Fix bug, use of the command-line "espeakedit --compile" produced phoneme data
in which vowels started at full amplitude instead of gradual increase.
Fix bug: SSML <break time=""> with values greater than 23 seconds were ignored
due to arithmetic overflow.
Phoneme definition files:
Added new function: ChangeNextPhoneme().
Language changes include: da, en, fa, gu, hu, it, nl, pt, ta
Started new language: eu (Basque).
*_rules files:
Allow $p_alt1, $p_alt2, etc in pre-conditions. The rule matches if the
characters before the current point in the word occur as an entry in the
*_list file with the corresponding $alt attibute.
Added condition $list. This is similar to $p_alt etc, but examines
all words in the *_list file which don't have the $only attribute.
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@374 d46cf337-b52f-0410-862d-fd96e6ae7743
Added rule attribute $noprefix for *_rules files. The rule is not used if a prefix has been removed from the word.
Language changes: en, or, pt, ta.
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@369 d46cf337-b52f-0410-862d-fd96e6ae7743
Language changes to: an, en, gu, it, pt, ta.
Superscripts and subscripts are spoken as single letters and numbers during njormal speech, and with "superscript" and "subscript" spoken when spelling.
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@367 d46cf337-b52f-0410-862d-fd96e6ae7743
Language changes:
an, en, es, fa, gu, it, pt, ta, te.
Voice files: added attribute 'dictdialect', eg:
dictdialect en-us
uses the US English voice to speak words or rules which are marked with _^_EN
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@366 d46cf337-b52f-0410-862d-fd96e6ae7743
Fix: Flush stderr() after espeak_CompileDictionary().
Fix: parallel formant parameters for Klatt synthesizer.
Changes for Malayalam numbers.
Language changes: en-us, nl, pt, sl, ta, ur.
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@359 d46cf337-b52f-0410-862d-fd96e6ae7743
Reinstate language "az" (Azerbaijani) as a test voice.
Language changes: ga, hy, pt, sk.
Generate character names for subscript and superscript letters and digits.
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@347 d46cf337-b52f-0410-862d-fd96e6ae7743
lang=zh: Fix: was speaking Chinese character names as character codes.
lang=hr, sr: Fix: speak Cyrillic character names in the local langauge.
Phoneme definitions: add new statement type "LengthAdd <signed number>".
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@335 d46cf337-b52f-0410-862d-fd96e6ae7743
Fix bug in SAPI5 version: speed suddenly changes to very slow when speaking a word which contains foreign characters.
Language changes: da, fa, hu, pt, ur.
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@313 d46cf337-b52f-0410-862d-fd96e6ae7743
Command line option --ipa can have option to include Tie or ZWJ characters to disabmiguate multi-letter phoneme names.
Recognize and ignore UTF8-BOM byte sequence at the start of phoneme definition files.
Languages changes: ta, nl, pt.
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@306 d46cf337-b52f-0410-862d-fd96e6ae7743
Fixes to speaking characters from foreign alphabets.
Allow $textmode and $text to translate between words in different alphabets.
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@305 d46cf337-b52f-0410-862d-fd96e6ae7743
Ignore character U+00AD (soft hyphen) in text.
Added stress rule code for Lang=kl (Greenlandic).
Changes to languages: it, ta, si
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@294 d46cf337-b52f-0410-862d-fd96e6ae7743
Added language "lt" Lithuanian.
Changes to languages: pt, ta, ka.
Allow ZLNJ character (U+200C) to appear as an alphabetic character (lang=fa).
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@285 d46cf337-b52f-0410-862d-fd96e6ae7743
Speak names of characters U+2800 to U+28FF, Braille dot symbols.
Language option to speak numbers which use characters '0' to '9' with the English voice.
Add some more characters to the vowel and consonant lists for Indian languages.
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@284 d46cf337-b52f-0410-862d-fd96e6ae7743
Add --version command-line option.
Allow hexadecimal numbers for UTF-8 bytes in rules in *_rules files.
wave_pulse.cpp, ensure wave_close is not called again while waiting.
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@278 d46cf337-b52f-0410-862d-fd96e6ae7743
Fixed crash on re-initializing libespeak.
lang=hu, experimental, if the last two words of a clause (before punctuation) both have capital letters, then make the last word unstressed. This is useful for people names, but may be wrong for other cases.
lang=en-us, distinguish phonemes [aI@] and [aI3] (compare "diet", "tired").
lang=de, fix automatic devoicing of word-final 'b', 'd', 'g'.
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@276 d46cf337-b52f-0410-862d-fd96e6ae7743
Fixed bug where SSML voice change didn't remember an initially specified Klatt variant.
Removed some unused variables and redundant initializations.
Patch to fix a spike from Klatt function setzeroabc()
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@275 d46cf337-b52f-0410-862d-fd96e6ae7743
Added runtime detection of pulseaudio, with fallback to portaudio if pulse is not running. Use build option "AUDIO=runtime" in the makefile.
Phoneme programs. Added condition "isTranslationGiven" to test whether a phoneme has been specified explicitly for a word in *_list.
*_rules files. Suffix rules are now considered only if there is a previous vowel in the word.
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@274 d46cf337-b52f-0410-862d-fd96e6ae7743
Keep the original voice variant across SSML xml:lang changes.
Use the original language sub-type (eg. en-sc) when SSML xml:lang changes to a compatible language.
SSML <emphasis>, increase volume for "strong" and "x-strong".
Fix bug where </emphasis> was ignored at end of clause.
Fix: $dot attribute was ignored for multi-word entries in *_list files.
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@269 d46cf337-b52f-0410-862d-fd96e6ae7743
Amplitude changes now apply also to mbrola voices.
"voicing" attribute is now used in mbrola voice files to set the voice loudness.
SSML commands such as <mark> and <prosody> were ignored if they were inside a multiword group (defined in *_list). This is now fixed so that the command is processed at the end of the group.
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@267 d46cf337-b52f-0410-862d-fd96e6ae7743
-x command-line option. Fixed possible buffer overflow when writing phoneme output.
Slavic languages, changed behaviour of [v] during voiced/unvoiced propagation.
When looking for a word boundary when breaking long clauses, allow for greater word size (Southern Indian languages have big words 3-byte UTF-8 character codes).
$combine. Don't combine with the nerxt word if it is translated using a different language.
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@262 d46cf337-b52f-0410-862d-fd96e6ae7743
lang=Turkish, fix pronunciation of suffixes, fix suffixes after names.
Windows: command option --stdout caused spurious 0x0d bytes to be included in speech output (CR/LF problem).
Fixed crash which could occur when a clause starts with a hyphen.
espeakedit: New input format for Bulgarian lexicon.
lang=de: Fix important pronunciation error (editing error) for "sage".
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@259 d46cf337-b52f-0410-862d-fd96e6ae7743
Phoneme definitions.
New keyword "prevoice", always add murmer before a voiced stop, not only after a vowel.
Fixed bug in nextVowel() condition.
Fixed bug is condition "isRhotic".
Command line options.
On Windows, -- means don't treat following arguments as options.
--help output now should the espeak-data path.
Tools->Lexicon. Redesign Lexicon processing for Russian and Bulgarian.
Numbers. Fixes for Hungarian ordinal numbers.
Fixed bug where double .. was ignored.
Fixed bug in reporting word event lengths in callbacks.
Fixed bug in reporting marker event positions in callbacks, when non-ascii text.
SSML.
Allow single as well as double quotes around strings.
Fix bugs in SSML "prosody rate".
Speed.
Only use Sonic speedups greater than x2.
Use custom eSpeak speed settings when Sonic speedup is used.
Mbrola. Look for Mbrola voice also at /usr/share/mbrola/xx/xx
espeakedit. Try "aplay" and then "play" to play sound.
Languages:
Kannada (kn) now in official release.
Esperanto (eo) change prosody.
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@257 d46cf337-b52f-0410-862d-fd96e6ae7743
Change in format of *_dict files. 'Word-boundary' at the start of a rule from *_rules is not implied by a new rule type RULE_PRE_ATSTART.
espeakedit: Improvements to Tools->Process Lexicon->Italian which makes the file dictsource/it_listx.
lang=hu, fixes for ordinal numbers.
Speed-up, add files sonic.cpp, sonic.h
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@255 d46cf337-b52f-0410-862d-fd96e6ae7743
Speed increase. Added the "sonic" routines by Bill Cox to reduce the length of eSpeak's sound output data, so as to make the speaking speed faster.
Ordinal numbers: Some languages (el,es) use the ordinal form of hundreds and tens part of a number, as well as the units part. Added language option NUM2_MULTIPLE_ORDINAL.
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@254 d46cf337-b52f-0410-862d-fd96e6ae7743
Fix: crash with very long numbers.
Very long nubers ( > 14 digits) now spoken as individual digits.
Fix: bracket was interfering with recognizing last-word in a clause.
Lang=el fix: all words were spoken as individual letters.
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@252 d46cf337-b52f-0410-862d-fd96e6ae7743
Phoneme definitions. Add "equivalents" table to translate phonemes from a foreign language (eg. English) into the equivalant phonemes of this language. This is used when *_rules and *_list files indicate that a word uses pronunciation rules from a different language.
Unpronouncable words can be recognised by rules in *_rule files. This feature is now used for en, de, es.
Lang=ta, hi. Letter-names for combining vowel characters are distinguished from stand-alone vowel characters by a preceding beep.
git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@251 d46cf337-b52f-0410-862d-fd96e6ae7743
jonsd