int) (val) & (unsigned int) 0x0000ff00U) <<  8) | \
+     (((unsigned int) (val) & (unsigned int) 0x00ff0000U) >>  8) | \
+     (((unsigned int) (val) & (unsigned int) 0xff000000U) >> 24)))
+
+#else
+
+#  define FIX_ENDIAN_USHORT(val) (val)
+#  define FIX_ENDIAN_UINT(val) (val)
 #endif
 
 
 	size = GetFileLength(fname);
 
 	f = fopen(fname,"rb");
-	if((f == NULL) || (size == 0))
+	if((f == NULL) || (size <= 0))
 	{
 		if(no_error == 0)
 		{
 
 
 	pw = (int *)data_dictlist;
-	length = pw[1];
+	length = FIX_ENDIAN_UINT(pw[1]);
 	if(size <= (N_HASH_DICT + sizeof(int)*2))
 	{
 		fprintf(stderr,"Empty _dict file: '%s\n",fname);
 		return(2);
 	}
-	if((pw[0] != N_HASH_DICT) || (length <= 0) || (length > 0x8000000))
+
+	if((FIX_ENDIAN_UINT(pw[0]) != N_HASH_DICT) ||
+	   (length <= 0) || (length > 0x8000000))
 	{
 		fprintf(stderr,"Bad data: '%s'\n",fname);
 		return(2);
 			pw = (unsigned int *)langopts.replace_chars;
 			while(*pw != 0)
 			{
-				*pw = reverse_word_bytes(*pw);
+				*pw = FIX_ENDIAN_UINT(*pw);
+				pw++;
+				*pw = FIX_ENDIAN_UINT(*pw);
 				pw++;
 			}
 #endif

src/numbers.cpp

 
 
 // characters U+00e0 to U+017f
-const short letter_accents_0e0[] = {
+const unsigned short letter_accents_0e0[] = {
 LETTER('a',M_GRAVE,0),   // U+00e0
 LETTER('a',M_ACUTE,0),
 LETTER('a',M_CIRCUMFLEX,0),
 
 
 // characters U+0250 to U+029F
-const short letter_accents_250[] = {
+const unsigned short letter_accents_250[] = {
 LETTER('a',M_TURNED,0),		// U+250
 LETTER(L_ALPHA,0,0),
 LETTER(L_ALPHA,M_TURNED,0),

src/readclause.cpp

 #include "voice.h"
 #include "translate.h"
 
+#ifdef PLATFORM_POSIX
+#include <unistd.h>
+#endif
+
 #include <locale.h>
 #define N_XML_BUF   256
 
 	if((f = fopen(fname,"rb")) != NULL)
 	{
 		int ix;
+		int fd_temp;
 		const char *resample;
 		int header[3];
 
 			else
 				resample = "polyphase";
 
-			sprintf(fname_temp,"%s.wav",tmpnam(NULL));
-			sprintf(command,"sox \"%s\" -r %d -w -s -c1 %s %s\n", fname, samplerate, fname_temp, resample);
-			if(system(command) == 0)
+			strcpy(fname_temp,"/tmp/espeakXXXXXX");
+			if((fd_temp = mkstemp(fname_temp)) >= 0)
 			{
-				fname = fname_temp;
+				close(fd_temp);
+//			sprintf(fname_temp,"%s.wav",tmpnam(NULL));
+				sprintf(command,"sox \"%s\" -r %d -w -s -c1 %s %s\n", fname, samplerate, fname_temp, resample);
+				if(system(command) == 0)
+				{
+					fname = fname_temp;
+				}
 			}
 		}
 	}

src/synthdata.cpp

 #include "translate.h"
 #include "wave.h"
 
-const char *version_string = "1.32  04.Mar.08";
+const char *version_string = "1.33  06.Mar.08";
 const int version_phdata  = 0x013200;
 
 int option_device_number = -1;

src/synthesize.cpp

 	int  ix;
 	static char buf[5];
 
-	for(ix=0; ix<3; ix++)
+	for(ix=0; ix<4; ix++)
 		buf[ix] = word >> (ix*8);
 	buf[4] = 0;
 	return(buf);

src/tr_languages.cpp

 
 	case L('t','a'):  // Tamil
 		{
-			static const short stress_lengths_ta[8] = {190, 190,  200, 200,  0, 0,  220, 220};
+			static const short stress_lengths_ta[8] = {200, 200,  210, 210,  0, 0,  230, 230};
 			static const unsigned char stress_amps_ta[8] = {18,18, 18,18, 20,20, 22,22 };
 
 			tr = new Translator();

src/translate.cpp

 				}
 
 				wordx[-1] = c_temp;
-				pfix = 1;
-				prefix_chars[0] = 0;
-				n_chars = prefix_type & 0x3f;
 
-				for(ix=0; ix < n_chars; ix++)    // num. of bytes to remove
+				if((prefix_type & SUFX_B) == 0)
 				{
-					prefix_chars[pfix++] = *wordx++;
+					for(ix=(prefix_type & 0xf); ix>0; ix--)    // num. of characters to remove
+					{
+						wordx++;
+						while((*wordx & 0xc0) == 0x80) wordx++;  // for multibyte characters
+					}
+				}
+				else
+				{
+					pfix = 1;
+					prefix_chars[0] = 0;
+					n_chars = prefix_type & 0x3f;
 
-					if((prefix_type & SUFX_B) && (ix == (n_chars-1)))
+					for(ix=0; ix < n_chars; ix++)    // num. of bytes to remove
 					{
-						prefix_chars[pfix-1] = 0;  // discard the last character of the prefix, this is the separator character
+						prefix_chars[pfix++] = *wordx++;
+	
+						if((prefix_type & SUFX_B) && (ix == (n_chars-1)))
+						{
+							prefix_chars[pfix-1] = 0;  // discard the last character of the prefix, this is the separator character
+						}
 					}
+					prefix_chars[pfix] = 0;
 				}
-				prefix_chars[pfix] = 0;
 				c_temp = wordx[-1];
 				wordx[-1] = ' ';
 				confirm_prefix = 1;
 
 				if(prefix_type & SUFX_B)
 				{
+// SUFX_B is used for Turkish, tr_rules contains "(Pb£
 					// retranslate the prefix part
 					char *wordpf;
 					char prefix_phonemes2[12];

src/transldlg.cpp

 #include "translate.h"
 #include "prosodydisplay.h"
 
+#ifdef PLATFORM_POSIX
+#include <unistd.h>
+#endif
+
+
 #include "wx/font.h"
 #include "wx/textctrl.h"
 #include "wx/button.h"
 	int clause_tone;
 	int clause_count;
 	FILE *f;
+	int  fd_temp;
 	char fname_temp[100];
 	static int n_ph_list;
 	static PHONEME_LIST ph_list[N_PH_LIST+1];
 	case T_RULES:
 	case MENU_SPEAK_RULES:
 		option_phonemes = 2;
+
+#ifdef PLATFORM_POSIX
+		strcpy(fname_temp,"/tmp/espeakXXXXXX");
+		if((fd_temp = mkstemp(fname_temp)) >= 0)
+		{
+			close(fd_temp);
+
+			if((f = fopen(fname_temp,"w+")) != NULL)
+			{
+				f_trans = f;   // write translation rule trace to a temp file
+			}
+		}
+#else
 		strcpy(fname_temp,tmpnam(NULL));
 		if((f = fopen(fname_temp,"w+")) != NULL)
 		{
 			f_trans = f;   // write translation rule trace to a temp file
 		}
+#endif
 	case T_TRANSLATE:
 	case MENU_SPEAK_TRANSLATE:
 		option_multibyte = espeakCHARS_AUTO;

src/voices.cpp

 
 	p = voice->languages;  // list of languages+dialects for which this voice is suitable
 
+	if(strcmp(spec_language,"mbrola")==0)
+	{
+		// only list mbrola voices
+		if(memcmp(voice->identifier,"mb/",3) == 0)
+			return(100);
+		return(0);
+	}
+
 	if(spec_n_parts == 0)
 	{
 		score = 100;
 	for(ix=0; ix<n_voices_list; ix++)
 	{
 		vp = voices_list[ix];
+
 		if(((control & 1) == 0) && (memcmp(vp->identifier,"mb/",3) == 0))
 			continue;
 

src/wavegen.cpp

 	if(embedded_value[EMBED_T] > 0)
 	{
 		// announcing punctuation
-		amp = embedded_value[EMBED_T] * 10;
-		delay = 130;
+		amp = embedded_value[EMBED_T] * 8;
+		delay = 60;
 	}
 
 	if(delay == 0)
 
 	echo_head = (delay * samplerate)/1000;
 	echo_length = echo_head;       // ensure completion of echo at the end of speech. Use 1 delay period?
+	if(amp == 0)
+		echo_length = 0;
 	if(amp > 20)
 		echo_length = echo_head * 2;    // perhaps allow 2 echo periods if the echo is loud.