[1.25.08] Fix to uft-8 recognizer.

lang=hu, set no secondary stress; [e:] [E] are closer, 
Don't consider mbrola voices during voice selection.
Fix voice variant not being set from SSML.


git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@23 d46cf337-b52f-0410-862d-fd96e6ae7743

dictsource/fr_list

 _17	d'is:'Et
 _18	d'is:'w^it
 _19	d'is:n'@f
-_20	v'E~ 
+_20	v'E~
 _21	v'E~te:E~
 _22	v'E~td'Y
 _23	v'E~t:tRw'a
 
 
 // conjonctions
-et	e $u $pause
+et	e $u $brk
-ou	$u $pause
+ou	$u $brk
-car	$u $pause
+car	$u $brk
-soit	$u $pause
+soit	$u $brk
-mais	$u $pause
+mais	$u $brk
-ou	u $u $pause
+ou	u $u $brk
-donc	$u $pause
+donc	$u $brk
-or	$u $pause
+or	$u $brk
-ni	$u $pause
+ni	$u $brk
-car	$u $pause
+car	$u $brk
-cependant  $u $pause
+cependant  $u $brk
-néanmoins  $u $pause
+néanmoins  $u $brk
-toutefois  $u $pause
+toutefois  $u $brk
-
+comme	   $u $brk
-
+lorsque	   $u $brk
-comme	   $u $pause
+puisque	   $u $brk
-lorsque	   $u $pause
+quand	   $u $brk
-puisque	   $u $pause
+que	   $u $brk
-quand	   $u $pause
+qui	   $u $brk
-que	   $u $pause
+quoique	   $u $brk
-qui	   $u $pause
+si	   $u $brk
-quoique	   $u $pause
+// (à cause que)   $u $brk
-si	   $u $pause
+// (ainsi que)     $u $brk
-
+// (à mesure que)  $u $brk
-// (à cause que)   $u $pause
+// (après que)     $u $brk
-// (ainsi que)     $u $pause
+// (à moins que)   $u $brk
-// (à mesure que)  $u $pause
+// (au lieu que)   $u $brk
-// (après que)     $u $pause
+// (aussitôt que)  $u $brk
-// (à moins que)   $u $pause
+// (pendant que)   $u $brk
-// (au lieu que)   $u $pause
-// (aussitôt que)  $u $pause
-// (pendant que)   $u $pause
 // (pourvu que)    $u $pause
 
 // pronouns
 leurs	$u+
 
 // prepositions
-de	$u $nounf
+de $u $nounf
 du	$u+ $nounf
 des	dEz2 $u $nounf
 ?2 des  dez2 $u+ $nounf
 au	$u $brk
 aux	$u $brk
 
-dans	$u+ $pause
+dans	$u+ $brk
-sur	$u+ $pause
+sur	$u+ $brk
-par	$u+ $pause
+par	$u+ $brk
-pour	$u+ $pause
+pour	$u+ $brk
 en	$u+ $brk
-(en effet) A~nEf'E $pause
+(en effet) A~nEf'E $brk
 
-vers	$u+ $pause
+vers	$u+ $brk
-avec	$u+ $pause
+avec	$u+ $brk
-sans	$u+ $pause
+sans	$u+ $brk
-sous	$u+ $pause
+sous	$u+ $brk
-entre	$u+ $pause
+entre	$u+ $brk
-contre	$u+ $pause
+contre	$u+ $brk
 
-durant	$pause
+durant	$brk
 
 // negative
 ne	$u
 fils	      fis
 mars	      maRs
 pays	      pEi
+pourcent    purs'A~
 sens	      sA~s
 c'est         sEt2 $u
 qu'à          ka

dictsource/hr_rules

 .group r
         r        R
      A) r (A     R
-     K) r (A     R
+     K) r (A     @-*
         rr       R
         r (_     R
 

dictsource/hu_list

 
 
 // Conjunctions
-és        $u $pause  // and
+és        $u $pause $strend2 // and
-de        $u $pause  // but
+de        $u+ $pause  // but
-vagy      $u $pause  // or
+vagy      $u+ $pause  // or
 
 mert      $pause     // because
 miatt     $pause     // because
 ha        $u $pause  // if
 
 
+// Prepositions etc.
+
+
 // Negative
 nem       $u+
 ne        $u+
 
 // Unstressed adverbs etc.
 néhány    $u+
+is        $u         // ??
 
 
 // Verbs

dictsource/hu_rules

 
 .group r
         r        R
-     A) r (A     R
+     A) r (A     *
      C) r (A     R
         r (C     r
         rr       R

dictsource/no_rules

         a        A:
         a (C%    A
         a (_     A
-     X) a (CCX   A  // single syllable with >= 2 final consonants
+     X) a (CCK   A  // single syllable with >= 2 final consonants
+                    // or initial syllable with 3 consonants
         ai       AI
         au       aU
 
         aa       o:   // å
         aa (C%   O
-     X) aa (CCX  O
+     X) aa (CCK  O
         ae       a:   // æ
         ae (C%   a
-     X) ae (CCX  a
+     X) ae (CCK  a
 
 
 .group b
 .group e
         e        e:
         e (C%    E
-     X) e (CCX   E
+     X) e (CCK   E
         ei       aI
         e (rC    a           //  ??
      &) e (_     @2
 .group i
         i        i:
         i (C%    I
-     X) i (CCX   I
+     X) i (CCK   I
      &) ig (_    I
 
 .group j
 .group o
         o        u:
         o (C%    O
-     X) o (CCX   O
+     X) o (CCK   O
         oi       OI
 
         oe       Y:   // ø
         oe (C%   W
-     X) oe (CCX  W
+     X) oe (CCK  W
         oey      Yy
 
         o (nd    U
 .group u
         u        u-:
         u (C%    u-
-     X) u (CCX   u-
+     X) u (CCK   u-
         ui       u-I
 
 .group v
 .group å
         å        o:
         å (C%    O
-     X) å (CCX   O
+     X) å (CCK   O
 
 .group æ
         æ        E:
         æ (r     a:
         æ (C%    a
-     X) æ (CCX   a
+     X) æ (CCK   a
 
 .group ø
         ø        Y:
         ø (C%    W
-     X) ø (CCX   W
+     X) ø (CCK   W
         øy       Yy
 
 .group ä
         ä        E:
         ä (C%    E
-     X) ä (CCX   E
+     X) ä (CCK   E
 
 .group ö
         ö        Y:
         ö (C%    W
-     X) ö (CCX   W
+     X) ö (CCK   W
 
 .group
         é        'e:

phsource/compile_report

   1  vowel/@_bck    hi
  13  vowel/e        en en_n af cy eo fr hu hr it pt pt_pt vi
   2  vowel/e#       en_sc
-  7  vowel/e_2      en_sc de hi hu sv no
+  6  vowel/e_2      en_sc de hi sv no
+  1  vowel/e_3      hu
   1  vowel/e_5      en_sc
   3  vowel/ee       fr pl sv
   2  vowel/e_e      en_sc is
   7  vowel/ee_1     en en_n pl it sv no zh_yue
-  4  vowel/ee_2     en cy hu nl
+  3  vowel/ee_2     en cy nl
   1  vowel/ee#_2    sv
   3  vowel/ee_3     af pt vi
   3  vowel/ee_6     en_n sk sv
- 11  vowel/e_mid    en_rp en_wm fr_ca hi sk cs hr es pt_pt no is
+ 12  vowel/e_mid    en_rp en_wm fr_ca hi hu sk cs hr es pt_pt no is
  10  vowel/e_mid2   af de fi nl sk ro el sw
   1  vowel/@_fnt    hr
   1  vowel/@_hgh    no

phsource/ph_hungarian

 phoneme E
   vowel starttype (e) endtype (e)
   length 130
-  formants vowel/ee_2
+  formants vowel/e_mid
 endphoneme
 
 phoneme e
 phoneme e:
   vowel starttype (e) endtype (e)
   length 240
-  formants vowel/e_2
+  formants vowel/e_3
 endphoneme
 
 phoneme A

src/espeakedit.cpp

 	// to something other than the default "C".  Then, not only Latin1 but also the
 	// other characters give the correct results with iswalpha() etc.
 	if(setlocale(LC_CTYPE,"en_US.UTF-8") == NULL)
-		setlocale(LC_CTYPE,"");
+	{
+		if(setlocale(LC_CTYPE,"UTF-8") == NULL)
+			setlocale(LC_CTYPE,"");
+	}
 
 
   // Create the main frame window

src/extras.cpp

 
 
 
-char* text[]=
-{
-  "<voice xml:lang=\"en\" variant=\"1\">eSpeak.</voice>",
-  " <voice xml:lang=\"en\" gender=\"male\" variant=\"1\"><prosody rate=\"1.0\"> normally would, </prosody></voice>"
-};
-
 
 #ifdef deleted
 void Test2()
 }
 #endif
 
+char* text1[]=
+{
+  "<voice xml:lang=\"en\" variant=\"2\">eSpeak.</voice>",
+  " <voice xml:lang=\"fr\" gender=\"male\" variant=\"3\"><prosody rate=\"1.0\"> normally would, </prosody></voice>"
+};
 
 void TestTest(int control)
 {//=======================
 //CharsetToUnicode("ISCII");
 
 
-return;
+//return;
 
 if(control==2)
 {
 
 	espeak_Initialize(AUDIO_OUTPUT_SYNCHRONOUS,100,NULL);
 	espeak_SetSynthCallback(TestSynthCallback);
-//	espeak_SetUriCallback(TestUriCallback);
-
-	voice.languages = "fr";
-	voice.gender = 0;
-	voice.age = 0;
-	voice.variant = 0;
-	ix = espeak_SetVoiceByName("pt-xx");
 
-	voice2 = espeak_GetCurrentVoice();
+  unsigned int unique_identifier=0;
+  int index=0;
+  espeak_Synth(text1[index], strlen(text1[index])+1, 0, POS_CHARACTER, 0,  espeakSSML|espeakCHARS_UTF8, &unique_identifier, NULL);
 
-//	espeak_SetVoiceByProperties(&voice);
+  espeak_SetParameter(espeakPUNCTUATION, 1, 0);
-	espeak_SetParameter(espeakRANGE,0,0);
+  espeak_Synchronize();
-	espeak_Synth(textbuf,ix+1,0,POS_CHARACTER,0,espeakCHARS_8BIT+espeakSSML,NULL,NULL);
+  //  espeak_Cancel();
+  espeak_SetParameter(espeakPUNCTUATION, 1, 0);
 
-//	for(voice.variant = 1; voice.variant < 9; voice.variant++)
+  index++;
-//	{
+  espeak_Synth(text1[index], strlen(text1[index])+1, 0, POS_CHARACTER, 0, espeakSSML|espeakCHARS_UTF8, &unique_identifier, NULL);
-//		espeak_SetVoiceByProperties(&voice);
-//		espeak_Synth(textbuf,ix+1,0,POS_CHARACTER,0,espeakCHARS_8BIT+espeakSSML,NULL,NULL);
-//	}
 }
 
 

src/readclause.cpp

 					break;
 				}
 
-				if(((c2 & 0xc0) != 0x80) && ((c2 & m) != 0))
+				if((c2 & 0xc0) != 0x80)
 				{
 					GetC_unget(c2);
 					break;
 			return(0);  // <s> or <p> without language spec, nothing to do
 	
 		ssml_sp = &ssml_stack[n_ssml_stack++];
-	
+
 		attrcopy_utf8(ssml_sp->language,lang,sizeof(ssml_sp->language));
 		attrcopy_utf8(ssml_sp->voice_name,name,sizeof(ssml_sp->voice_name));
 		ssml_sp->voice_variant = attrnumber(variant,1,0)-1;

src/speak.cpp

    setlocale(LC_CTYPE,"ISO8859-1");
 #else
 	if(setlocale(LC_CTYPE,"en_US.UTF-8") == NULL)
-		setlocale(LC_CTYPE,"");
+	{
+		if(setlocale(LC_CTYPE,"UTF-8") == NULL)
+			setlocale(LC_CTYPE,"");
+	}
 #endif
 
 

src/speak_lib.cpp

   // to something other than the default "C".  Then, not only Latin1 but also the
   // other characters give the correct results with iswalpha() etc.
 #ifdef PLATFORM_RISCOS
-  setlocale(LC_CTYPE,"ISO8859-1");
+	setlocale(LC_CTYPE,"ISO8859-1");
 #else
-  if(setlocale(LC_CTYPE,"en_US.UTF-8") == NULL)
+	if(setlocale(LC_CTYPE,"en_US.UTF-8") == NULL)
-    setlocale(LC_CTYPE,"");
+	{
+		if(setlocale(LC_CTYPE,"UTF-8") == NULL)
+			setlocale(LC_CTYPE,"");
+	}
 #endif
 
   init_path(path);

src/synthdata.cpp

 #include "translate.h"
 #include "wave.h"
 
-const char *version_string = "1.25.06  23.May.07";
+const char *version_string = "1.25.08  25.May.07";
 const int version_phdata  = 0x012501;
 
 int option_device_number = -1;

src/tr_languages.cpp

 
 			tr->langopts.vowel_pause = 0x10;
 			tr->langopts.stress_rule = 0;
-			tr->langopts.stress_flags = 0x16;
+			tr->langopts.stress_flags = 0x36;
+			tr->langopts.unstressed_wd1 = 2;
 			tr->langopts.param[LOPT_REGRESSIVE_VOICING] = 0x4;  // don't propagate over word boundaries
 			tr->langopts.param[LOPT_IT_DOUBLING] = 1;
 			tr->langopts.long_stop = 130;

src/voices.cpp

 	if((v != NULL) && (variant > 0))
 	{
 		sprintf(buf,"!variant%d",variant);
-		LoadVoice(buf,2);
+		v = LoadVoice(buf,2);
 	}
 	return(v);
 }
 }  // end of ScoreVoice
 
 
-static int SetVoiceScores(espeak_VOICE *voice_select, espeak_VOICE **voices)
+static int SetVoiceScores(espeak_VOICE *voice_select, espeak_VOICE **voices, int control)
-{//=========================================================================
+{//======================================================================================
+// control: bit0=1  include mbrola voices
 	int ix;
 	int score;
 	int nv;           // number of candidates
 	nv = 0;
 	for(ix=0; ix<n_voices_list; ix++)
 	{
+		vp = voices_list[ix];
+		if(((control & 1) == 0) && (memcmp(vp->identifier,"mb/",3) == 0))
+			continue;
+
 		if((score = ScoreVoice(voice_select, n_parts, lang_len, voices_list[ix])) > 0)
 		{
-			voices[nv++] = vp = voices_list[ix];
+			voices[nv++] = vp;
 			vp->score = score;
 		}
 	}
 		espeak_ListVoices(NULL);   // create the voices list
 
 	// select and sort voices for the required language
-	nv = SetVoiceScores(voice_select,voices);
+	nv = SetVoiceScores(voice_select,voices,0);
 
 	if(nv == 0)
 	{
 	if(voice_spec)
 	{
 		// select the voices which match the voice_spec, and sort them by preference
-		SetVoiceScores(voice_spec,voices);
+		SetVoiceScores(voice_spec,voices,1);
 		return((const espeak_VOICE **)voices);
 	}