Presentation is loading. Please wait.

Presentation is loading. Please wait.

Introduction to VoiceXML 2.0 Rob Marchand Director of Product Management VoiceGenie Technologies Inc.

Similar presentations


Presentation on theme: "Introduction to VoiceXML 2.0 Rob Marchand Director of Product Management VoiceGenie Technologies Inc."— Presentation transcript:

1

2 Introduction to VoiceXML 2.0 Rob Marchand Director of Product Management VoiceGenie Technologies Inc.

3 Introduction to VoiceXML Audience o Managers and programmers with little experience with VoiceXML Attendees will learn o The basic principles of VoiceXML, o Just enough syntax to design and code simple speech applications requiring voice menus and voice forms.

4 VoiceXML in the Marketplace VoiceXML 2.0 is now ratified as a Recommendation (e.g., official standard) by the W3C Hundreds of millions of VoiceXML calls are answered every day VoiceXML is the standard for building speech-enabled applications

5 W3C and VoiceXML Forum W3C manages the technical evolution and development of the VoiceXML language VoiceXML Forum focuses on providing best practices, certification testing, resources and tools Together the W3C and VoiceXML Forum accelerate the adoption of VoiceXML-based speech applications

6 Outline Motivation for VoiceXML W3C Speech Interface Framework Languages Dialog—VoiceXML 2.0 Speech Synthesis—SSML Grammars—SRGS Semantic Interpretation—SI Call Control

7 Motivation for Speech Applications Users access Web sites from any telephone, anywhere, any time. Speaking and listening are the natural usage modes for phones.

8 Speech-enabled Applications Are Possible Now Increased computing power at less expense o Due to improved chip design and manufacturing techniques Improved speech recognition o Due to refinements to basic speech recognition algorithms Improved dialog design using voice o Minimizes the number of words and phrases that the speech recognizer must process at any point during the dialog

9 Strength of VoiceXML Applications Traditional system-directed dialogs for novice users Mixed initiative dialogs for experienced users Novice users smoothly become experienced users at their own pace

10 Limitations of VoiceXML Applications No special analysis of speech input o Not suitable for training speech skills—Reading, ESL, singing, etc. VUI conversational bandwidth is slower than GUI conversational bandwidth o Using a VUI is like drinking from Lake Superior with a straw

11 Exercise 1 Name or describe a speech application you could use at work. Name or describe a speech application you or family member can use at home.

12 XML o XML = eXtensible Markup Language o Elements are surrounded by tags Welcome to the voice system o Elements may be nested Welcome to Ajax Travel we have the cheapest fares o Elements may have attributes o Because “ ”, and “&” have special meanings “<” in place of “<” “>” in place of “>” “&” in place of “&”.

13 Outline Motivation for VoiceXML W3C Speech Interface Framework Languages Dialog—VoiceXML 2.0 Speech Synthesis—SSML Grammars—SRGS Semantic Interpretation—SI Call Control

14 DB Multimedia Files Audio Files Web Server HTML Scripts VoiceXML Scripts Grammars Speech Server/Gateway Web Browser Capture Voice ASR DTMF Replay Audio TTS Database Server Voice Browser Documents

15 W3C Speech Interface Framework Speech Synthesis GrammarOther VoiceXML 2.0 Call Control Semantic Interpretation

16 Status of W3C Speech Interface Languages Call Control Semantic Interpret- ration Recommendation Proposed Recommendation Candidate Recommendation Last Call Working Draft Requirements Working Draft V 3 Synthesis Grammar VoiceXML 2.0 VoiceXML 2.1 PLS

17 Outline Motivation for VoiceXML W3C Speech Interface Framework Languages Dialog—VoiceXML 2.0 Speech Synthesis—SSML Grammars—SRGS Semantic Interpretation—SI Call Control

18 VoiceXML 2.0 Fragment … Which account savings or checking savings checking CD certificate of deposit $ = “CD” …. … Dialog Language (VoiceXML 2.0) Speech Synthesis Markup Language (SSML) Speech Recognition Grammar Speci

19 VoiceXML 2.0 Fragment … Which account savings or checking savings checking CD certificate of deposit $ = “CD” …. … Dialog Language (VoiceXML 2.0) Speech Synthesis Markup Language (SSML) Speech Recognition Grammar Specification erpretation (SI)

20 VoiceXML 2.0 Fragment … Which account savings or checking savings checking CD certificate of deposit $ = “CD” …. … Dialog Language (VoiceXML 2.0) Speech Synthesis Markup Language (SSML) Speech Recognition Grammar Specification (SRGS) Semantic Interpretation (SI)

21 VoiceXML 2.0 Fragment … Which account savings or checking savings checking CD certificate of deposit $ = “CD” …. … Dialog Language (VoiceXML 2.0) Speech Synthesis Markup Language (SSML) Speech Recognition Grammar Specification (SRGS) Semantic Interpretation (SI)

22 VoiceXML 2.0 features Menus, forms, sub-dialogs o,, Inputs o Speech recognition o Recording o Keypad Output o Audio files o Text-to-speech Variables o Events –,,,, Transition and submission –, –Telephony –Connection control –, –Telephony information –Platform –Objects –Performance –Fetch

23 A Typical Voice Menu Do you want to listen, next, prior, buy, or exit? listen next prior buy exit Exercise 2: Write a menu that asks the user a “yes/no” question to confirm that the user wants to buy the audio “three blind mice

24 Answer to Exercise 2 A “yes/no” menu Do you want to buy three blind mice now? yes no

25 Typical Form Fill-In Welcome to the electronic payment system. Please enter your credit card number? Please enter your expiration date Exercise 3: Write a form that solicits the month, day, and year for the user’s birth date.

26 Answer to Exercise 3 When were you born? What month? What day of the month? What year

27 Event Handlers Deal with exceptional or error conditions Control mechanism for dialog turn retries o … o o … Shorthand notation available o …, etc. Scoped according to where they occur o,, etc.

28 Adding Event Handlers When were you born? ….. ….. What month? …..

29 Adding Event Handlers When were you born? ….. ….. What month? …..

30 Adding Event Handlers When were you born? ….. ….. What month? …..

31 Default Event Handlers Sorry, no help is available. I did not understand, please try again I did not hear anything, please speak again

32 Exercise 4 Write event handlers for the month field ____________________ __________________________ ___________________________________

33 Answer to Exercise 4 Write event handlers for the month field Which month, for example, January February, or March? Say the name of the month you were born in In what month were you born?

34 Outline Motivation for VoiceXML W3C Speech Interface Framework Languages Dialog—VoiceXML 2.0 Speech Synthesis—SSML Grammars—SRGS Semantic Interpretation—SI Call Control

35 Speech Synthesis ML Structure Analysis Text Normali- zation Text-to- Phoneme Conversion Prosody Analysis Waveform Production Markup support: p, s Non-markup behavior: infer structure by automated text analysis

36 Before and after Structure Analysis Before structure analysis o Dr. Smith lives at 214 Elm Dr. He weights 214 lb. He plays bass guitar. He also likes to fish; last week he caught a 19 lb. bass. After structure analysis He plays bass guitar. He also likes to fish; last week he caught a 19 lb. bass. Dr. Smith lives at 214 Elm Dr. He weights 214 lb.

37 Speech Synthesis ML Structure Analysis Text Normali- zation Text-to- Phoneme Conversion Prosody Analysis Waveform Production Markup support: say-as for dates, times, etc. sub for aliasing Non-markup behavior: automatically identify and convert constructs Markup support: p, s Non-markup behavior: infer structure by automated text analysis

38 After Text Normalization Dr. Smith lives at 214 Elm Dr. He weights 214 lb. He plays bass guitar. He also likes to fish; last week he caught a 19 lb. bass.

39 Dr. Smith lives at 214 Elm Dr. He weighs 214 lb. He plays bass guitar. He also likes to fish; last week he caught a 19 lb. bass.

40 Speech Synthesis ML Structure Analysis Text Normali- zation Text-to- Phoneme Conversion Prosody Analysis Waveform Production Markup support: phoneme, say-as Non-markup behavior: look up in pronunciation dictionary Markup support: say-as for dates, times, etc. sub for aliasing Non-markup behavior: automatically identify and convert constructs Markup support: p, s Non-markup behavior: infer structure by automated text analysis

41 After text-to-phoneme conversion Dr. Smith lives at 214 Elm Dr. He weighs 214 lb. He plays bass guitar. He also likes to fish; last week he caught a 19 lb. bass.

42 Speech Synthesis ML Structure Analysis Text Normali- zation Text-to- Phoneme Conversion Prosody Analysis Waveform Production Markup support: emphasis, break, prosody Non-markup behavior: automatically generate prosody through analysis of document structure and sentence syntax Markup support: phoneme, say-as Non-markup behavior: look up in pronunciation dictionary Markup support: say-as for dates, times, etc. sub for aliasing Non-markup behavior: automatically identify and convert constructs Markup support: p, s Non-markup behavior: infer structure by automated text analysis

43 Prosody Analysis (Initial text) Environmental control menu. Do you want to adjust the lighting or temperature?

44 Prosody Analysis (Add pause at phrase boundaries) Environmental control menu Do you want to adjust the lighting or temperature?

45 Prosody analysis (De-emphasize familiar words) Environmental control menu Do you want to adjust the lighting or temperature?

46 Prosody Analysis (pause to let the listener catch up) Environmental control menu do you want to adjust the lighting or temperature?

47 Prosody Analysis (Add emphasis to focus listener’s attention) Environmental control menu do you want to adjust the lighting or temperature?

48 Speech Synthesis ML Structure Analysis Text Normali- zation Text-to- Phoneme Conversion Prosody Analysis Waveform Production Markup support: voice, audio* Markup support: emphasis, break, prosody Non-markup behavior: automatically generate prosody through analysis of document structure and sentence syntax Markup support: phoneme, say-as Non-markup behavior: look up in pronunciation dictionary Markup support: say-as for dates, times, etc. sub for aliasing Non-markup behavior: automatically identify and convert constructs Markup support: paragraph, sentence Non-markup behavior: infer structure by automated text analysis *audio icons, branding, advertising

49 Waveform Production Environmental control menu. Do you want to adjust the lighting or temperature

50 Exercise 5 ( insert SSML commands ) Welcome to Ajax Bank do you want to withdraw or deposit funds?

51 Answer to Exercise 5 Welcome to Ajax Bank do you want to withdraw or deposit funds?

52 Outline Motivation for VoiceXML W3C Speech Interface Framework Languages Dialog—VoiceXML 2.0 Speech Synthesis—SSML Grammars—SRGS Semantic Interpretation—SI Call Control

53 Grammars Describe what the user may say at a point in the dialog Enable the speech recognition engine to work faster and more accurately Consist of one or more “rules”

54 Example Grammar zero ten one two three four five six seven eight nine XML form of grammars

55 Example Grammar zero ten one two three four five six seven eight nine Rule describing single digits Rule describing digits zero through ten

56 Example Grammar zero ten one two three four five six seven eight nine Grammar processor should start with the “zero_to_ten” rule

57 Example Grammar zero ten one two three four five six seven eight nine This is a grammar used by the speech recognizer. (There may also be grammars for DTMF recognizers.)

58 Example Grammar zero ten one two three four five six seven eight nine describes alternatives

59 Example Grammar zero ten one two three four five six seven eight nine Rule element references another rule

60 Example Grammar zero ten one two three four five six seven eight nine Exercise 6: Write a grammar for that recognizes the digits zero to nineteen

61 Answer to Exercise 6 Write a grammar for zero to nineteen zero ten eleven twelve thirteen fourteen fifteen sixteen seventeen eighteen nineteen one two three four five six seven eight nine

62 More Grammar Elements Repeat and optional very good Sequence Twenty Garbage James Lewis

63 Exercise 7 Write a grammar for that recognizes the digits zero to thirty-nine

64 Answer to Exercise 7 Write a grammar for zero to thirty-nine zero twenty twenty thirty thirty one two three four five six seven eight nine ten eleven twelve thirteen fourteen fifteen sixteen seventeen eighteen nineteen

65 Reusing existing grammars <grammar type = "application/srgs+xml" root = "size” src = “http://www.example.com/size.grxml"/>

66 Outline Motivation for VoiceXML W3C Speech Interface Framework Languages Dialog—VoiceXML 2.0 Speech Synthesis—SSML Grammars—SRGS Semantic Interpretation—SI Call Control

67 Semantic Interpretation To create smart voice user interfaces, we need to extract the semantic information from speech utterances Example: o Utterance: “I want to fly from Dublin to Paris” o Semantic Interpretation: { origin: “Dublin” destination: “Paris” }

68 Semantic Interpretation ASR Grammar with Semantic Interpretation Scripts Semantic Interpretation Processor VoiceXML Interpreter Application text ECMAScript object fourteen

69 Semantic Interpretation ASR Grammar with Semantic Interpretation Scripts VoiceXML Interpreter Application text fourteen ECMAScript object Semantic Interpretation Processor

70 Semantic Interpretation ASR Grammar with Semantic Interpretation Scripts VoiceXML Interpreter Application text fourteen fourteen $.quantity=“14”; ECMAScript object Semantic Interpretation Processor

71 Semantic Interpretation ASR Grammar with Semantic Interpretation Scripts VoiceXML Interpreter Application text fourteen { quantity: “14” } fourteen $.quantity=“14”; ECMAScript object Semantic Interpretation Processor

72 Semantic Interpretation ASR Grammar with Semantic Interpretation Scripts VoiceXML Interpreter Application text fourteen quantity = “14” fourteen { quantity: “14” } fourteen $.quantity=“14”; ECMAScript object Semantic Interpretation Processor

73 Semantic Interpretation Semantic Interpretation defines the content of s in SRGS grammars Two kinds of syntax for contents: o Semantic Literals (literal values) o Semantic Scripts (ECMAScript)

74 Semantic Interpretation Semantic Literals example: coca cola coke cola coke black fizzy stuff coke coke

75 Semantic Interpretation Semantic Literals example: coca cola coke cola coke black fizzy stuff coke coke Default Assignment

76 Semantic Interpretation Semantic Scripts employ ECMAScript Advantages: Richer structure (objects) Ability to perform computations

77 Semantic Interpretation Example grammar rule with Script Syntax: small $.size = "small"; medium $.size = "medium"; large $.size = “large"; green $.color = "green"; blue $.color = "blue"; white $.color = "white"; ECMAScript structure: action: { size: "large" color: "white" } Large white

78 Semantic Interpretation Example grammar rule with Script Syntax: What is $.total = $digit; plus $.total = $.total + $digit; ECMAScript structure: calculator: { total: 6 } What is 1+ 2+ 3?

79 Exercise 8 Fill in the contents of Grammar rule: from savings ________________________ checking ________________________ to savings ________________________ checking ________________________ ECMAScript structure: transfer: { source_account: "savings" target_account: “checking" } From savings to checking

80 Answer to Exercise 8 From savings to checking Grammar rule: from savings $.source_account = “savings"; checking $.source_account = “checking"; to savings $.target_account = “savings"; checking $.target_account = “checking"; ECMAScript structure: transfer: { source_account: "savings" target_account: “checking" }

81 Outline Motivation for VoiceXML W3C Speech Interface Framework Languages Dialog—VoiceXML 2.0 Speech Synthesis—SSML Grammars—SRGS Semantic Interpretation—SI Call Control

82 CCXML Provides call control support for VoiceXML and other dialog languages Separate interpreter from VoiceXML o Lives on its own thread o Handles asynchronous events May be used to create standalone applications Replaces and currently in VoiceXML 2.0 (or provides the underlying support for them)

83 CCXML VoiceXML

84 CCXML VoiceXML + CCXML

85 CCXML Features o Multi-party conferencing (human and machine) o Sophisticated multi-call handling and control o Support for async external messages and events o More sophisticated call control than VoiceXML o Call control protocol independence Goal to support very high density and performance

86 CCXML

87 CCXML

88 CCXML

89 CCXML

90 CCXML

91 CCXML A new participant has entered the conference.

92 Exercise 9 Announce when a caller leaves

93 Answer to Exercise 9

94 A participant has left the conference. Answer to Exercise 9

95 Example Applications with CCXML-VoiceXML Alerts o Stock value changes, order is available, flight is delayed, road closure, school closure Conference o Add additional person to the conference o Whisper o Eject Find me o Try alternative telephone numbers Instant messaging o Notify me when John calls in to access his e-mail Control home applications o Turn on/off coffee pot, oven, air conditioner, lights, arm/disarm the security system Call Center/Customer Care Applications

96 VoiceXML 2.1 VoiceXML’s success and popularity resulted in many implementations early in the standardization process Additional, innovative features were conceived after VoiceXML 2.0 content was agreed Goals of VoiceXML 2.1: o Ensure portability by specifying a set of commonly implemented extensions o Backwards-compatible with VoiceXML 2.0 o Follow a “fast track” to standardization

97 VoiceXML 2.1 Standardized extensions: o Locate barge-in occurrences within prompts o Interact directly with XML-based infrastructure o Access recognition utterances for analysis o Increase performance be reducing server round-trips o Extended call transfer types

98 Summary W3C Speech Interface Framework o Dialog—VoiceXML o Grammar—SRGS o Synthesis—SSML o Semantic Interpretation—SI o Call Control—CCXML Can work together or separately See http://www.w3.org/voice/ for detailshttp://www.w3.org/voice/

99 Resources

100 Industry Organizations World Wide Web Consortium o http://www.w3.orghttp://www.w3.org W3C Voice Browser Working Group o http://www.w3.org/voice/http://www.w3.org/voice/ W3C Multi-Modal Working Group o http://www.w3.org/2002/mmi/http://www.w3.org/2002/mmi/ VoiceXML Forum o http://www.voicexml.orghttp://www.voicexml.org SALT Forum: o http://www.saltforum.orghttp://www.saltforum.org Speech Technology Magazine o http://www.amcommexpos.com/http://www.amcommexpos.com/

101 Books James A. Larson, VoiceXML—An Introduction to Developing Speech Applications, 2002, Upper Saddle River, NJ: Prentice Hall. Eve Astrid Andersson, et.al., Early Adopter Voice, 2001, Birmingham UK: Vrox. Bruce Balentine & David P. Morgan, How to Build a Speech Recognition Application: A Style Guide for Telephony Dialogues, 1999, San Ramon, CA: Enterprise Integration Group. Rick Beasley et. al., Voice Application Development with Voice, 2002, Indianapolis: Sams. Bob Edgar, The Voice Handbook, 2001, New York: CMP. Susan Weinschenk & Dean T. Barker, Designing Effective Speech Interfaces, 2000, New York: John Wiley & Sons. Chetan Sharma & Jeff Kunins, Voice: Strategies and Techniques for Effective Voice Application Development with Voice 2.0, 2002, New York: John Wiley. Michael H. Cohen, James P. Giangola, & Jennifer Balogh, Voice User Interface Design, 2004, Addison Wesley.

102 Tutorials and Articles VoiceXML Forum o http://www.voicexmlforum.org/ http://www.voicexmlforum.org/ VoiceXML Review o http://www.voicexmlreview.org/ http://www.voicexmlreview.org/ World of VoiceXML o http://www.kenrehor.com/voicexml/ http://www.kenrehor.com/voicexml/

103 Online Voice SDKs NameURL BeVocal Cafehttp://cafe.bevocal.com Hey Anita FreeSpeechhttp://www.heyanita.com Tellme Studiohttp://studio.tellme.com VoiceGenie Developer Workshop http://developer.voicegenie.com Voxeo Communityhttp://www.voxeo.com Voxpilot voxbuilderhttp://www.voxbuilder.com

104 Downloadable Voice Interpreters NameURL IBM WebSphere Voice Server SDK http://www.ibm.com/software/voice

105 Public VoiceXML Interpreters InterpreterSourceURL OpenVXI - VoiceXML Interpreter Carnegie-Melon University Department of Computer Science Speech Group http://www.speech.c s.cmu.edu/openvxi/i ndex.html PublicVoiceXML - VoiceXML platform Public Voice Lab Vienna, Austria http://www.publicvoi cexml.org/

106 Introduction to VoiceXML Questions?


Download ppt "Introduction to VoiceXML 2.0 Rob Marchand Director of Product Management VoiceGenie Technologies Inc."

Similar presentations


Ads by Google