[1.29.29] Add inital draft of language "ta", Tamil

lang=en: Don't treat "Mc" prefix as a separate word, combine it with the following word.


git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@113 d46cf337-b52f-0410-862d-fd96e6ae7743

dictsource/dict_phonemes

@@ -262,6 +262,10 @@ l    m    n    N    p    r    s    t
 tS   v    w    z    
 
 
+Dictionary ta_dict
+
+
+
 Dictionary pt_dict
 
 &    &/   &U~  &~   @    @-   a    A    

dictsource/en_list

@@ -169,7 +169,7 @@ i.e	aIi:_!   $pause
 irc	$abbrev
 lbs	paUndz
 ltd	lImItId
-mc	m@k
+mc	m@k  $alt2   // combine with the following word
 oem	$abbrev
 ok	$abbrev
 omg	$abbrev

dictsource/en_rules

@@ -3227,7 +3227,9 @@
         li) mb (e        m
          o) mb (e        m
         du) mb (e        m
-         _) m (c         m@
+         _) mc           m@k
+         _) mc (A        mak
+         _) mc (-A       mak
             mn (_        m
          _) mn           n
         &B) me (_        mI2
@@ -4965,6 +4967,7 @@
          t) y (rann      I
         ph) ys (i        Iz
             yr (_        @
+            yre (_       aI3
          c) y (ni        I
          l) y (ri        I
          p) y (ra        I

dictsource/mk_list

@@ -162,37 +162,7 @@ _1M4  bili'on
 _dpt  _:z'api*k&_
 _0and i
 
-// Roman numerals
+
-ii	$abbrev
-iii	$abbrev
-iv	$abbrev
-vii	$abbrev
-viii	$abbrev
-ix	$abbrev
-xi	$abbrev
-xii	$abbrev
-xiii	$abbrev
-xiv	$abbrev
-xvi	$abbrev
-xvii	$abbrev
-xviii	$abbrev
-xix	$abbrev
-xxi	$abbrev
-xxii	$abbrev
-xxiii	$abbrev
-xxiv	$abbrev
-xxvi	$abbrev
-xxvii	$abbrev
-xxviii	$abbrev
-xxix	$abbrev
-xxxi	$abbrev
-xxxii	$abbrev
-xxxiii	$abbrev
-xxxiv	$abbrev
-xxxvi	$abbrev
-xxxvii	$abbrev
-xxxviii  $abbrev
-xxxix	$abbrev
 
 // Abbreviations
 

dictsource/ta_list

@@ -0,0 +1,4 @@
+ 
+// This file is UTF8 encoded 
+// Spelling to phoneme words and exceptions for Tamil
+

dictsource/ta_rules

@@ -0,0 +1,173 @@
+
+// This file is UTF8 encoded 
+// Spelling to phoneme rules for Tamil
+
+.replace
+௦	0	// Tamil numbers
+௧	1
+௨	2
+௩	3
+௪	4
+௫	5
+௬	6
+௭	7
+௮	8
+௯	9
+
+
+.group 0xe0ae
+	ஂ		// anusvara
+
+	ஃ		// visarga
+
+	அ	a
+
+	ஆ	a:
+
+	இ	i
+
+	ஈ	i:
+
+	உ	u
+
+	ஊ	u:
+
+	எ	e
+
+	ஏ	e:
+
+	ஐ	E:
+
+	ஒ	o
+
+	ஓ	o:
+
+	ஔ	aU
+
+// consonants
+
+	க	ga   // inter-vocalic, unless there is virama before or after
+	க (B	g
+	க (் 	g
+     ்) க	ga
+     ்) க (B	g
+     _) க	ka
+     _) க (B	k
+	க்க	ka
+	க்க (B	k
+
+	ங	Na
+	ங (B	N
+
+	ச	dZa
+	ச (B	dZ
+     _) ச	sa
+     _) ச (B	s
+	ச்ச	tSa
+	ச்ச (B	tS
+
+	ஜ	dZa
+	ஜ (B	dZ
+
+	ஞ	n^a
+	ஞ (B	n^
+
+	ட	d.a
+	ட (B	d.
+     _) ட	t.a
+     _) ட (B	t.
+	ட்ட	t.a
+	ட்ட (B	t.
+
+	ண	n.a
+	ண (B	n.
+
+	த	da
+	த (B	d
+     _) த	ta
+     _) த (B	t
+	த்த	ta
+	த்த (B	t
+
+	ந	na
+	ந (B	n
+
+	ன	na
+	ன (B	n
+
+	ப	ba
+	ப (B	b
+     _) ப	pa
+     _) ப (B	p
+	ப்ப	pa
+	ப்ப (B	p
+
+	ம	ma
+	ம (B	m
+
+	ய	ja
+	ய (B	j
+
+	ர	ra
+	ர (B	r
+
+	ற	Ra
+	ற (B	R
+	ற் (ற	t	// RR -> tR
+
+	ல	la
+	ல (B	l
+
+	ள	l/2a    // this should be [l.] not [l/2]
+	ள (B	l/2
+
+	ழ	z.a
+	ழ (B	z.
+
+	வ	va
+	வ (B	v
+
+	ஶ	Sa
+	ஶ (B	S
+
+	ஷ	s.a
+	ஷ (B	s.
+
+	ஸ	sa
+	ஸ (B	s
+
+	ஹ	ha
+	ஹ (B	h
+
+
+// combining vowels
+
+	ா	a:
+
+	ி	i
+
+.group 0xe0af
+	ீ	i:
+
+	ு	u
+
+	ூ	u:
+
+	ெ	e
+
+	ே	e:
+
+	ை	E:
+
+	ொ	o
+
+	ோ	o:
+
+	ௌ	aU
+
+	்		// virama
+
+	ௗ	:	// aU length mark
+
+.group
+	$	dola

espeak-data/voices/grc

@@ -4,5 +4,5 @@ gender male
 
 stressLength 170 170  190 190  0 0  230 240
 dictrules 1
-words 1
+words 3
 

espeak-data/voices/ta

@@ -0,0 +1,4 @@
+name tamil-test
+language ta
+gender male
+

phsource/compile_report

@@ -1,4 +1,4 @@
-41 phoneme tables
+42 phoneme tables
           new total
     base  96   96
    base2  24  115
@@ -16,6 +16,7 @@
       fr  34  118
    fr_ca  11  118
       hi  50  132
+      ta  12  134
       hu  23  112
       nl  26  117
       pl  15  107
@@ -23,7 +24,7 @@
       cs   6  125
       hr  20  130
       mk   3  131
-      sr  11  130
+      sr  14  130
       ru  38  126
       it  17  118
       la  21  114
@@ -427,6 +428,7 @@ r3/r_trill      [R2] base
                 [r] cy
                 [r] nl
                 [R] sk
+                [r*] sr
                 [x] pt
 r3/r_trill2     [R] base
                 [R] hr
@@ -438,6 +440,7 @@ r3/r_trill3.wav  [R3] base
                 [r] nl
 r3/r_trill.wav  [R2] base
                 [R] sk
+                [r*] sr
                 [x] pt
 r3/r_u          [(u)] base
 r3/rx           [*] base
@@ -699,6 +702,7 @@ ustop/t_dnt     [t] base
                 [t] hu
                 [t] pl
                 [t] hr
+                [t] sr
                 [t] ru
                 [t;] ro
                 [t[] vi
@@ -741,7 +745,6 @@ ustop/ts_pzd_   [c] hi
 ustop/ts_pzd2   [c] hi
                 [c] hu
 ustop/t_sr      [d] sr
-                [t] sr
 ustop/ts_rfx    [ts.] zh
 ustop/ts_rfx_unasp  [ts.] zh
 ustop/ts_sr.wav  [tS;] sr
@@ -1106,6 +1109,7 @@ vowel/a         [a] en_n
                 [a] zh
 vowel/a#        [a/] base2
                 [a2] en_sc
+                [&] sr
                 [&] pt
                 [&/] pt
                 [a2] en_wi
@@ -1120,7 +1124,8 @@ vowel/a_2       [a] base2
                 [A:] en_wi
                 [A@] en_wi
                 [aa] en_wi
-vowel/a#_2      [&] hr
+vowel/a#_2      [a] ta
+                [&] hr
                 [a2] sv
                 [&] is
                 [a/] sw
@@ -1150,6 +1155,7 @@ vowel/a_5       [a/] it
                 [a] sw
 vowel/aa        [a] fi
                 [A] fr_ca
+                [a:] ta
                 [A] no
                 [A:] no
                 [aa] zhy
@@ -1185,6 +1191,8 @@ vowel/e         [e] base2
                 [e] eo
                 [e] fr
                 [E:] fr
+                [e] ta
+                [e:] ta
                 [e] hr
                 [e/] it
                 [e:] la
@@ -1267,6 +1275,7 @@ vowel/i         [i] base2
                 [i:] cy
                 [i] eo
                 [I] fr
+                [i:] ta
                 [i:] hu
                 [i] pl
                 [i] it
@@ -1301,6 +1310,7 @@ vowel/i#_5      [y] pt_pt
                 [i/] pt_pt
                 [y] ro
 vowel/i_6       [i:] en_us
+                [i] ta
                 [i] hr
 vowel/i_en      [i:] en
                 [i] vi
@@ -1338,6 +1348,7 @@ vowel/ii_5      [I] en_wi
                 [I2] en_wi
 vowel/ii_6      [I] en_wm
                 [I2] en_wm
+                [I] sr
 vowel/ii_en     [i] en
 vowel/@_low     [3] en_rp
                 [@] en_rp
@@ -1353,6 +1364,8 @@ vowel/o         [o] base2
                 [o:] en
                 [o:] de
                 [o] hi
+                [o] ta
+                [o:] ta
                 [o] it
                 [o/] it
                 [o:] la
@@ -1450,7 +1463,8 @@ vowel/u_2       [u1] fi
                 [u] sk
                 [U] hr
                 [u] ro
-vowel/u#_2      [u-] sv
+vowel/u#_2      [u-] ta
+                [u-] sv
 vowel/u_3       [yU] ro
 vowel/u_4       [u:] en_n
 vowel/u#_4      [U] en_sc
@@ -1465,6 +1479,7 @@ vowel/u_bck     [u] base2
                 [u] fi
                 [u] fr_ca
                 [u:] hi
+                [u:] ta
                 [u:] hu
                 [u] nl
                 [u] pl
@@ -1489,6 +1504,7 @@ vowel/uu_4      [U] fi
 vowel/uu_bck    [U] en_n
                 [U] fr_ca
                 [U] hi
+                [u] ta
                 [u] hu
                 [U] pt
                 [U] no

phsource/ph_serbian

@@ -13,14 +13,39 @@
 // Only phonemes which differ from the inherited phonemes need
 // to be included here.
 
+phoneme I
+  vowel starttype (i) endtype (i)
+  length 120
+  formants vowel/ii_6
+  linkout ;
+endphoneme
+
+
+phoneme &
+  vowel starttype (a) endtype (a)
+  length 140
+  formants vowel/a#
+endphoneme
+
+
+phoneme r*   // this is [R] from Slovak/Czech
+  liquid
+  vowelin  f1=0  f2=1700 -300 300  f3=-300 80
+  vowelout f1=2  f2=1700 -300 300  f3=-300 80 brk
+  formants r3/r_trill+r3/r_trill.wav%50
+  trill
+  lengthmod 6
+endphoneme
+
+
 
 phoneme t        // dental variant of /t/
   vls dnt stop
   vowelin f1=0  f2=1600 -300 300  f3=-100 80
   vowelout f1=0 f2=1600 -300 250  f3=-100 80  rms=20
   lengthmod 2
-  wave  ustop/t_sr%50
+  wave  ustop/t_dnt%50
-  before _ ustop/t_sr%35
+  before _ ustop/t_dnt%50
   switchvoicing d
 endphoneme
 
@@ -97,8 +122,8 @@ phoneme tS
   vls pla afr sibilant
   vowelin f1=0  f2=2300 200 400  f3=-100 80
   lengthmod 2
-  wave ustop/tsh_sr.wav
+  wave ustop/tsh_sr.wav%50
-  before _ ustop/tsh_sr.wav
+  before _ ustop/tsh_sr.wav%50
   switchvoicing dZ
 endphoneme
 
@@ -113,12 +138,14 @@ phoneme tS;
 endphoneme
 
 
+
+
 phoneme Z
   vcd pla frc sibilant
   vowelin  f1=0  f2=2000 0 300  f3=-200 80
   vowelout f1=2  f2=2000 0 300  f3=-200 80 brk
-  formants voc/zh+ufric/sh_sr.wav
+  formants voc/zh+ufric/sh_sr.wav%60
-  before _ voc/zh+ufric/sh_sr.wav
+  before _ voc/zh+ufric/sh_sr.wav%50
   lengthmod 6
   switchvoicing S
 endphoneme
@@ -130,8 +157,8 @@ phoneme S
   vowelout f1=0  f2=2100 100 300  f3=-100 80
   length 100
   lengthmod 3
-  wave ufric/sh_sr.wav
+  wave ufric/sh_sr.wav%70
-  before _ ufric/sh_sr.wav
+  before _ ufric/sh_sr.wav%60
   switchvoicing Z
 endphoneme
 

phsource/ph_tamil

@@ -0,0 +1,84 @@
+
+// Phonemes for Tamil
+// This inherits from ph_hindi
+
+
+
+phoneme a
+  vowel  starttype (@)  endtype (@)
+  length 130
+  formants vowel/a#_2
+endphoneme
+
+
+phoneme a:
+  vowel  starttype (a)  endtype (a)
+  length 270
+  formants vowel/aa
+endphoneme
+
+
+phoneme e
+  vowel  starttype (e)  endtype (e)
+  length 130
+  formants vowel/e
+endphoneme
+
+
+phoneme e:
+  vowel  starttype (e)  endtype (e)
+  length 270
+  formants vowel/e
+endphoneme
+
+
+phoneme i
+  vowel  starttype (i)  endtype (i)
+  length 130
+  formants vowel/i_6
+endphoneme
+
+
+phoneme i:
+  vowel  starttype (i)  endtype (i)
+  length 270
+  formants vowel/i
+endphoneme
+
+
+
+phoneme o
+  vowel  starttype (o)  endtype (o)
+  length 130
+  formants vowel/o
+endphoneme
+
+
+phoneme o:
+  vowel  starttype (o)  endtype (o)
+  length 270
+  formants vowel/o
+endphoneme
+
+
+phoneme u
+  vowel  starttype (u)  endtype (u)
+  length 130
+  formants vowel/uu_bck
+endphoneme
+
+
+phoneme u:
+  vowel  starttype (u)  endtype (u)
+  length 270
+  formants vowel/u_bck
+endphoneme
+
+
+phoneme u-
+  vowel starttype (u) endtype (u)
+  length 130
+  formants vowel/u#_2
+endphoneme
+
+

phsource/ph_zh

@@ -0,0 +1,407 @@
+//based on "phonemes"
+//====================================================
+//  Tone Numbers
+//====================================================
+
+phoneme 11   //  tone: low level
+  stress
+  tone 15 10 envelope/p_fall NULL
+endphoneme
+
+phoneme 21   //  tone: low fall
+  stress
+  tone 20 10 envelope/p_fall NULL
+endphoneme
+
+phoneme 214   //  tone: fall rise
+  stress
+  tone 20 40 envelope/p_214 NULL
+endphoneme
+
+phoneme 22   //  tone: mid-low level
+  stress
+  tone 21 19 envelope/p_fall NULL
+endphoneme
+
+phoneme 33   //  tone: mid level
+  stress
+  tone 31 29 envelope/p_fall NULL
+endphoneme
+
+phoneme 35   //  tone: mid rise
+  stress
+  tone 30 50 envelope/p_rise NULL
+endphoneme
+
+phoneme 44   //  tone: mid-high level
+  stress
+  tone 41 39 envelope/p_level NULL
+endphoneme
+
+phoneme 51   //  tone: high fall
+  stress
+  tone 50 10 envelope/p_fall NULL
+endphoneme
+
+phoneme 53   //  tone: high fall
+  stress
+  tone 50 30 envelope/p_fall NULL
+endphoneme
+
+phoneme 55   //  tone: high level
+  stress
+  tone 50 50 envelope/p_level NULL
+endphoneme
+
+//====================================================
+//  Consonants
+//====================================================
+
+phoneme p
+  vls blb stop
+  vowelin  f1=0  f2=1000 -50 -100  f3=-200 80 amp=11
+  vowelout f1=0  f2=1000 -500 -350  f3=-300 80 rms=22
+  lengthmod 2
+  wave ustop/p_unasp
+endphoneme
+
+phoneme ph
+  vls blb stop
+  vowelin  f1=0  f2=1000 -50 -100  f3=-200 80 amp=11
+  vowelout f1=0  f2=1000 -500 -350  f3=-300 80 rms=22
+  lengthmod 2
+  wave ustop/p_
+endphoneme
+
+phoneme m //not clear before some vowel
+  vcd blb nasal
+  vowelout f1=2  f2=1000 -500 -350  f3=-200 80 brk
+  lengthmod 2
+  formants m/_m
+endphoneme
+
+phoneme f //not clear before some vowel
+  vls lbd frc
+  vowelout f1=0  f2=1000 -500 -350  f3=-200 80
+  lengthmod 2
+  wave ufric/f
+endphoneme
+
+phoneme t
+  vls alv stop
+  vowelin f1=0  f2=1700 -300 300  f3=-100 80
+  vowelout f1=0 f2=1700 -300 250  f3=-100 80  rms=20
+  lengthmod 2
+  wave  ustop/t_unasp
+endphoneme
+
+phoneme th
+  vls alv stop
+  vowelin f1=0  f2=1700 -300 300  f3=-100 80
+  vowelout f1=0 f2=1700 -300 250  f3=-100 80  rms=20
+  lengthmod 2
+  wave  ustop/t_
+endphoneme
+
+phoneme n
+  vcd alv nasal
+  vowelout f1=2 f2=1700 -300 250  f3=-100 80  rms=20 brk
+  length 30
+  lengthmod 2
+  formants n/_n
+  before _ n/n_
+  before (i) n/ni
+endphoneme
+
+phoneme l
+  liquid
+  vowelin f1=0  f2=2700 400 600  f3=300 80 rate len=20
+  lengthmod 2
+  formants l/_l
+  before (i) l/li
+endphoneme
+
+phoneme k
+  vls uvl stop
+  vowelin f1=1  f2=1700 0 200  f3=-300 80 f4
+  vowelout f1=1 f2=1700 0 200  f3=-300 80 f4 rms=30
+  lengthmod 2
+  wave ustop/k_unasp_
+endphoneme
+
+phoneme kh
+  vls uvl stop
+  vowelin f1=1  f2=1700 0 200  f3=-300 80 f4
+  vowelout f1=1 f2=1700 0 200  f3=-300 80 f4 rms=30
+  lengthmod 2
+  wave ustop/k_
+endphoneme
+
+phoneme tS;
+  vls pal afr sibilant palatal
+  vowelin f1=0  f2=2700 400 600  f3=300 80
+  lengthmod 2
+  wave ustop/tsh_pzd_unasp
+endphoneme
+
+phoneme tS;h
+  vls pal afr sibilant palatal
+  vowelin f1=0  f2=2700 400 600  f3=300 80
+  lengthmod 2
+  wave ustop/tsh_pzd
+endphoneme
+
+phoneme S;
+  vls pla frc sibilant
+  vowelin f1=0  f2=2700 400 600  f3=300 80
+  lengthmod 2
+  wave ufric/sh_pzd
+endphoneme
+
+phoneme ts
+  vls alv afr sibilant
+  vowelin f1=0  f2=1700 -300 300  f3=-100 80
+  vowelout f1=0 f2=1700 -300 250  f3=-100 80  rms=20
+  lengthmod 2
+  wave ustop/ts_unasp
+endphoneme
+
+phoneme tsh
+  vls alv afr sibilant
+  vowelin f1=0  f2=1700 -300 300  f3=-100 80
+  vowelout f1=0 f2=1700 -300 250  f3=-100 80  rms=20
+  lengthmod 2
+  wave ustop/ts
+endphoneme
+
+phoneme ts.
+  vls pla frc sibilant
+  vowelin  f1=0  f2=1800 -100 300  f3=-300 80
+  vowelout f1=0  f2=1800 -100 300  f3=-300 80
+  length 100
+  lengthmod 3
+  wave ustop/ts_rfx_unasp
+endphoneme
+
+phoneme ts.h
+  vls pla frc sibilant
+  vowelin  f1=0  f2=1800 -100 300  f3=-300 80
+  vowelout f1=0  f2=1800 -100 300  f3=-300 80
+  length 100
+  lengthmod 3
+  wave ustop/ts_rfx
+endphoneme
+
+phoneme j
+  liquid
+  palatal
+  length 40
+  formants j/_j
+  lengthmod 1
+endphoneme
+
+phoneme w
+  liquid
+  length 40
+  lengthmod 1
+  formants w/_w
+endphoneme
+
+//====================================================
+//  Vowels and Diphtnongs
+//====================================================
+
+phoneme N //should be more different from n
+  vcd vel nasal
+  length 10
+  formants NULL
+  before _ nn/nn_
+  after _ nn/_nn
+  lengthmod 1
+  after (i) nn/inn
+endphoneme
+
+phoneme a
+  vowel starttype (a) endtype (a)
+  length 250
+  formants vowel/a
+endphoneme
+
+phoneme A
+  vowel starttype (a) endtype (a)
+  length 250
+  formants vowel/aa
+  before N vowel/aa_2
+endphoneme
+
+phoneme ai
+  vowel starttype (a) endtype (i)
+  length 250
+  formants vdiph/ae
+endphoneme
+
+phoneme Au
+  vowel starttype (a) endtype (u)
+  length 250
+  formants vwl_zh/aau
+endphoneme
+
+phoneme @
+  vowel  starttype (@)  endtype (@)
+  length 230
+  formants vowel/3_2%120
+  before N vowel/@_bck
+endphoneme
+
+phoneme @r  //tmp use V3_r
+  vowel  starttype (@)  endtype (@)
+  length 250
+  formants vowelr/V3_r
+endphoneme
+
+phoneme E
+  vowel  starttype (e)  endtype (e)
+  length 250
+  formants vowel/ee_2
+endphoneme
+
+phoneme ei
+  vowel starttype (e) endtype (i)
+  length 250
+  formants vdiph/ei
+endphoneme
+
+phoneme i
+  vowel  starttype (i)  endtype (i)
+  length 250
+  formants vowel/i
+  before N vowel/ii_2
+endphoneme
+
+phoneme i[ //after ts tsh s
+  vowel  starttype (i)  endtype (i)
+  length 250
+  formants vwl_zh/i_dnt
+endphoneme
+
+phoneme i. //after ts. ts.h s. z.
+  vowel  starttype (i)  endtype (i)
+  length 250
+  formants vwl_zh/i_rfx
+endphoneme
+
+phoneme iA  //try more before N
+  vowel  starttype (i)  endtype (a)
+  length 250
+  formants vwl_zh/iaa
+endphoneme
+
+phoneme iAu
+  vowel  starttype (i)  endtype (u)
+  length 250
+  formants vwl_zh/iaau
+endphoneme
+
+phoneme iE
+  vowel  starttype (i)  endtype (e)
+  length 250
+  formants vwl_zh/iee
+endphoneme
+
+phoneme io
+  vowel starttype (i) endtype (o)
+  length 250
+  formants vdiph2/iioo
+endphoneme
+
+phoneme iou
+  vowel starttype (i) endtype (u)
+  length 250
+  formants vwl_zh/iou
+endphoneme
+
+phoneme o
+  vowel starttype (o) endtype (o)
+  length 250
+  formants vowel/o
+endphoneme
+
+phoneme o-
+  vowel  starttype (o)  endtype (o)
+  length 250
+  formants vowel/o-
+endphoneme
+
+phoneme ou
+  vowel starttype (o) endtype (u)
+  length 250
+  formants vdiph/8u
+endphoneme
+
+phoneme u //try more before N
+  vowel starttype (u) endtype (u)
+  length 250
+  formants vowel/u
+  before N vowel/o
+endphoneme
+
+phoneme uA
+  vowel starttype (u) endtype (a)
+  length 250
+  formants vwl_zh/uaa
+endphoneme
+
+phoneme ua
+  vowel starttype (u) endtype (a)
+  length 250
+  formants vdiph2/oa
+endphoneme
+
+phoneme uai
+  vowel starttype (u) endtype (i)
+  length 250
+  formants vwl_zh/uai
+endphoneme
+
+phoneme u@
+  vowel starttype (u) endtype (@)
+  length 250
+  formants vdiph2/o@
+endphoneme
+
+phoneme uei
+  vowel starttype (u) endtype (i)
+  length 250
+  formants vwl_zh/uei
+endphoneme
+
+phoneme uo
+  vowel starttype (u) endtype (o)
+  length 250
+  formants vwl_zh/uo
+endphoneme
+
+phoneme y  //try more before N
+  vowel  starttype (i)  endtype (u)
+  length 250
+  formants vowel/y
+  before N vdiph2/yu
+endphoneme
+
+phoneme y&
+  vowel  starttype (i)  endtype (e)
+  length 250
+  formants vwl_zh/y&
+endphoneme
+
+phoneme yE  //try more so not similar to iE
+  vowel  starttype (i)  endtype (e)
+  length 250
+  formants vwl_zh/yee
+endphoneme
+
+phoneme y@
+  vowel  starttype (i)  endtype (@)
+  length 250
+  formants vdiph2/y#@
+endphoneme

phsource/phonemes

@@ -1172,6 +1172,9 @@ include ph_french_ca
 phonemetable hi base
 include ph_hindi
 
+phonemetable ta hi
+include ph_tamil
+
 phonemetable hu base
 include ph_hungarian
 

src/dictionary.cpp

@@ -976,7 +976,7 @@ void Translator::SetWordStress(char *output, unsigned int dictionary_flags, int
 	int done;
 
 	unsigned char vowel_stress[N_WORD_PHONEMES/2];
-	char syllable_type[N_WORD_PHONEMES/2];
+	char syllable_weight[N_WORD_PHONEMES/2];
 	unsigned char phonetic[N_WORD_PHONEMES];
 
 	static char consonant_types[16] = {0,0,0,1,1,1,1,1,1,1,0,0,0,0,0,0};
@@ -1040,7 +1040,7 @@ void Translator::SetWordStress(char *output, unsigned int dictionary_flags, int
 				// followed by two consonants, a long consonant, or consonant and end-of-word
 				weight++;
 			}
-			syllable_type[ix] = weight;
+			syllable_weight[ix] = weight;
 			ix++;
 		}
 	}
@@ -1164,16 +1164,14 @@ void Translator::SetWordStress(char *output, unsigned int dictionary_flags, int
 		{
 			int wt;
 			int max_weight = -1;
-			int min_weight = 10;
 			int prev_stressed;
 
-			for(ix = 1; ix < vowel_count; ix++)
+			// find the heaviest syllable, excluding the final syllable
+			for(ix = 1; ix < (vowel_count-1); ix++)
 			{
 				if(vowel_stress[ix] == 0)
 				{
-					if((wt = syllable_type[ix]) < min_weight)
+					if((wt = syllable_weight[ix]) >= max_weight)
-						min_weight = wt;
-					if(wt >= max_weight)
 					{
 						max_weight = wt;
 						prev_stressed = stressed_syllable;
@@ -1181,14 +1179,19 @@ void Translator::SetWordStress(char *output, unsigned int dictionary_flags, int
 					}
 				}
 			}
-			if(max_weight > min_weight)
+
+			if((syllable_weight[vowel_count-1] == 2) &&  (max_weight< 2))
 			{
-				// different weights, don't stress the final syllable
+				// the only double=heavy syllable is the final syllable, so stress this
-				if((stressed_syllable == vowel_count-1) && (syllable_type[vowel_count-2] > 0))
+				stressed_syllable = vowel_count-1;
-				{
-					stressed_syllable = vowel_count-2;
-				}
 			}
+			else
+			if(max_weight <= 0)
+			{
+				// all syllables, exclusing the last, are light. Stress the first syllable
+				stressed_syllable = 1;
+			}
+
 			vowel_stress[stressed_syllable] = 4;
 			max_stress = 4;
 		}
@@ -1253,7 +1256,7 @@ void Translator::SetWordStress(char *output, unsigned int dictionary_flags, int
 				if((stress == 3) && (langopts.stress_flags & 0x20))
 					continue;      // don't use secondary stress
 
-				if((v > 1) && (langopts.stress_flags & 0x40) && (syllable_type[v]==0) && (syllable_type[v+1]>0))
+				if((v > 1) && (langopts.stress_flags & 0x40) && (syllable_weight[v]==0) && (syllable_weight[v+1]>0))
 				{
 					// don't put secondary stress on a light syllable which is followed by a heavy syllable
 					continue;

src/intonation.cpp

@@ -43,7 +43,7 @@ typedef struct {
 	short pitch2;
 } SYLLABLE;
 
-SYLLABLE syllable_tab[N_PHONEME_LIST];
+SYLLABLE *syllable_tab;
 
 
 static int tone_pitch_env;    /* used to return pitch envelope */
@@ -242,7 +242,7 @@ static int drops_0[8] = {0x400,0x400,0x700,0x700,0x700,0xa00,0x1800,0x0e00};
 static short oflow[] = {0, 20, 12, 4, 0};
 static short oflow_emf[] = {5, 24, 15, 10, 5};
 static short oflow_less[] = {1, 17, 10, 5, 1};
-static short back_emf[] = {36, 32, 0};
+static short back_emf[] = {35, 32, 0};
 
 typedef struct {
 	unsigned char pitch_env0;     /* pitch envelope, tonic syllable at end */
@@ -285,7 +285,7 @@ static TONE_TABLE tone_table[N_TONE_TABLE] = {
    20, 25,   34, 20,  drops_0, 3, 3,   5, oflow, NULL, 15, 29, 0},
 
    {PITCHfall, 41, 4,  PITCHfall, 41, 27,              // exclamation
-   20, 25,   34, 24,  drops_0, 3, 4,   5, oflow_emf, back_emf, 18, 5, 0},
+   20, 25,   34, 24,  drops_0, 3, 4,   5, oflow_emf, back_emf, 16, 5, 0},
 
    {PITCHfall, 38, 2,  PITCHfall, 42, 30,              // statement, emphatic
    20, 25,   34, 22,  drops_0, 3, 3,   5, oflow, NULL, 15, 5, 0},
@@ -376,7 +376,7 @@ static void count_pitch_vowels(int start, int end, int clause_end)
 
 	if(no_tonic)
 	{
-		tone_posn = tone_posn2 = end-1;
+		tone_posn = tone_posn2 = end;  // next position after the end of the truncated clause
 	}
 	else
 	if(last_primary >= 0)
@@ -880,6 +880,10 @@ void Translator::CalcPitches(int clause_type)
 	int count_primary;
 	int ph_end=n_phoneme_list;
 
+	SYLLABLE syllable_tab2[N_PHONEME_LIST];
+
+	syllable_tab = syllable_tab2;   // don't use permanent storage. it's only needed during the call of CalcPitches()
+
 	if(langopts.intonation == 1)
 	{
 		CalcPitches_Tone(clause_type);
@@ -892,7 +896,7 @@ void Translator::CalcPitches(int clause_type)
 		option = 0;
 
 	group_tone_emph = group_tone = punct_to_tone[option][clause_type]; 
-	group_tone_emph = punct_to_tone[option][4];   // emphatic form of statement
+	group_tone_emph = punct_to_tone[option][5];   // emphatic form of statement
 
 	if(clause_type == 4)
 		no_tonic = 1;       /* incomplete clause, used for abbreviations such as Mr. Dr. Mrs. */
@@ -907,6 +911,7 @@ void Translator::CalcPitches(int clause_type)
 		if(p->synthflags & SFLAG_SYLLABLE)
 		{
 			syllable_tab[n_st].flags = 0;
+			syllable_tab[n_st].env = PITCHfall;
 			syllable_tab[n_st++].stress = p->tone;  // stress level
 
 			if(p->tone >= 4)

src/readclause.cpp

@@ -1967,22 +1967,15 @@ if(option_ssml) parag=1;
 		}
 		ix += utf8_out(c1,&buf[ix]);    //	buf[ix++] = c1;
 
-		if(!iswalnum(c1) && (ix > (n_buf-20)))
+		if(((ix > (n_buf-20)) && !IsAlpha(c1) && !iswdigit(c1))  ||  (ix >= (n_buf-2)))
 		{
 			// clause too long, getting near end of buffer, so break here
+			// try to break at a word boundary (unless we actually reach the end of buffer).
 			buf[ix] = ' ';
 			buf[ix+1] = 0;
 			UngetC(c2);
 			return(CLAUSE_NONE);
 		}
-		if(ix >= (n_buf-2))
-		{
-			// reached end of buffer, must break now
-			buf[n_buf-2] = ' ';
-			buf[n_buf-1] = 0;
-			UngetC(c2);
-			return(CLAUSE_NONE);
-		}
 	}
 	buf[ix] = ' ';
 	buf[ix+1] = 0;

src/synthdata.cpp

@@ -35,7 +35,7 @@
 #include "translate.h"
 #include "wave.h"
 
-const char *version_string = "1.29.26  18.Dec.07";
+const char *version_string = "1.29.29  20.Dec.07";
 const int version_phdata  = 0x012924;
 
 int option_device_number = -1;

src/synthesize.cpp

@@ -33,9 +33,6 @@
 #include "voice.h"
 #include "translate.h"
 
-#define    PITCHfall   0
-#define    PITCHrise   1
-
 
 extern FILE *f_log;
 static void SmoothSpect(void);

src/synthesize.h

@@ -18,7 +18,7 @@
  ***************************************************************************/
 
 
-#define N_PHONEME_LIST  700    // enough for source[] full of text, else it will truncate
+#define N_PHONEME_LIST  1000    // enough for source[N_TR_SOURCE] full of text, else it will truncate
 
 #define MAX_HARMONIC  400           // 400 * 50Hz = 20 kHz, more than enough
 #define N_SEQ_FRAMES   25           // max frames in a spectrum sequence (real max is ablut 8)

src/tr_english.cpp

@@ -43,6 +43,7 @@ Translator_English::Translator_English() : Translator()
 	langopts.stress_rule = 0;
 
 	langopts.numbers = 0x41 + NUM_ROMAN;
+	langopts.param[LOPT_COMBINE_WORDS] = 5;       // allow "mc" to cmbine with the following word
 }
 
 
@@ -52,7 +53,7 @@ static unsigned char initials_bitmap[86] = {
  0x20, 0x24, 0x20, 0x80, 0x10, 0x00, 0x00, 0x00,
  0x00, 0x28, 0x08, 0x00, 0x88, 0x22, 0x04, 0x00,  // 16
  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x88, 0x22, 0x04, 0x00, 0x02, 0x00, 0x00,  // 32
+ 0x00, 0x88, 0x22, 0x04, 0x00, 0x02, 0x00, 0x04,  // 32
  0x20, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
  0x00, 0x28, 0x8a, 0x03, 0x00, 0x00, 0x40, 0x00,  // 48
  0x02, 0x00, 0x41, 0xca, 0x9b, 0x06, 0x20, 0x80,

src/tr_languages.cpp

@@ -45,6 +45,7 @@
 #define OFFSET_GREEK  0x380
 #define OFFSET_CYRILLIC 0x420
 #define OFFSET_DEVANAGARI  0x900
+#define OFFSET_TAMIL  0xb80
 
 
 static const unsigned int replace_cyrillic_latin[] = 
@@ -292,7 +293,7 @@ Translator *SelectTranslator(const char *name)
 			tr->charset_a0 = charsets[19];   // ISCII
 			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
+			tr->langopts.stress_rule = 6;      // stress on last heaviest syllable, excluding final syllable
 			tr->langopts.stress_flags =  0x10004;   // use 'diminished' for unstressed final syllable
 			tr->langopts.numbers = 0x811;
 			tr->langopts.numbers2 = 0x100;
@@ -314,9 +315,14 @@ Translator *SelectTranslator(const char *name)
 		{
 			static const unsigned char stress_amps_hr[8] = {16,16, 20,20, 20,24, 24,22 };
 			static const short stress_lengths_hr[8] = {180,160, 200,200, 0,0, 220,230};
+			static const short stress_lengths_sr[8] = {160,150, 200,200, 0,0, 250,260};
 
 			tr = new Translator();
-			SetupTranslator(tr,stress_lengths_hr,stress_amps_hr);
+
+			if(name2 == L('s','r'))
+				SetupTranslator(tr,stress_lengths_sr,stress_amps_hr);
+			else
+				SetupTranslator(tr,stress_lengths_hr,stress_amps_hr);
 			tr->charset_a0 = charsets[2];   // ISO-8859-2
 
 			tr->langopts.stress_rule = 0;
@@ -590,6 +596,30 @@ SetLengthMods(tr,3);  // all equal
 		}
 		break;
 
+	case L('t','a'):
+		{
+			static const short stress_lengths_ta[8] = {190, 190,  210, 210,  0, 0,  230, 250};
+			static const unsigned char stress_amps_ta[8] = {17,14, 20,19, 20,24, 24,22 };
+
+			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_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));
+			SetLetterBitsRange(tr,LETTERGP_A,0x05,0x14);   // vowel letters
+			SetLetterBitsRange(tr,LETTERGP_A,0x3e,0x4c);   // vowel signs
+			SetLetterBitsRange(tr,LETTERGP_B,0x3e,0x4d);   // vowel signs, and virama
+			SetLetterBitsRange(tr,LETTERGP_C,0x15,0x39);   // the main consonant range
+			tr->langopts.param[LOPT_UNPRONOUNCABLE] = 1;   // disable check for unpronouncable words
+		}
+		break;
+
 	case L('t','r'):   // Turkish
 		{
 			static const unsigned char stress_amps_tr[8] = {16,16, 20,20, 20,24, 24,22 };

src/translate.cpp

@@ -88,8 +88,8 @@ static int embedded_ix;
 static int embedded_read;
 unsigned int embedded_list[N_EMBEDDED_LIST];
 
-// the source text of a single clause
+// the source text of a single clause (UTF8 bytes)
-#define N_TR_SOURCE    350
+#define N_TR_SOURCE    700
 static char source[N_TR_SOURCE+40];     // extra space for embedded command & voice change info at end
 
 int n_replace_phonemes;
@@ -339,12 +339,12 @@ unsigned char *length_mod_tabs[6] = {
  };
 
 
-/* index by 0=. 1=, 2=?, 3=! 4=none */
+/* index by 0=. 1=, 2=?, 3=! 4=none, 5=emphasized */
-static unsigned char punctuation_to_tone[4][5] = {
+static unsigned char punctuation_to_tone[4][6] = {
-	{0,1,2,3,4},
+	{0,1,2,3,0,4},
-	{5,6,2,0,4},
+	{5,6,2,0,0,4},
-	{0,1,2,3,0},
+	{0,1,2,3,0,0},
-	{0,1,2,3,0} };
+	{0,1,2,3,0,0} };
 
 
 void SetLengthMods(Translator *tr, int value)
@@ -369,6 +369,9 @@ int IsAlpha(unsigned int c)
 	if((c >= 0x901) && (c <= 0x957))
 		return(1);    // Devanagari  vowel signs and other signs
 
+	if((c >= 0xb81) && (c <= 0xbe5))
+		return(1);    // Devanagari  vowel signs and other signs
+
 	if((c >= 0x300) && (c <= 0x36f))
 		return(1);   // combining accents
 

src/translate.h

@@ -25,7 +25,7 @@
 
 #define N_WORD_PHONEMES  160          // max phonemes in a word
 #define N_WORD_BYTES     160          // max bytes for the UTF8 characters in a word
-#define N_CLAUSE_WORDS   256          // max words in a clause
+#define N_CLAUSE_WORDS   300          // max words in a clause
 #define N_RULE_GROUP2    120          // max num of two-letter rule chains
 #define N_HASH_DICT     1024
 #define N_CHARSETS        20