RealSpeak Telecom Software Development Kit

Transcription

1 RealSpeak Telecom Software Development Kit User s Guide for German V4.0

2 EV AL U AT I O N AG R E E M E N T READ THE FOLLOWING CAREFULLY BEFORE USING THE SOFTWARE. PRIOR TO RECEIVING THE SOFTWARE, YOU HAVE SIGNED EITHER A LICENSE AGREEMENT OR A NON-DISCLOSURE AGREEMENT WITH SCANSOFT, CONCERNING THE REALSPEAK SOLO DEVELOPMENT SYSTEM OR SOME OF ITS COMPONENTS. THEREFORE, REFERENCE IS MADE TO THESE AND AS SUCH THE CONTRACTUAL TERMS AND CONDITIONS SHALL APPLY TO THE SOFTWARE. IF YOU HAVE NOT SIGNED SUCH AN AGREEMENT AND IF YOU USE THE SOFTWARE, SCANSOFT WILL ASSUME THAT YOU AGREED TO BE BOUND BY THE EVALUATION AGREEMENT SPECIFIED HEREUNDER. IF YOU DO NOT ACCEPT THE TERMS OF THIS EVALUATION AGREEMENT, YOU MUST RETURN THE PACKAGE UNUSED TO SCANSOFT WITHIN SEVEN (7) DAYS AFTER RECEIPT. 1. Grant of Rights In consideration of a possible commercial relationship, ScanSoft hereby grants to you, the LICENSEE, who accepts, a non-exclusive right to internaly evaluate and test the software program ( the Software ). 2. Ownership of Software ScanSoft retains title, interests and ownership of the Software recorded on the original disk(s) and all subsequent copies of the Software and Documentation, regardless of the form or media in or on which the original and other copies may exist. ScanSoft reserves all rights not expressly granted to LICENSEE. 3. Copy Restrictions This Software and the accompanying documentation are copyrighted. Unauthorized copying of the Software, including Software that has been merged or included with other software, or of the documentation is expressly forbidden. LICENSEE may be held legally responsible for any intellectual property infringement that is caused or encouraged by his failure to abide by the terms of this agreement. LICENSEE is allowed to make two (2) copies of the Software solely for backup purposes, provided that the copyright notice is included on the backup copy. 4. Use Restrictions LICENSEE agrees not to use the Software for any other purpose than internally evaluating the Software. LICENSEE may physically transfer the Software from one computer to another, provided that the Software is used on only one computer at a time. LICENSEE may not modify, adapt, translate, reverse engineer, decompile, disassemble or create derivative works based on the Software.LICENSEE may not modify, adapt, translate or create derivative works based on the documentation provided by ScanSoft. The Software may not be transferred to anyone without the prior written consent of ScanSoft. In no event may LICENSEE transfer, assign, lease, sell or otherwise dispose of the Software and Documentation on a temporary or permanent basis except as expressly provided herein. 5. Warranty THE SOFTWARE IS PROVIDED AS IS WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. ScanSoft shall have no liability to LICENSEE or any third party for any claim, loss or damage of any kind, including but not limited to lost profits, punitive, incidental, consequential or special damages, arising out of or in connection with the use or performance of the Software and accompanying documentation. 6. Termination This agreement is effective until terminated. ScanSoft reserves the right to terminate this agreement automatically if any provision of this agreement is violated. LICENSEE may terminate this agreement by returning the Software and the accompanying documentation to ScanSoft, along with a written warranty stating that all copies have been returned.

3 Copyright (C) ScanSoft, Inc RealSpeak Telecom Software Development Kit User's Guide for German V4.0 December 2005-December 2005 No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording or any information retrieval system, without the written permission of ScanSoft. Trademarks MS-DOS, WINDOWS, MICROSOFT VISUAL C++, BORLAND C++ and Sound Blaster are registered trademarks of their respective owners. ScanSoft is a registered trademark. All rights reserved.

4 Document History Date Type of change SoftwareVersion July 2004 Reformatting of document V4.0 Update for version 4.0 November 2004 Update for version 4.0 V4.0 December 2005 Update SSML Preprocesor chapter and Using Control Sequences section of Chapter I V4.0

5 Table of Contents GERMAN TEXT-TO-SPEECH SYSTEM... 8 Introduction...8 Preparing a text for Text-To-Speech...8 Native Character Set...8 Using Control Sequences...8 Quick Reference of the RealSpeak native Control Sequences for German...10 Entering phonetic input...15 How to proceed...15 Lexical stress and sentence accents in phonetic input...16 The German L&H+ & UNIPA Phonetic Alphabet...18 Using a User Dictionary...20 Using the Microsoft SAPI5 Lexicon...20 User Lexicons...20 Application Lexicons...20 The German SAPI5 Phoneme List...20 Notes on the German Text-To-Speech System...24 Cardinal Numbers...24 Decimal Numbers...24 Ordinal Numbers...25 Roman Numbers...25 Telephone Numbers...25 Bank Account Numbers...25 Dates...26 Time Indications...26 Currencies...26 Abbreviations and Acronyms PREPROCESSOR Introduction Header Processing...30 Header Field Extraction...30 Header Field Reading...32 From Field...32 Date Field...33 Subject Field body processing...35 Message Extraction...35 Text Normalization...36 Language Specific Text Normalization...39 Umlaut...39 English Words...39 Customizing the Preprocessor...40 Customizing the Header...40 From Field...40 Subject Field...42 Customizing the Body...43

6 SSML PREPROCESSOR Introduction...45 German specific SSML markup...45 XML encoding types for German SML Specifics for German...46 CUSTOM G2P DICTIONARIES Introduction...51 Proper names Custom G2P dictionary...51 APPENDICES Appendix A: German voice names...53

7 RealSpeak Telecom SDK Chapter I German Text-To-Speech System User s Guide for German V4.0

8 Chapter I German Text-to-Speech system German Text-To-Speech System Introduction This section provides operational instructions for the ScanSoft Text-To-Speech system for German. It reviews the functionality of the system, and describes how the user can customize the pronunciation of input texts. This part also describes issues that are particular to the German Text-To-Speech system. It introduces the German phonetic alphabet and discusses a number of the languagespecific features of the German Text-To-Speech system. Preparing a text for Text-To-Speech Native Character Set Using Control Sequences In general, there are four ways to intervene in the pronunciation of text: By using control sequences By entering phonetic input By using a user dictionary or a user ruleset By using one of the supported API s These mechanisms are described in the Programmer s Guide. In this part, however, the specifications for German are fully described. The native character set of the German TTS system is Windows-1252 which has the printable characters in the ASCII range as a subset. Note that TTS input encoded in another supported character set is converted to the native character set for that language before it is processed internally. Consequently, input must be representable in the native character set even if it is encoded in another character set supported by the API. For a description of the various supported markup languages (independent from the language), refer to the Programmer's Guide. Remark: <ESC> represents the escape character \x1b (decimal 27) that generates the ASCII character 27 (Hex 1B). User sguide for German Chapter I/8

9 Chapter I German Text-to-Speech system Below, you find a quick reference table for the RealSpeak native control sequences for German. The language-specific support for the SSML markup language is described in the SSML Preprocesor chapter. User sguide for German Chapter I/9

10 German Text-to-Speech system Chapter I Quick Reference of the RealSpeak native Control Sequences for German Sequence Description Range Default Delimiter <ESC> \vol=x\ <ESC> \rate=x\ <ESC> \rate_wpm= xxx\ <ESC>Mx <ESC>Wx Volume (x : ) 0 = silence 10 = low 100 = high 80 No <ESC>\vol=10\ Ich kann sehr leise sprechen, <ESC>\vol=90\ aber auch sehr laut. Speech Rate (x : ) 10 = slow 100 = fast 50 No Ich kann <ESC>\rate=70\ sehr schnell sprechen <ESC>\rate=20\ oder aber ganz langsam. Word per minute (xxx: ) Voice-specific (see subsequent table) Voicespecific No Ich kann <ESC>\rate_wpm=350\ sehr schnell sprechen <ESC>\rate_wpm=110\ oder aber ganz langsam. Read mode; some read modes are not supported in e- mail mode x = 0..3: 0 = character-bycharacter 1 = word-by-word (not supported in e- mail mode) 2 = sentence-bysentence 3 = line-by-line (not supported in mode) <ESC>M0 Demonstration (The word "Demonstration" will be spelled.) <ESC>M1 Dies ist eine Demonstration. (This sentence will be read word by word.) <ESC>M2 Dies ist eine Demonstration. (This sentence will be read as one sentence.) Wait Period Sie hören jetzt <ESC>W5 eine kurze Pause. x= = no wait period 1 = 200 ms 9 = 1800 ms 2 Yes 2 yes User sguide for German Chapter I/10

11 Chapter I German Text-to-Speech system Sequence Description Range Default Delimiter <ESC> \Pause=xxx \ <ESC>" <ESC>C Long Pause Man kann die Dauer der Pause ganz genau <ESC>\pause =1500\ definieren msec No Sentence Accent <ESC>"Jakob hat gestern angerufen. (nicht Peter) Jakob hat <ESC>"gestern angerufen. (nicht heute) No Note: Manually inserted sentence accents may have no effect in RealSpeak. The RealSpeak synthesis module may indeed have reasons to override the requested sentence accent, and thus not realize it. Continuation Der 50. Besucher. Der 50. <ESC>C Besucher. No In the first of the above examples, the text-to-speech system will detect an end-of-sentence after 50 and will inappropriately split the input into two separate sentences. In the second example, a continuation sequence is inserted in order to make the system pronounce the entire input as one sentence. <ESC>E <ESC>/+ End-of-Message Yes Dies ist der erste Satz <ESC>E und dies ist der zweite. In the above example, the sequence <ESC>E forces the system to pronounce the two halves of the input separately. Phonetic Input No (L&H+ phonetic alphabet) <ESC>/+ Stat<ESC>/+ User sguide for German Chapter I/11

12 Chapter I German Text-to-Speech system Sequence Description Range Default Delimiter <ESC>%x <ESC> \tn=s\ <ESC>F <ESC>@c <ESC>\do main=s\ <ESC>\do main\ Preprocessing Mode text = standard text mode = mode Yes <ESC>%text Sie hören eine Nachricht von Hans Kessler. <ESC>% Von: kessler@alpha.ee.upenn.edu (Hans Kessler) Guide text normalization; limited support in mode s = string: address=address mode (not supported in mode) normal=standard mode spell=spell mode Normal The text normalization types corresponding with the SSML <say-as> types are also supported in standard text mode (not in mode), see the SSML Preprocesor chapter for more details. <ESC>\tn=address\ Prof. Dr. Meier, GH Bonn, Gebäude III, 2. St., PF 360, Bonn <ESC>\tn=normal\ Dr. GH Meier No Reset to Default Yes <ESC>\vol=10\ Dies ist der niedrigste Wert der Lautstärke. <ESC>F Und dies ist der normale Wert. <ESC>\rate=90\ Dies ist der höchste Wert der Geschwindigkeit. <ESC>F Und dies ist der normale Wert. Declare the c = character No part-of-speech (not supported in German) Enable the s = string: the name Yes extension of the extension (only if a custom g2p has been loaded) Disable the last Yes extension User sguide for German Chapter I/12

13 Chapter I German Text-to-Speech system Sequence Description Range Default Delimiter <ESC>\voic e=s\ <ESC>\mrk =n\ Set the voice (if there is more than 1 voice is available) Insert a bookmark s = string: the name of the voice n = Yes No <ESC>\p\ Hallo, ich bin \mrk=2000\ Steffi Insert a paragraph boundary Yes <ESC>\aud io="s"\ Herr Krueger \p\ Kirchstrasse 12 \p\ Berlin Insert an audio file; not supported in e- mail mode s = string: the URI of a document with an appropriate MIME type Yes User sguide for German Chapter I/13

14 German Text-to-Speech system Chapter I NOTE The speech rate is language, gender and technology dependent. It can be set in 9 discrete steps. The values given here are in words per minute. Rate level Rate (wpm) Steffi RealSpeak User sguide for German Chapter I/14

15 German Text-to-Speech system Chapter I Entering phonetic input How to proceed To switch from orthographic to phonetic mode, insert <ESC>/+ to use the L&H+ phonetic alphabet. The phonetic input mode remains active until the command is explicitly reset by entering <ESC>/+ again. The phonetic input string is composed of symbols of the L&H+ phonetic alphabet (see phonetic table below). Examples are given below in the phonetic table. In addition to the phonetic symbols, it is advised to use the following characters in the phonetic input string: L&H + Symbol ' (ASCII 39, Hex 27) Meaning primary word stress Special characters As in: <ESC>/+ mo:.'dern <ESC>/+ (Adjektiv 'modern') vs. <ESC>/+ 'mo:.d$rn <ESC>/+ (Verb 'modern') '2 secondary word stress <ESC>/+ 'te.nis.'2spi:.l$r <ESC>/+ (Tennisspieler) " (ASCII 34, Hex 22) sentence accent <ESC>/+Es_gIbt_"tsva&i_?ak."t&s En.t$_?In_'di:.z$m_ 'zats <ESC>/+ (Es gibt ZWEI AKZENTE in diesem Satz.). syllable boundary <ESC>/+ 'bu:x.sta:.b$ <ESC>/+ (Buchstabe) # silence (pause) <ESC>/+?ER_"fRa:k.t$_#_vi:_"ge:t_?Es <ESC>/+ (Er fragte: wie geht es?) Note that the use of punctuation marks remains useful within phonetic input to assure a correct intonation. Each punctuation mark needs to be preceded by an asterisk. User sguide for German Chapter I/15

16 Chapter I German Text-to-Speech system <ESC>/+ "vil.ko.m$n*,_"ko.m$n_zi:_he. "Ra&in*. <ESC>/+ (Willkommen, kommen Sie herein.) Punctuation Marks L&H+ Symbol Meaning - Word delimiter *. End of declarative *, Comma *! End of exclamation *? End of question *; Semicolon *: Colon Lexical stress and sentence accents in phonetic input In phonetic input strings, lexical stress and sentence accents can be indicated manualy by the user, by using a single quote ( ) or double quote ( ) respectively. Note that manually inserted lexical stress or sentence accents may have no effect in RealSpeak. The RealSpeak synthesis module may indeed have reasons to override the requested stress/accent and thus not realize it. 1. The Text-To-Speech system will automatically convert all lexical stress marks into sentence accents in case no manually added sentence accents are found in the phonetic input string. Example: <ESC>/+IC_'ha:.b$_zi:_'hO&y.t$_nOx_g$.'SpRO.x$n*.<ES C>/+ is the same as <ESC>/+IC_"ha:.b$_zi:_"hO&y.t$_nOx_g$."SpRO.x$n*. <ESC>/+ (Ich habe sie heute noch gesprochen.) 2. If phonetic input contains at least one manually added sentence accent, no additional sentence accents are assigned by the Text-To-Speech system. Therefore, only those words marked with " will get a sentence accent. As a consequence, a message containing only one manual sentence accent will have an almost flat intonation on all the other words. User sguide for German Chapter I/16

17 Chapter I German Text-to-Speech system Example: <ESC>/+IC_'ha:.b$_zi:_"hO&y.t$_nOx_g$.'SpRO.x$n*. <ESC>/+ (Only one sentence accent will be realized.) 3. Phonetic input can also be combined with orthographic input. If no sentence accents are found in the input text (indicated by <ESC>" in orthographic input, or by " in phonetic input), the Text-To-Speech system will automatically assign sentence accents. In the orthographic part of the input, the Text-To-Speech system will realize these sentence accents on the basis of part-of-speech and syntactic information. In the phonetic part of the input, all lexical stress marks (if any) will be converted into sentence accents. If there are no lexical stress marks, no sentence accent will be realized for the phonetic part of the input (see point 1 above). If the user has manually specified one or more sentence accents, no additional sentence accents will be realized (see point 2 above). Er hat heute noch mit <ESC>/+'klIn.t$n <ESC>/+ gesprochen. (No sentence accents are found; the Text-To-Speech system will automatically assign sentence accents.) Er hat heute noch mit <ESC>/+"klIn.t$n <ESC>/+ gesprochen. (A sentence accent is specified in the phonetic part of the input text. No additional sentence accents will be realized.) Er hat <ESC>"heute noch mit <ESC>/ +'klin.t$n <ESC>/+ gesprochen. (A sentence accent is specified in the orthographic part of the input text. No additional sentence accents will be realized. Hence, the lexical stress that is specified in the phonetic part will NOT be converted into a sentence accent.) Er hat <ESC>"heute noch mit <ESC>/+"klIn.t$n <ESC>/+ gesprochen. (Two sentence accents were specified; no additional sentence accents will be realized.) User sguide for German Chapter I/17

18 German Text-to-Speech system Chapter I The German L&H+ & UNIPA Phonetic Alphabet L&H+ Symbol L&H+ Transcription Vowels and Diphthongs UNIPA Symbol UNIPA Transcription As in: a Stat a Stat Stadt a: va:.g$n a: va:.g$n Wagen E lets.t$ E lets.t$ Letzte e: ke:.l$ e: ke:.l$ Kehle I milc I milc Milch i: Ri:.z$ i: Ri:.z$ Riese O fol O fol voll o: gro:s o: gro:s groß U kunst U kunst Kunst u: fu:s u: fu:s Fuß Y kys.t$ Y kys.t$ Küste y: gry:n y: gry:n grün E+ le+.s$n E= le=.s$n löschen e+ Se+n e= Se=n schön E: 'fe:.r$ E: 'fe:.r$ Fähre $ tas.t$ $ tas.t$ Taste a&u 'ba&um a+u 'ba+um Baum O&y 'ho&y.t$ O+y 'ho+y.t$ heute a&i 'ta&il a+i 'ta+il Teil A%~ REs.to:.'RA%~ A%~ REs.to:.'RA%~ Restaurant O%~ bal.'ko%~ O%~ bal.'ko%~ Balkon E%~ 'te%~ E%~ 'te%~ Teint E+%~ par.'fe+%~ E=%~ par.'fe=%~ Parfum User sguide for German Chapter I/18

19 German Text-to-Speech system Chapter I L&H+ Symbol L&H+ Transcription Consonants UNIPA Symbol UNIPA Transcription As in: p 'post p 'post Post b 'ba&in b 'ba+in Bein t 'tin.t$ t 'tin.t$ Tinte d 'dic d 'dic dich k 'kla&in k 'kla+in klein g 'li:.g$n g 'li:.g$n liegen f 'felt f 'felt Feld v 'vax v 'vax wach s 'fels s 'fels Fels S 'Sne: S 'Sne: Schnee z 'za:l z 'za:l Saal Z ZUR. 'na:l Z ZUR. 'na:l Journal C 'milc C 'milc Milch x 'bax x 'bax Bach h 'hant h 'hant Hand j 'je:.mant j 'je:.mant jemand l 'lict l 'lict Licht R 'Ra&i.z$ R 'Ra+i.z$ Reise m 'man m 'man Mann n 'nor.d$n n 'nor.d$n Norden nk 'RInK nk 'RInK Ring? b$.'?ax.t$n? b$.'?ax.t$n beachten t&s 't&su:k t+s 't+su:k Zug p&f 'p&fe:rt p+f 'p+fe:rt Pferd t&s 't&sil.p$n t+s 't+sil.p$n tschilpen d&z 'd&zes d+z 'd+zes Jazz User sguide for German Chapter I/19

20 Chapter I German Text-to-Speech system NOTE Using a User Dictionary Note that the L&H+alphabet is not SSML compliant. For SSML, use the UNIPA alphabet. For information on how to create and use user dictionaries, please refer tothe User Configuration chapterin the RealSpeak Telecom Programmer s Guide. Using the Microsoft SAPI5 Lexicon Microsoft SAPI5 provides lexicons so that users and applications can specify pronunciation and part of speech information for particular words. As such, all SAPI compliant Text-To-Speech engines should use these lexicons to guarantee uniformity of pronunciation and part of speech information. There are two types of lexicons in SAPI: user lexicons and application lexicons. User Lexicons Each user who logs into a computer will have a User Lexicon. Initially, this lexicon is empty; words can be added either programmatically, or by using an engine's add/remove words UI component (for example, the sample application Dictation Pad provides an Add/Remove Words dialog). Application Lexicons Applications can create and ship their own lexicons of specialized words. These lexicons are fixed and cannot be edited. Detailed information on how to use the MS SAPI5 lexicons can be found in the manual Microsoft Speech SDK V5.1, chapter ISpLexicon Interface. The German SAPI5 Phoneme List To add entries to the lexicon, the user should use a set of language specific phonemes. The language specific phoneme list for German is given below. User sguide for German Chapter I/20

21 Chapter I German Text-to-Speech system NOTE Note that the Microsoft Speech SDK V5.1 only provides a phoneme set for American English. The German SAPI5 phoneme set mentioned below has been developed by ScanSoft, based on the symbols available for American English. The phoneme list below is therefore not to be considered as an official phoneme set defined by Microsoft SAPI5. SAPI5 Symbols SAPI Symbol PhoneID Example SAPI Transcription A 13 Satz Z A TS A : Tat T A : T AW 14 Haus H AW S AX 15 bitte B IH T AX AX 15 besser B EH S AX R AY 16 Eis AY S EH 20 Gesetz G AX Z EH TS EY 22 Beet, spät B EY T, S P EY T EH : (Buchstabe) Ä EH : OE 34 plötzlich P L OE TS L IH X EU 21 blöd B L EU T IH 26 Sitz Z IH TS IY 27 lieb L IY P OH 35 Trotz T R OH TS OW 36 Boot B OW T OY 37 Kreuz K R OY TS UY 48 hübsch H UY P SH UE 45 süss (süß) Z UE S UH 46 Schutz SH UH TS UW 47 Blut B L UW T B 17 Bein B AY N D 18 Deich, dank D AY X, D A NG K CH 19 deutsch D OY CH F 23 fast F A S T User sguide for German Chapter I/21

22 Chapter I German Text-to-Speech system SAPI5 Symbols SAPI Symbol PhoneID Example SAPI Transcription G 24 Gunst G UH N S T H 25 Hand H A N T JH 28 dschungel JH UH NG AX L K 29 Kunst K UH N S T L 30 Leim L AY M M 31 mein M AY N N 32 nein N AY N NG 33 Ding D IH NG P 38 Pein P AY N PF 39 Pfahl PF A : L R 40 Reim R AY M S 41 Tasse T A S AX SH 42 waschen V A SH AX N T 43 Teich T AY X TS 44 Zahl TS A : L V 49 was V A S X 50 sicher Z IH X AX R X 50 Buch B UW X Y 51 Jahr Y A : AX R Z 52 Hase H A : Z AX ZH 53 Genie ZH EY N IY IY AX R Tier T IY AX R IH AX R Wirt V IH AX R T UE AX R Tür T UE AX R UY AX R Türke T UY AX R K AX EY AX R schwer SH W EY AX R EH AX R Berg B EH AX R K EH : AX R Bär B EH : AX R EU AX R Föhr F EU AX R OE AX R Wörter V OE AX T AX R A : AX R Haar H A : AX R A AX R hart H A AX R T UW AX R Kur K UW AX R UH AX R kurz K UH AX R TS OW AX R Ohr OW AX R OH AX R dort D OH AX R T User sguide for German Chapter I/22

23 Chapter I German Text-to-Speech system SAPI5 Symbol SAPI Phone ID SAPI5 Symbols Meaning SAPI Transcription - 1 syllable boundary (hyphen) B IH - T AX! Sentence terminator (exclamation B IH T AX! 2 mark) & B IH T AX & B EH S 3 word boundary AX R, B IH T AX, _ B EH S 4 Sentence terminator (comma) AX R. 5 Sentence terminator (period) B IH T AX.? 6 Sentence terminator (question mark) AY S? _ B IH T AX, _ B EH S Silence (underscore) 7 AX R 1 9 primary stress 1 B IH - T AX 2 10 secondary stress ~ 11 nasalization : 12 lengthen ^ 8 Verein F EH AX R ^ AY N User sguide for German Chapter I/23

24 Chapter I German Text-to-Speech system Notes on the German Text-To-Speech System Cardinal Numbers Decimal Numbers The German Text-To-Speech system has been designed in order to pronounce correctly any input written according to the rules of German orthography. The following cases, however, require special attention. Cardinal numbers up to 15 digits are pronounced as full numbers. Periods may be used to separate groups of digits or Decimal numbers may consist of up to 15 digits before or after the comma. Periods may be used to separate groups of digits in the string before the comma. The digits after the comma are pronounced one by one. NOTE 9550, ,5 Numerals that are normally pronounced as full numbers, can also be pronounced digit by digit by using the control sequence <ESC>\tn=spell\ in front of the numeral. User sguide for German Chapter I/24

25 Chapter I German Text-to-Speech system Ordinal Numbers A cardinal number smaller than 32 followed by a period is pronounced as an ordinal number if it is not in a sentence initial or sentence final position. am 15. Mai 1998 Roman Numbers Roman numbers smaller than 10 that cannot be interpreted as single letters are pronounced as full numbers (i.e. II, III, IV, VI, VII, VIII, IX). Abteilung IV Garzweiler II Telephone Numbers In order to ensure a correct pronunciation of telephone numbers, it is recommended to use parentheses to separate country code and/or area code from the remainder of the telephone number. Also, use spaces to separate groups of digits. Telephone numbers written in this format will always be pronounced in groups of two or three digits, with a pause at the place of the space. (041) (03 35) Bank Account Numbers To have a bank account number correctly pronounced (in groups of 2 or 3 digits), use hyphens between groups of digits. To have the number pronounced digit by digit, switch to spell mode (<ESC>\tn=spell\). User sguide for German Chapter I/25

26 Chapter I German Text-to-Speech system LG Stuttgart BLZ Dates Dates can be written as structured groups of digits, separated by periods or slashes. Day (1 or 2 digits)/month (1 or 2 digits)/year (2 digits) Day (1 or 2 digits).month (1 or 2 digits).year (2 or 4 digits) 17/12/ am Time Indications Time indications will be pronounced correctly when written in one of the following formats: 9:15 09:15 Uhr 4.00 Uhr Uhr Currencies The German Text-To-Speech system correctly handles the German currency indication DM when written in one of the following ways: DM DM DM 10,- The Austrian, Swiss and American currencies are handled correctly: 19 Sfr 130 ÖS. 5 $ Other currency indications have to be written in full words: User sguide for German Chapter I/26

27 Chapter I German Text-to-Speech system 15 Franken 10 Gulden Currencies up to 15 digits will be correctly pronounced. Periods may be used to separate groups of digits DM DM Decimal digits in combination with currency indications are also supported. Decimal currency amounts up to 15 digits will be pronounced correctly. 1999,50 DM 1.999,50 DM Abbreviations and Acronyms The German Text-To-Speech system contains a dictionary with the most common abbreviations and acronyms, such as: bzw.: beziehungsweise MwSt.: Mehrwertsteuer usw.: und so weiter Some abbreviations are not case-sensitive: uppercase and lowercase are both accepted. Examples of case-sensitive abbreviations are: Hrsg.: Herausgeber hrsg.: Herausgegeben So.: Sonntag Abbreviations that are NOT in the dictionary: will be spelled if they consist of consonants only (with or without punctuation) will be spelled if the abbreviation contains one or more vowels, separated by periods will be spoken as full words if the abbreviation contains one or more vowels and is not separated by periods User sguide for German Chapter I/27

28 RealSpeak Telecom SDK Chapter II Preprocessor User s Guide for German V4.0

29 Chapter II Preprocessor Preprocessor Introduction The ScanSoft preprocessor (EMPP) is developed to analyze a specific type of text, namely messages. messages differ from any average type of text in both their structure and contents. An message consists of two clearly distinguishable parts: the header and the body. A substantial part of the header contains routing and administrative information, which is irrelevant to the user. Both the header and the body contain all kinds of specific text features, e.g. addresses, emoticons such as smileys, etc. Furthermore, informal writing is often combined with a lack of grammatical conventions. Spelling rules are frequently violated, punctuation is often omitted, etc. Although the standard ScanSoft Text-To-Speech system can handle special text items (abbreviations, numbers, dates, etc.), it is not capable of correctly handling all specific text features. These text features are therefore dealt with by the preprocessor. The EMPP transforms specific information into a format that complies with the rules of the standard ScanSoft Text-To-Speech system. The EMPP is a plug-in preprocessing module of the ScanSoft Text-To-Speech system. It replaces the preprocessor of the standard Text-To-Speech system. In the following sections you will find a description of the functioning of the ScanSoft preprocessor as well as an overview of its features. The preprocessor has two main tasks: processing of the header and processing of the body of the message. The input to the EMPP consists of one or more messages. In order to process the header, the EMPP extracts relevant header fields and then provides an intelligent header field reading. User sguide for German Chapter II/29

30 Chapter II Preprocessor During the processing of the body, the text is divided into smaller text units, called text-to-speech messages, which are synthesized by the Text-To-Speech system. Text normalization is applied to specific text features such as addresses, proper names, emoticons, URLs (Universal Resource Locators), etc. For the text normalization of an message, the ScanSoft EMPP applies linguistic rules and performs dictionary look-up, in order to yield an adequate phonetic transcription. The EMPP also supports the ScanSoft user dictionary mechanism, which allows the user to customize the output of the processing. Header Processing Header Field Extraction An message consists of two clearly distinguishable parts: the header and the body. The EMPP detects the header and extracts the relevant header fields. Information that is of no interest to the user (such as routing information) is not retained. The EMPP extracts the following header fields: From Field Date Field Subject Field Contains the sender s name and/or addres Contains the date and time of sending Optionally contains the subject of the The extraction of the header fields is based on the detection of specific keywords in the header. The supported keywords for the extraction of the header fields are listed below: From Field Date Field Subject Field: From: Author: Sender: De: Von: Date: Enviado: Gesendet: Subject: Subj: Asunto: Betreff: User sguide for German Chapter II/30

31 Preprocessor Chapter II The following is an example of header field extraction. The original header holds information that is irrelevant to the user. After extraction of date, sender and subject, the processed header merely mentions the Date field, the From field and the Subject field.: Original header: Path: news.be.innet.net!inbe.net!news.nl.innet.net!innl.net!hunter.prem ier.net! s.com!howland.erols.net!news.sprintlink.net!news- peer.sprintlink.net!uunet!in3.uu.net!01-newsfeed.univie.ac.at!02- newsfeed.univie.ac.at!news.ecrc.de!news00.btx.dtag.de!not-formail From: (Ulrike Noska) Newsgroups: de.etc.sprache.deutsch Subject: Re: Neue Wörter Date: 19 Oct :08:38 GMT Organization: Telekom Online Internet Gateway Lines: 10 Message-ID: news00.btx.dtag.de> Mime-Version: 1.0 Content-Type: text/plain; charset=iso Content-Transfer-Encoding: 8bit X-Sender: (Ulrike Noska) X-Mailer: Mozilla 2. 01DT [de] (Win16; I) Extracted header fields: From: (Ulrike Noska) Subject: Re: Neue Wörter Date: 19 Oct :08:38 GMT User sguide for German Chapter II/31

32 Chapter II Preprocessor Header Field Reading After the header fields have been extracted, they are processed by the EMPP. The header field keywords (see above) are replaced by an introductory message. The remainder of the header fields is processed by the EMPP in order to allow the Text-To-Speech system to intelligently read the fields. From Field The From field keyword is replaced by the introductory message Nachricht von:. Author: Thomas Kärtner is pronounced: Nachricht von: Thomas Kärtner The remainder of the From field is further processed by the EMPP. The EMPP supports From fields that either consist of a) a proper name b) a proper name and an address c) an address a) - b) In case the From field contains a proper name, this name and only this name is sent to the Text-To-Speech system. This means that if both a name and an address are found in the From field, the address will not be read by the Text-To-Speech system. From: Alex Wunneburger From: "Jo Hans Kessler" <kessler@alpha. ee.upenn.edu> From: Peter Kuppelwieser at IepXchgPO From: Udo Kohl/LHS/IEP/BE From: heikof@elis. rug.ac.be ( Heiko Felsmann) User sguide for German Chapter II/32

33 Chapter II Preprocessor are pronounced: Nachricht von: Alex Wunneburger Nachricht von: Jo Hans Kessler Nachricht von: Peter Kuppelwieser Nachricht von: Udo Kohl Nachricht von: Heiko Felsmann c) In case the From field contains only an address, the EMPP extracts the name out of the address and expands the domain that is contained in the address. In other words, the address is not read literally. Author: Author: wien.at!kunze are pronounced: Nachricht von: thomas b at Bundestag punkt d e Nachricht von: kunze at Wien punkt a t Date Field The Date field keyword is replaced by the introductory message Datum: The Date field contains the date and time of sending. The EMPP supports multiple date and time formats, which are transformed to a uniform format that complies with the rules for date and time indications of the ScanSoft Text-To-Speech system. The EMPP only pronounces the date. The EMPP supports dates in the following formats: Date: 12/23/ :15 PM Date: 13 Mar :50 AM are pronounced: Datum: 23. Dezember 2002 Datum: 13. März 1996 User sguide for German Chapter II/33

34 Chapter II Preprocessor Subject Field The Subject field keyword is replaced by the introductory message Betreff:. The Subject field can contain all kinds of data, but may also be empty. The EMPP searches for keywords that are typical of the subject field (e.g. RE, FYI, FW). Subject: RE: Wer holt mich ab? Subject: FYI: Stundenplan fuer die naechsten Monate Subject: FW: Ich habe es gefunden!!!! are pronounced: Betreff: Antwort auf: Wer holt mich ab? Betreff: zu Ihrem Interesse: Stundenplan fuer die naechsten Monate Betreff: weitergeleitete Nachricht: Ich habe es gefunden!!!! User sguide for German Chapter II/34

35 Preprocessor Chapter II body processing Message Extraction The preprocessor splits the body of the message into text-to-speech messages. This is done on the basis of a number of criteria, such as punctuation, capitalization, layout, intelligent abbreviation handling, etc. The following examples illustrate some criteria for splitting the text into text-to-speech messages: Using sentence final punctuation and capital letters Schreiben Sie mir bitte, wenn Sie noch Fragen haben. Meine - Adresse ist annap@henkel.co.at Using layout Themen der Versammlung: 1) Neue Kollegen 2) Organisation des Umzugs 3) Colloquium in Berlin Using intelligent abbreviation handling Bitte kontaktieren Sie Dr. A. M. Sander. User sguide for German Chapter II/35

36 Chapter II Preprocessor Text Normalization An message typically contains specific text features, such as addresses, URLs, file names, emoticons, etc. The EMPP transforms these specific features into a format that complies with the rules of the standard text normalization of the ScanSoft Text-To-Speech system. The following are examples of specific text normalization: Support for multiple address formats plknapp@unix.government.de Support for URLs (Universal Resource Locators) gopher://viertel.com/1 Support for file names ldb001.tse sysinfo.exe lipedu.xls Processing of emoticons :-) is pronounced : haha :-x is pronounced : Nicht weitersagen! :-( is pronounced :och Processing of overuse of punctuation PASS AUF!!!!!!!: INTERNET VIRUS!!!!!!!! Ich habe immer noch Probleme! #%&#@$<! Kannst Du mir helfen? becomes: PASS AUF! INTERNET VIRUS! Ich habe immer noch Probleme! Kannst Du mir helfen? User sguide for German Chapter II/36

37 Chapter II Preprocessor Normalization of lay-out lines (e.g. part of an signature); not active when in spell mode. These sequences of identical characters are not pronounced: o 10 or more identical digits o a word consisting of 5 or more identical US-ASCII encoded letters of the modern Latin alphabet o a sequence of 3 or more identical US-ASCII characters that are no letters, no digits, no sentencefinal punctuation marks (.?!) and no white spaces; e.g. '&', '#', '%', '*', '-' oooooooooooooooooooooooooooooo will be removed. User sguide for German Chapter II/37

38 Chapter II Preprocessor Processing of Question/Answer (FAQ) Q. Ich habe eine neue -Adresse. Wie kann ich meine EcoLink mails weiter empfangen? A. Schick einfach eine mit Deiner alten und Deiner neuen Adresse an becomes: Frage: Ich habe eine neue -Adresse. Wie kann ich meine EcoLink mails weiter empfangen? Antwort: Schick einfach eine mit Deiner alten und Deiner neuen Adresse an Processing of inserted mail Ralf> Na, ich hab' mir 'ne neue Stereoanlage Ralf>gekauft und bin ziemlich Ralf> enttäuscht. Die klingt zwar nicht schlecht, Ralf> aber die Lautstärke- Ralf> regelung wirkt nicht. Ich weiss nicht, was Ralf>ich damit machen soll. Heide> Bring sie gleich zurück und tausch sie um! becomes: Ralf: Na, ich hab' mir 'ne neue Stereoanlage gekauft und bin ziemlich enttäuscht. Die klingt zwar nicht schlecht, aber die Lautstärke-regelung wirkt nicht. Ich weiss nicht, was ich damit machen soll. Heide: Bring sie gleich zurück und tausch sie um! User sguide for German Chapter II/38

39 Chapter II Preprocessor Language Specific Text Normalization Umlaut English Words The L&H preprocessor for German is able to convert umlauted characters. In messages, umlauts are written in several ways: über ueber gelähmt All above-mentioned ways of writing the umlaut are supported by the L&H preprocessor. Since is an international medium, German messages will inevitably contain a lot of English words, which might refer to Internet, electronic mail or soft- and hardware. The typical jargon is handled by the exceptions dictionary of the preprocessor. This dictionary is a lexicon for terminology and provides the Text-To-Speech system with an adequate German transcription for a number of English words. asciitext attachment banner forwarded frame hypertext iexplorer image /+'as.ki:.tekst /+$.'tet&s.m$nt /+'be.n$r /+'fo:r.var.d$t /+'fre:m /+'ha&i.p$r.tekst /+'?In.t$R.nEt.Eks.'plo:R$R /+'?I.m$t&S User sguide for German Chapter II/39

40 Chapter II Preprocessor Customizing the Preprocessor The preprocessor supports the standard ScanSoft Text-To- Speech SDK user dictionary mechanism, which allows the user to customize the output of the preprocessor. The user dictionary is consulted both during the header processing and the body processing. For more information on how to build and use user dictionaries, see the User Configuration chapterof theprogrammer s Guide. Customizing the Header The user dictionary is consulted during the header processing while reading the From field and the Subject field. From Field The From field either consists of a) a proper name b) a proper name and an address c) an address a) In case the From field contains only a proper name, the name is passed to the user dictionary. If the lookup is successful, the proper name is substituted by the replacement string. If not, the name is further processed by the header reading module. If the user dictionary contains the following line: Johann /+'jo:.han the following From field: From: Johann Strauss Becomes: Nachricht von: */+'jo:.han*/+ Strauss User sguide for German Chapter II/40

41 Chapter II Preprocessor b) In case the From field contains a proper name and an address, the EMPP first passes the address to the user dictionary. If the lookup is successful, both the proper name and the address are substituted by the replacement string. If not, the EMPP passes the proper name to the user dictionary. If this lookup is successful, the name and the address are substituted by the replacement string. If not, the name is further processed by the header reading module. The address will not be read by the Text-To-Speech system. If the user dictionary contains the following lines: pschmidt@lhs.be Peter, vom Tennis Heinz Mein bester Freund Schreve /+ 'scre:.v$ the following From fields: From: "Alex Van Schreve" <VanSchreve@ALPHA.ee.upenn.edu> Author: heinz@elis.rug.ac.be (Heinz) From: pschmidt@lhs.be (P. Schmidt) become: Nachricht von: Alex Van */+ 'scre:.v$*/+ Nachricht von: Mein bester Freund Nachricht von: Peter, vom Tennis c) In case the From field contains only an address, the complete address is looked up in the user dictionary. If the lookup is successful, a proper name is added to the From field. If not, only the domain part is sent to the user dictionary. The EMPP first calls the dictionary for the complete domain part. If the lookup is successful, the complete domain part is substituted by the replacement string. Otherwise, the EMPP cuts off the leftmost sublevel domain and repeats the lookup and matching procedures for the remainder of the domain part. If the lookup is successful, the remainder of the domain part is substituted by the replacement string. This procedure is repeated until the top level domain is encountered. If none of the lookups is successful, the address is further processed by the header reading module. User sguide for German Chapter II/41

42 Chapter II Preprocessor NOTE If the user dictionary contains the following lines: Mein bester Freund postbank.de Deutsche Postbank AG the following From fields: Sender: From: become: Nachricht von: Mein bester Freund Nachricht von: anja gihl at Deutsche Postbank AG To allow a correct processing of the From field, the replacement string in the user dictionary should not contain an address or a domain. Subject Field Every word in the Subject field is sent to the user dictionary. If the lookup is successful, the replacement string is sent directly to the Text-To-Speech system. If not, the Subject field is further processed by the header reading module. If the user dictionary contains the following lines: ECAI /+?e:.tse:.?a:.'?i: IDT I D T the following Subject fields: Subject: ECAI Versammlung am Donnerstag Subject: Windows und IDT User sguide for German Chapter II/42

43 Chapter II Preprocessor are pronounced: Betreff: */+?e:.tse:.?a:.'?i:*/+ Versammlung am Donnerstag Betreff: Windows und I D T Customizing the Body When the user dictionary has been loaded, the EMPP will call the dictionary for every word of the body. If the word is found in the user dictionary, it is substituted by the replacement string. If not, the body is further processed by the body processing module. If the user dictionary contains the following line: IH : Iris Heller The word "IH" in the following sentence: Ich bleibe hier und IH wird die Kunden besuchen. is replaced by the corresponding string found in the user dictionary: Ich bleibe hier und Iris Heller wird die Kunden besuchen. User sguide for German Chapter II/43

44 RealSpeak Telecom SDK Chapter III SSML Preprocessor User s Guide for German V4.0

45 Chapter III SSML Preprocessor SSML Preprocessor Introduction SSML (Speech Synthesizer Markup Language) is part of a set of markup specifications by the W3C for voice browsers. General information regarding the RealSpeak SSML processor can be found in the SSML Support chapter of the Programmer s Guide. The RealSpeak Telecom SDK provides a built-in preprocessor that supports a large portion of the SSML 1.0 September 2004 Recommendation (REC). Moreover RealSpeak extends SSML with a number of Scansoft specific elements/attributes. The set supported by Scansoft is caled ScanSoft SSML (4SML). The section below describes language-specific SSML support included in the RealSpeak Telecom V4.0 German language version. German specific SSML markup XML encoding types for German The encoding is specified in the XML text declaration ("<?xml?>") by the encoding declaration which is of the form encoding="<encodingname>". E.g. <?xml version="1.0" encoding="utf-8"?> RealSpeak Telecom V4.0 German supports: Windows-1252 and ISO (ISO Latin1) The Unicode encoding UTF-8, UTF-16 and UCS-4 (Note that the alias "ISO UCS-4" is not supported) Any coding character set supported by the ICU component as long as the input text only contains characters that can be transcoded to the native coded character set, being Windows For more information about the character sets supported by ICU, take a look at the ICU website and User sguide for German Chapter III/45

46 Chapter III SSML Preprocessor NOTE 4SML Specifics for German Encoding names are parsed case-insensitive; hyphens and underscores are ignored For reasons of compatibility with the standard German system, the parallel text control sequence (<esc> sequence) is listed where applicable. As such, a similar TTS behavior can be created or combined with non-ssml driven text input. 4SML Tags Comment Corresponding control sequence High-level and document structure tags xml:lang Text normalization tags <say-as interpretas= xxx > <say-as interpretas= number format= cardinal > <say-as interpretas= number format= digits > <say-as interpretas= number format= decimal > <say-as interpretas= number > <say-as interpretas= number format= ordinal > <say-as interpretas= number format= telephone > de-de for German. Attribute of speak, paragraph, sentence and voice. ; limited support in mode. In mode the only supported interpretas value is spel. <esc>\tn=number_cardinal\ <esc>\ tn=number_digits\ <esc>\ tn=number_decimal\ <esc>\ tn=number\ <esc>\ tn=number_ordinal\ <esc>\ tn=number_telephone\ User sguide for German Chapter III/46

47 Chapter III SSML Preprocessor 4SML Tags Comment Corresponding control sequence <say-as interpretas= number format= telephone detail= punctuation > <say-as interpretas= ordinal> <say-as interpretas= acronym > <say-as interpretas= acronym detail= strict > <say-as interpretas= measure > <say-as interpretas= letters > <say-as interpretas= letters detail= strict > <say-as interpretas= words > <say-as interpretas= date > <say-as interpretas= date format= mdy > <say-as interpretas= date format = dmy > <say-as interpretas= date format= ymd > <say-as interpretas= date format= ym > <say-as interpret-as date format= my > <say-as interpretas= date format= dm > <say-as interpretas= date format= md > <say-as interpretas= date format= y > <esc>\ tn=number_telephone_punctuation \ <esc>\ tn=ordinal\ <esc>\ tn=acronym\ <esc>\ tn=acronym_strict\ <esc>\ tn=measure\ <esc>\ tn=letters\ <esc>\ tn=letters_strict\ <esc>\ tn=words\ <esc>\ tn=date\ <esc>\ tn=date_mdy\ <esc>\ tn=date_dmy\ <esc>\ tn=date_ymd\ <esc>\ tn=date_ym\ <esc>\ tn=date_my\ <esc>\ tn=date_dm\ <esc>\ tn=date_md\ <esc>\ tn=date_y\ User sguide for German Chapter III/47

48 Chapter III SSML Preprocessor 4SML Tags Comment Corresponding control sequence <say-as interpretas= date format= m > <say-as interpretas= date format= d > <say-as interpretas= time > <say-as interpretas= time format= h > <say-as interpretas= time format= hm > <say-as interpret as= time format= hms > <say-as interpretas= duration format= hms > <say-as interpretas= duration format= hm > <say-as interpretas= duration format= ms > <say-as interpretas= duration format= h > <say-as interpretas= duration format= m > <say-as interpretas= duration format= s > <say-as interpretas= duration > <say-as interpretas= curency > <say-as interpretas= telephone > <say-as interpretas= telephone detail= punctuation > <say-as interpretas= address > <esc>\ tn=date_m\ <esc>\ tn=date_d\ <esc>\ tn=time\ <esc>\ tn=time_h\ <esc>\tn=time_hm\ <esc>\ tn=time_hms\ <esc>\ tn=duration_hms\ <esc>\ tn=duration_hm\ <esc>\ tn=duration_ms\ <esc>\ tn=duration_h\ <esc>\ tn=duration_m\ <esc>\ tn=duration_s\ <esc>\ tn=duration\ <esc>\ tn=currency\ <esc>\ tn=telephone\ <esc>\ tn=telephone_punctuation\ <esc>\ tn=address\ User sguide for German Chapter III/48

49 Chapter III SSML Preprocessor 4SML Tags Comment Corresponding control sequence <say-as interpretas= spell > <say-as interpretas= name > <say-as interpretas= net format= > <say-as interpretas= net format= uri > <say-as interpretas= net > Pronunciation tags <phoneme alphabet= unipa > See section the German L&H+ and UNIPA phonetic alphabets for an overview of the alphabet. <esc>\ tn=spell\ <esc>\ tn=name\ <esc>\ tn=net_ \ <esc>\ tn=net_uri\ <esc>\ tn=net\ User sguide for German Chapter III/49

50 RealSpeak Telecom SDK Chapter IV Custom G2P Dictionaries User s Guide for German V4.0

51 Chapter IV Custom G2P Dictionaries Custom G2P Dictionaries Introduction ScanSoft's RealSpeak system now offers support for custom G2P dictionaries. A custom G2P dictionary module is an add-on module specifically designed to improve the quality of pronunciation for specific kinds of words. One example of a custom G2P dictionary module currently available from Scansoft is the propernames. module, which is described below. Check with Scansoft for the availability of other custom G2P dictionary modules. Proper names Custom G2P dictionary The standard German RealSpeak system correctly pronounces a number of common proper names. However, given the complexity of the grapheme to phoneme conversion for proper names, a dedicated module is used to guarantee the same quality for proper names as for common words. The propernames module contains a lexicon of around 85K proper names (including company/brand names, countries, cities, and first/last names of people). The propernames customg2p module is dynamically enabled/disabled by using the 4SML tags <ssft=domain type= propernames >, or <say-as type= name >, or the Scansoft escape sequences. Example: Mr. VanderHoff went to Abbeville, and also to Bois-de-Boulogne. <sft=domain type= propernames > Mr. VanderHof went to Abbeville, and also to Bois-de-Boulogne </ssft=domain>. Mr. VanderHoff went to Abbeville, and also to <say-as type= name > Bois-de-Boulogne </say-as>. User sguide for German Chapter IV/51