EP1184839A2 - Grapheme-phoneme conversion - Google Patents

Grapheme-phoneme conversion Download PDF

Info

Publication number
EP1184839A2
EP1184839A2 EP01117869A EP01117869A EP1184839A2 EP 1184839 A2 EP1184839 A2 EP 1184839A2 EP 01117869 A EP01117869 A EP 01117869A EP 01117869 A EP01117869 A EP 01117869A EP 1184839 A2 EP1184839 A2 EP 1184839A2
Authority
EP
European Patent Office
Prior art keywords
grapheme
word
interface
partial words
words
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
EP01117869A
Other languages
German (de)
French (fr)
Other versions
EP1184839B1 (en
EP1184839A3 (en
Inventor
Horst-Udo Hain
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Siemens AG
Original Assignee
Siemens AG
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Siemens AG filed Critical Siemens AG
Publication of EP1184839A2 publication Critical patent/EP1184839A2/en
Publication of EP1184839A3 publication Critical patent/EP1184839A3/en
Application granted granted Critical
Publication of EP1184839B1 publication Critical patent/EP1184839B1/en
Anticipated expiration legal-status Critical
Expired - Lifetime legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L13/00—Speech synthesis; Text to speech systems
    • G10L13/08—Text analysis or generation of parameters for speech synthesis out of text, e.g. grapheme to phoneme translation, prosody generation or stress or intonation determination

Definitions

  • the invention relates to a method, a computer program product, a data carrier and a computer system for grapheme-phoneme conversion of a word that as a whole is not is contained in a pronunciation dictionary.
  • Speech processing methods in general are, for example from US 6 029 135, US 5 732 388, DE 19636739 C1 and DE 19719381 C1 known.
  • a speech synthesis system that is Font-to-speech or grapheme-phoneme conversion of the crucial words. Error with Sounds, syllable borders and the emphasis on words can be heard directly, can lead to incomprehensibility and at worst Case even distort the meaning of a statement.
  • a remedy can be a morphological decomposition in this case create.
  • a word that is not found in the dictionary is into its morphological components such as prefixes, stems and Suffixes are broken down and these components are searched for in the lexicon.
  • a morphological decomposition is just about long words problematic because of the number of possible The word length increases. It also requires an excellent knowledge of the word formation grammar of a Language.
  • words that are not in a pronunciation dictionary can be found using out-of-vocabulary methods (OOV method), e.g. with neural networks, transcribed.
  • OOV method out-of-vocabulary methods
  • results as the phonetic conversion of whole words with the help of a pronunciation dictionary.
  • the word can also be broken down into partial words.
  • the partial words can be pronounced with the help of a pronunciation dictionary or an OOV procedure.
  • the found Partial transcriptions can be linked together. However, this leads to errors at the separation points between the Part transcriptions.
  • the object of the invention is to join together To improve partial transcriptions. This task is accomplished by a method, a computer program product, a data carrier and a computer system according to the independent claims.
  • the computer program is called a computer program product understood as a tradable product, in what form whatever, e.g. on paper, on a computer-readable data medium, distributed over a network, etc.
  • the transcriptions of the partial words are lined up one after the other, with at least one interface between the transcriptions of the partial words.
  • the at least an interface-bounding phoneme of the partial words be determined.
  • the Grapheme-phoneme conversion of the graphemes in the context of the respective Interface using a neural network new be calculated.
  • a pronunciation lexicon has the advantage of to deliver "correct" transcription. However, it fails when unknown words occur.
  • Neural networks can provide a transcription for any character string, but may make significant mistakes.
  • the invention combines the security of the lexicon with the flexibility of neural networks.
  • the transcription of the partial words can be done in different ways take place, e.g. with an out-of-vocabulary treatment (OOV treatment).
  • OOV treatment out-of-vocabulary treatment
  • a fairly reliable way is for the word in a database, the phonetic transcriptions of words contains to search for partial words. As Transcription is then made for one found in the database Partial word the phonetic recorded in the database Transcription chosen. For most words or Subwords for useful results.
  • the word next to the found subword is at least one has another component that is not in the database can be registered using OOV treatment be transcribed phonetically.
  • OOV treatment can be done using a statistical method, e.g. by means of a neural Network, or rule-based.
  • the word is advantageously divided into partial words of a certain Minimum length disassembled so that the largest possible partial words are found and there will be little reworking.
  • FIG. 1 shows a grapheme-phoneme conversion of a word suitable computer system.
  • This assigns a processor (processor, CPU) 20, a working memory (RAM) 21, a program memory (program memory, ROM) 22, a hard disk controller (hard disc controller, HDC) 23, which is a hard drive (hard disk) 30 controls, and an interface controller (I / O controller) 24 on.
  • Processor 20, memory 21, program memory 22, hard disk controller 23 and interface controllers 24 are on a bus that CPU bus 25, for exchanging data and commands with one another coupled.
  • the computer also has an input / output bus (I / O bus) 26 on the various input and output devices couples with the interface controller 24.
  • To the Input and output devices count e.g. a general Input and output interface (I / O interface) 27, a display device (display) 28, a keyboard (keyboard) 29 and a mouse 31.)
  • the remaining gaps in the preferred embodiment closed by a neural network.
  • the task of filling in the gaps is easier, because at least the left phoneme context is assumed to be safe can be, since it comes from the pronunciation dictionary comes.
  • the input of the previous phonemes stabilized thus the output of the neural network for the one to be filled Gap, because the phoneme to be generated not only from the letters, but also depends on the previous phoneme.
  • the ending ⁇ -ig> at the end of the syllable is spoken like [IC] in the phonetic spelling SAMPA, so like [I] (untensioned short non-rounded front vowel) followed by the first sound [C] (voiceless palatal fricative).
  • the prefix ⁇ er-> is spoken like [Er], with an [E] (untensioned short unrounded semi-open front vowel, open "e") and one [r] (central sonorant).
  • Remedy can be created here by using a neural Net calculated the last according to the left transcription. there the question arises, which letters at the end of the left transcription used to determine the last sound should be.
  • a special pronunciation dictionary is used for this decision used.
  • the special feature of this lexicon is that it contains information about which grapheme group for which Heard out loud. How to create the lexicon is in Horst-Udo Hain: "Automation of the Training Procedures for Neural Networks Performing Multi-Lingual Grapheme to Phoneme Conversion ". Eurospeech 1999, pp. 2087-2090.
  • the neural network can now use the existing one right context ⁇ wise> new about phoneme and syllable boundary decide at the end of the word.
  • the result in this case is that Phoneme [g] in front of which a syllable limit is set.
  • the syllable boundary is in the right place and the ⁇ g> is also transcribed as [g] and not as [C].
  • the first according to the right transcription is redetermined using the same scheme.
  • the correct transcription for ⁇ er-> from ⁇ weise> at this point is [6] and not [Er].
  • Two sounds have to be revised here, which is why in the preferred embodiment two sounds are always revised.

Abstract

A grapheme to phoneme conversion system using a computer programme selects graphemes either from a pronunciation dictionary for a whole word or by segmenting the word and iterating the grapheme selection to take account of combinations.

Description

Die Erfindung betrifft ein Verfahren, ein Computerprogrammprodukt, einen Datenträger und ein Computersystem zur Graphem-Phonem-Konvertierung eines Worts, das als Ganzes nicht in einem Aussprachelexikon enthalten ist.The invention relates to a method, a computer program product, a data carrier and a computer system for grapheme-phoneme conversion of a word that as a whole is not is contained in a pronunciation dictionary.

Sprachverarbeitungsverfahren im Allgemeinen sind beispielsweise aus US 6 029 135, US 5 732 388, DE 19636739 C1 und DE 19719381 C1 bekannt. Bei einem Sprachsynthese-System ist die Schrift-zu-Sprache- bzw. Graphem-Phonem-Konvertierung der zu sprechenden Wörter von entscheidender Bedeutung. Fehler bei Lauten, Silbengrenzen und der Wortbetonung sind direkt hörbar, können zur Unverständlichkeit führen und im schlimmsten Fall sogar den Sinn einer Aussage verdrehen.Speech processing methods in general are, for example from US 6 029 135, US 5 732 388, DE 19636739 C1 and DE 19719381 C1 known. In a speech synthesis system, that is Font-to-speech or grapheme-phoneme conversion of the crucial words. Error with Sounds, syllable borders and the emphasis on words can be heard directly, can lead to incomprehensibility and at worst Case even distort the meaning of a statement.

Die beste Qualität erhält man, wenn das zu sprechende Wort in einem Aussprachelexikon enthalten ist. Die Verwendung solcher Lexika bereitet jedoch Probleme. Auf der einen Seite erhöht die Anzahl der Einträge den Suchaufwand. Auf der anderen Seite ist es gerade bei Sprachen wie dem Deutschen nicht möglich, alle Wörter in einem Lexikon zu erfassen, da die Möglichkeiten der Kompositabildung nahezu unbeschränkt sind.The best quality is obtained when the word to be spoken in a pronunciation dictionary is included. The use of such However, encyclopedias cause problems. Increased on one side the number of entries the search effort. On the other hand it is not the case with languages like German possible to capture all the words in a lexicon because the Possibilities of composite formation are almost unlimited.

Abhilfe kann in diesem Fall eine morphologische Zerlegung schaffen. Ein Wort, das nicht im Lexikon gefunden wird, wird in seine morphologischen Bestandteile wie Präfixe, Stämme und Suffixe zerlegt, und diese Bestandteile werden im Lexikon gesucht. Eine morphologische Zerlegung ist jedoch gerade bei langen Wörtern problematisch, weil die Anzahl der möglichen Zerlegungen mit der Wortlänge steigt. Sie erfordert außerdem ein ausgezeichnetes Wissen über die Wortbildungsgrammatik einer Sprache. Daher werden Wörtern, die nicht in einem Aussprachelexikon gefunden werden, mit Out-Of-Vocabulary-Verfahren (OOV-Verfahren), z.B. mit Neuronalen Netzen, transkribiert. Solche OOV-Behandlungen sind allerdings relativ rechenintensiv und führen in aller Regel zu schlechteren Ergebnissen als die phonetische Konvertierung ganzer Wörter mit Hilfe eines Aussprachelexikons. Zur Bestimmung der Aussprache eines Worts, das nicht in einem Aussprachelexikon enthalten ist, kann das Wort auch in Teilwörter zerlegt werden. Die Teilwörter können mit Hilfe eines Aussprachelexikons oder eines OOV-Verfahrens transkribiert werden. Die gefundenen Teiltranskriptionen können aneinander gehängt werden. Dies führt jedoch zu Fehlern an den Trennstellen zwischen den Teiltranskriptionen.A remedy can be a morphological decomposition in this case create. A word that is not found in the dictionary is into its morphological components such as prefixes, stems and Suffixes are broken down and these components are searched for in the lexicon. However, a morphological decomposition is just about long words problematic because of the number of possible The word length increases. It also requires an excellent knowledge of the word formation grammar of a Language. Hence words that are not in a pronunciation dictionary can be found using out-of-vocabulary methods (OOV method), e.g. with neural networks, transcribed. However, such OOV treatments are relative computationally intensive and generally lead to worse ones Results as the phonetic conversion of whole words with the help of a pronunciation dictionary. To determine pronunciation of a word that is not in a pronunciation dictionary is included, the word can also be broken down into partial words. The partial words can be pronounced with the help of a pronunciation dictionary or an OOV procedure. The found Partial transcriptions can be linked together. However, this leads to errors at the separation points between the Part transcriptions.

Aufgabe der Erfindung ist es, das Aneinanderfügen von Teiltranskriptionen zu verbessern. Diese Aufgabe wird durch ein Verfahren, ein Computerprogrammprodukt, einen Datenträger und ein Computersystem gemäß den unabhängigen Ansprüchen gelöst.The object of the invention is to join together To improve partial transcriptions. This task is accomplished by a method, a computer program product, a data carrier and a computer system according to the independent claims.

Dabei wird unter einem Computerprogrammprodukt das Computerprogramm als handelbares Produkt verstanden, in welcher Form auch immer, z.B. auf Papier, auf einem computerlesbaren Datenträger, über ein Netz verteilt, etc.The computer program is called a computer program product understood as a tradable product, in what form whatever, e.g. on paper, on a computer-readable data medium, distributed over a network, etc.

Erfindungsgemäß wird bei der Graphem-Phonem-Konvertierung eines Worts, das als Ganzes nicht in einem Aussprachelexikon enthalten ist, zunächst das Wort in Teilwörter zerlegt. Anschließend wird eine Graphem-Phonem-Konvertierung der Teilwörter durchgeführt. According to the invention in the grapheme-phoneme conversion one Words that as a whole are not in a pronunciation dictionary is included, the word is first broken down into subwords. Subsequently becomes a grapheme-phoneme conversion of the subwords carried out.

Die Transkriptionen der Teilwörter werden hintereinander aufgereiht, wobei sich mindestens eine Schnittstelle zwischen den Transkriptionen der Teilwörter ergibt. Die an die mindestens eine Schnittstelle grenzenden Phoneme der Teilwörter werden bestimmt.The transcriptions of the partial words are lined up one after the other, with at least one interface between the transcriptions of the partial words. The at least an interface-bounding phoneme of the partial words be determined.

Dabei besteht die Möglichkeit, nur das letzte Phonem des in der zeitlichen Reihenfolge der Aussprache vor der Schnittstelle liegenden Teilworts zu berücksichtigen. Besser ist es jedoch, wenn sowohl das genannte als auch das erste Phonem der folgenden Silbe für die erfindungsgemäße Sonderbehandlung ausgewählt werden. Noch bessere Ergebnisse werden erzielt, wenn weitere angrenzende Phoneme einbezogen werden, z.B. ein oder zwei Phoneme vor der Schnittstelle und zwei nach der Schnittstelle.It is possible to only use the last phoneme of the in the chronological order of the pronunciation before the interface partial word to be considered. It is better however, if both the aforementioned and the first phoneme the following syllable for the special treatment according to the invention to be selected. Even better results are achieved if other adjacent phonemes are included, e.g. on or two phonemes before the interface and two after the Interface.

Anschließend werden diejenigen Grapheme der Teilwörter bestimmt, die die an die mindestens eine Schnittstelle grenzenden Phoneme erzeugen. Dies kann mittels eines Lexikons erfolgen, das angibt, durch welche Grapheme diese Phoneme erzeugt wurden. Wie das Lexikon zu erstellen ist, ist in Horst-Udo Hain: "Automation of the Training Procedures for Neural Networks Performing Multi-Lingual Grapheme to Phoneme Conversion", Eurospeech 1999, S. 2087-2090, ausgeführt.Then those graphemes of the partial words are determined, which border on the at least one interface Generate phonemes. This can be done using a lexicon, which indicates by which graphemes these phonemes are generated were. How to create the lexicon is in Horst-Udo Hain: "Automation of the Training Procedures for Neural Networks Performing Multi-Lingual Grapheme to Phoneme Conversion ", Eurospeech 1999, pp. 2087-2090.

Danach wird die Graphem-Phonem-Konvertierung der bestimmten Grapheme im Kontext, das heißt in Abhängigkeit des Kontexts, der jeweiligen Schnittstelle neu berechnet. Dies ist nur möglich, weil klar ist, welches Phonem durch welches Graphem bzw. welche Grapheme erzeugt wurde. After that, the grapheme-phoneme conversion is determined Graphemes in context, that is depending on the context, of the respective interface recalculated. This is only possible because it is clear which phoneme by which grapheme or which grapheme was generated.

Die Schnittstellen zwischen den Teiltranskriptionen werden somit gesondert behandelt. Gegebenenfalls werden Änderungen an den vorher ermittelten Teiltranskriptionen vorgenommen. Ein für ein Sprachsynthese-System nicht unerheblicher Vorteil der Erfindung ist die Beschleunigung der Berechnung. Während Neuronale Netze für die Konvertierung der 310000 Wörter eines typischen Lexikons für die deutsche Sprache ca. 80 Minuten benötigen, geschieht dies mit dem erfindungsgemäßen Ansatz in nur 25 Minuten.The interfaces between the partial transcriptions are thus treated separately. If necessary, changes made on the previously determined partial transcriptions. A not inconsiderable advantage for a speech synthesis system the invention is the acceleration of the calculation. While Neural networks for converting the 310000 words one typical lexicon for the German language about 80 minutes need, this is done with the inventive approach in only 25 minutes.

In einer vorteilhaften Weiterbildung der Erfindung kann die Graphem-Phonem-Konvertierung der Grapheme im Kontext der jeweiligen Schnittstelle mittels eines Neuronalen Netzes neu berechnet werden. Ein Aussprachelexikon hat den Vorteil, die "richtige" Transkription zu liefern. Es versagt jedoch, wenn unbekannte Wörter auftreten. Neuronale Netze können hingegen für jede beliebige Zeichenkette eine Transkription liefern, machen dabei aber unter Umständen erhebliche Fehler. Die Weiterbildung der Erfindung kombiniert die Sicherheit des Lexikons mit der Flexibilität der Neuronalen Netze.In an advantageous development of the invention, the Grapheme-phoneme conversion of the graphemes in the context of the respective Interface using a neural network new be calculated. A pronunciation lexicon has the advantage of to deliver "correct" transcription. However, it fails when unknown words occur. Neural networks, however, can provide a transcription for any character string, but may make significant mistakes. Continuing education the invention combines the security of the lexicon with the flexibility of neural networks.

Die Transkription der Teilwörter kann auf verschiedene Weise erfolgen, z.B. mittels einer Out-of-Vocabulary-Behandlung (OOV-Behandlung). Ein recht zuverlässiger Weg besteht darin, für das Wort in einer Datenbank, die phonetische Transkriptionen von Wörtern enthält, nach Teilwörtern zu suchen. Als Transkription wird dann für ein in der Datenbank gefundenes Teilwort die in der Datenbank verzeichnete phonetische Transkription gewählt. Dies führt für die meisten Wörter bzw. Teilwörter zu brauchbaren Ergebnissen.The transcription of the partial words can be done in different ways take place, e.g. with an out-of-vocabulary treatment (OOV treatment). A fairly reliable way is for the word in a database, the phonetic transcriptions of words contains to search for partial words. As Transcription is then made for one found in the database Partial word the phonetic recorded in the database Transcription chosen. For most words or Subwords for useful results.

Falls das Wort neben dem gefundenen Teilwort mindestens einen weiteren Bestandteil aufweist, der nicht in der Datenbank verzeichnet ist, kann dieser mittels einer OOV-Behandlung phonetisch transkribiert werden. Die OOV-Behandlung kann mittels eines statistischen Verfahrens, z.B. mittels eines Neuronalen Netzes, oder regelbasiert erfolgen. Vorteilhafterweise wird das Wort in Teilwörter einer gewissen Mindestlänge zerlegt, damit möglichst große Teilwörter gefunden werden und entsprechend wenig Nachbesserungen anfallen.If the word next to the found subword is at least one has another component that is not in the database can be registered using OOV treatment be transcribed phonetically. The OOV treatment can be done using a statistical method, e.g. by means of a neural Network, or rule-based. The word is advantageously divided into partial words of a certain Minimum length disassembled so that the largest possible partial words are found and there will be little reworking.

Weitere vorteilhafte Weiterbildungen der Erfindung sind in den Unteransprüchen gekennzeichnet.Further advantageous developments of the invention are in marked the subclaims.

Im folgenden wird die Erfindung anhand von Ausführungsbeispielen näher erläutert, die in den Figuren schematisch dargestellt sind. Im einzelnen zeigt:

Fig. 1
ein zur Graphem-Phonem-Konvertierung geeignetes Computersystem; und
Fig. 2
eine schematische Darstellung des erfindungsgemäßen Verfahrens.
The invention is explained in more detail below with the aid of exemplary embodiments which are shown schematically in the figures. In detail shows:
Fig. 1
a computer system suitable for grapheme-phoneme conversion; and
Fig. 2
a schematic representation of the method according to the invention.

Fig. 1 zeigt ein zur Graphem-Phonem-Konvertierung eines Worts geeignetes Computersystem. Dies weist einen Prozessor (processor, CPU) 20, einen Arbeitsspeicher (RAM) 21, einen Programmspeicher (programm memory, ROM) 22, einen Festplatten-Controller (hard disc controller, HDC) 23, der eine Festplatte (hard disk) 30 steuert, und einen Schnittstellen-Controller (I/O controller) 24 auf. Prozessor 20, Arbeitsspeicher 21, Programmspeicher 22, Festplatten-Controller 23 und Schnittstellen-Controller 24 sind über einen Bus, den CPU-Bus 25, zum Austausch von Daten und Befehlen miteinander gekoppelt. Ferner weist der Computer einen Ein-/Ausgabe-Bus (I/O Bus) 26 auf, der verschiedene Ein- und Ausgabeeinrichtungen mit dem Schnittstellen-Controller 24 koppelt. Zu den Ein- und Ausgabeeinrichtungen zählen z.B. eine allgemeine Ein- und Ausgabe-Schnittstelle (I/O interface) 27, eine Anzeigeeinrichtung (display) 28, eine Tastatur (keyboard) 29 und eine Maus 31.)1 shows a grapheme-phoneme conversion of a word suitable computer system. This assigns a processor (processor, CPU) 20, a working memory (RAM) 21, a program memory (program memory, ROM) 22, a hard disk controller (hard disc controller, HDC) 23, which is a hard drive (hard disk) 30 controls, and an interface controller (I / O controller) 24 on. Processor 20, memory 21, program memory 22, hard disk controller 23 and interface controllers 24 are on a bus that CPU bus 25, for exchanging data and commands with one another coupled. The computer also has an input / output bus (I / O bus) 26 on the various input and output devices couples with the interface controller 24. To the Input and output devices count e.g. a general Input and output interface (I / O interface) 27, a display device (display) 28, a keyboard (keyboard) 29 and a mouse 31.)

Betrachten wir als Beispiel für die Graphem-Phonem-Konvertierung das deutsche Wort "überflüssigerweise".Let us consider an example of the grapheme-phoneme conversion the German word "unnecessarily".

Zunächst wird versucht, das Wort in Teilwörter zu zerlegen, die Bestandteile eines Aussprache-Lexikons sind. Um die Anzahl der möglichen Zerlegungen auf ein sinnvolles Maß zu beschränken, wird für die gesuchten Bestandteile eine Mindestlänge vorgegeben. Für die deutsche Sprache haben sich 6 Buchstaben als Mindestlänge in der Praxis bewährt.First we try to break the word down into subwords, are the components of a pronunciation dictionary. By the number to limit the possible decompositions to a reasonable degree, becomes a minimum length for the components sought specified. There are 6 letters for the German language Proven as a minimum length in practice.

Alle gefundenen Bestandteile werden in einer verketteten Liste abgespeichert. Bei mehreren Möglichkeiten wird immer der längste Bestandteil bzw. der Pfad mit den längsten Bestandteilen verwendet.All components found are in a linked list stored. If there are several options, the longest component or the path with the longest components used.

Werden nicht alle Teile des Worts als Teilwörter im Aussprachelexikon gefunden, so werden die verbleibenden Lücken im bevorzugten Ausführungsbeispiel durch ein Neuronales Netz geschlossen. Im Gegensatz zur Standardanwendung des Neuronalen Netzes, bei der die Transkription für das ganze Wort erstellt werden muss, ist die Aufgabe beim Auffüllen der Lücken einfacher, weil zumindest der linke Phonemkontext als sicher angenommen werden kann, da er ja aus dem Aussprachelexikon stammt. Die Eingabe der vorhergehenden Phoneme stabilisiert somit die Ausgabe des Neuronalen Netzes für die zu füllende Lücke, da das zu generierende Phonem nicht nur von den Buchstaben, sondern auch vom vorhergehenden Phonem abhängt. Not all parts of the word are used as partial words in the pronunciation dictionary are found, the remaining gaps in the preferred embodiment closed by a neural network. In contrast to the standard application of the neural Network in which the transcription is created for the whole word the task of filling in the gaps is easier, because at least the left phoneme context is assumed to be safe can be, since it comes from the pronunciation dictionary comes. The input of the previous phonemes stabilized thus the output of the neural network for the one to be filled Gap, because the phoneme to be generated not only from the letters, but also depends on the previous phoneme.

Ein Problem beim Aneinanderhängen der Transkriptionen aus dem Lexikon sowie bei der Bestimmung der Transkription für die Lücken mittels eines Neuronalen Netzes besteht darin, daß in einigen Fällen der letzte Laut der vorhergehenden, linken Transkription verändert werden muss. Dies ist bei dem betrachteten Wort "überflüssigerweise" der Fall. Es wird im Lexikon als Ganzes nicht gefunden, dafür aber das Teilwort "überflüssig" und das Teilwort "erweise".A problem with the concatenation of the transcriptions from the Encyclopedia as well as in determining the transcription for the Gaps by means of a neural network is that in in some cases the last sound of the previous left Transcription needs to be changed. This is with the one under consideration Word "unnecessarily" the case. It will be in the lexicon not found as a whole, but the sub-word "superfluous" and the subword "prove".

Im Folgenden werden Grapheme zur besseren Unterscheidung in spitzen Klammern <> eingeschlossen und Phoneme in eckigen Klammern [].The following are graphemes for better differentiation in angle brackets <> enclosed and phonemes in square Brackets [].

Die Endung <-ig> am Silbenende wird gesprochen wie [IC], dargestellt in der Lautschrift SAMPA, also wie [I] (ungespannter kurzer ungerundeter vorderer Vokal) gefolgt vom Ich-Laut [C] (stimmloser palataler Frikativ). Die Vorsilbe <er-> wird gesprochen wie [Er], mit einem [E] (ungespannter kurzer ungerundeter halboffener vorderer Vokal, offenes "e") und einem [r] (zentraler Sonorant).The ending <-ig> at the end of the syllable is spoken like [IC] in the phonetic spelling SAMPA, so like [I] (untensioned short non-rounded front vowel) followed by the first sound [C] (voiceless palatal fricative). The prefix <er-> is spoken like [Er], with an [E] (untensioned short unrounded semi-open front vowel, open "e") and one [r] (central sonorant).

Beim einfachen Verketten der Transkriptionen wird sinnvollerweise automatisch eine Silbengrenze zwischen den beiden Wörtern eingefügt, dargestellt durch einen Bindestrich "-". Es ergibt sich somit als Gesamttranskription des Worts <überflüssigerweise> [y: - b6 - flY - sIC - Er - vaI - z@] statt richtigerweise [y: - b6 - flY - sI - g6 - vaI - z@] mit einem [g] (stimmhafter velarer Plosiv) und einem [6] (nichtbetonter zentraler halboffener Vokal mit velarer Färbung) sowie einer verschobenen Silbengrenze. Somit wären an der Wortgrenze Laut und Silbengrenze falsch.When the transcriptions are simply concatenated, it makes sense to automatically insert a syllable boundary between the two words, represented by a hyphen "-". It is thus the total transcription of the word <superfluous> [y: - b6 - flY - sIC - Er - vaI - z @] instead of right [y: - b6 - flY - sI - g6 - vaI - z @] with a [g] (voiced velar plosive) and a [6] (unstressed central semi-open vowel with velar coloring) and a shifted syllable boundary. Thus, at the word boundary, the sound and syllable boundary would be wrong.

Abhilfe kann hier geschaffen werden, indem ein Neuronales Netz den letzten Laut der linken Transkription berechnet. Dabei stellt sich aber die Frage, welche Buchstaben am Ende der linken Transkription zur Bestimmung des letzten Lautes herangezogen werden sollen.Remedy can be created here by using a neural Net calculated the last according to the left transcription. there the question arises, which letters at the end of the left transcription used to determine the last sound should be.

Für diese Entscheidung wird ein spezielles Aussprachelexikon benutzt. Die Besonderheit an diesem Lexikon besteht darin, daß es die Information enthält, welche Graphemgruppe zu welchem Laut gehört. Wie das Lexikon zu erstellen ist, ist in Horst-Udo Hain: "Automation of the Training Procedures for Neural Networks Performing Multi-Lingual Grapheme to Phoneme Conversion". Eurospeech 1999, S. 2087-2090, ausgeführt.A special pronunciation dictionary is used for this decision used. The special feature of this lexicon is that that it contains information about which grapheme group for which Heard out loud. How to create the lexicon is in Horst-Udo Hain: "Automation of the Training Procedures for Neural Networks Performing Multi-Lingual Grapheme to Phoneme Conversion ". Eurospeech 1999, pp. 2087-2090.

Der Eintrag für "überflüssig" hat in diesem Lexikon die Form ü - b er - f l ü - ss i g y: - b 6 - f l y - s I C The entry for "superfluous" has the form in this lexicon ü - b he - f l ü - ss i G y: - b 6 - f l y - s I C

Damit kann eindeutig bestimmt werden, aus welcher Graphemgruppe der letzte Laut entstanden ist, nämlich aus dem <g>.It can be used to clearly determine from which grapheme group the last sound emerged, namely from the <g>.

Das Neuronale Netz kann nun mit Hilfe des jetzt vorhandenen rechten Kontextes <erweise> neu über Phonem und Silbengrenze am Wortende entscheiden. Das Ergebnis ist in diesem Falle das Phonem [g], vor dem eine Silbengrenze gesetzt wird. The neural network can now use the existing one right context <wise> new about phoneme and syllable boundary decide at the end of the word. The result in this case is that Phoneme [g] in front of which a syllable limit is set.

Jetzt ist die Silbengrenze an der richtigen Stelle und das <g> wird auch als [g] transkribiert und nicht als [C].
Der erste Laut der rechten Transkription wird nach dem gleichen Schema neu bestimmt. Die richtige Transkription für <er-> von <erweise> ist an dieser Stelle [6] und nicht [Er]. Hier sind gleich zwei Laute zu revidieren, weshalb im bevorzugten Ausführungsbeispiel stets zwei Laute revidiert werden.Now the syllable boundary is in the right place and the <g> is also transcribed as [g] and not as [C].
The first according to the right transcription is redetermined using the same scheme. The correct transcription for <er-> from <weise> at this point is [6] and not [Er]. Two sounds have to be revised here, which is why in the preferred embodiment two sounds are always revised.

Im Ergebnis erhält man die korrekte phonetische Transkription an dieser Schnittstelle.The result is the correct phonetic transcription at this interface.

Weitere Verbesserungen sind zu erzielen, wenn für das Ausfüllen der Transkriptionslücken nicht das Standardnetz verwendet wird, das zur Konvertierung ganzer Wörter trainiert wurde, sondern ein speziell zum Ausfüllen der Lücken trainiertes Netz. Zumindest in den Fällen, bei denen der rechte Phonemkontext auch vorhanden ist, bietet sich ein Spezialnetz an, das unter Verwendung des rechten Phonemkontextes über den zu generierenden Laut entscheidet.Further improvements can be achieved when filling in the transcription gaps do not use the standard network who has been trained to convert whole words, but one specially trained to fill in the gaps Network. At least in those cases where the right phoneme context a special network is also available, that using the right phoneme context over the to generating sound decides.

Claims (11)

Verfahren zur Graphem-Phonem-Konvertierung eines Wortes, das als Ganzes nicht in einem Aussprachelexikon enthalten ist, mit folgenden Schritten: a) das Wort wird in Teilwörter zerlegt, b) eine Graphem-Phonem-Konvertierung der Teilwörter wird durchgeführt, c) die durch die Konvertierung erhaltenen Transkriptionen der Teilwörter werden hintereinander aufgereiht, wobei sich mindestens eine Schnittstelle zwischen den Transkriptionen der Teilwörter ergibt, d) die an die mindestens eine Schnittstelle grenzenden Phoneme der Teilwörter werden bestimmt, e) es werden diejenigen Grapheme der Teilwörter bestimmt, die die an die mindestens eine Schnittstelle grenzenden Phoneme erzeugen, f) die Graphem-Phonem-Konvertierung der bestimmten Grapheme wird im Kontext der jeweiligen Schnittstelle neu berechnet. A method for converting a grapheme-phoneme of a word that is not contained as a whole in a pronunciation dictionary, with the following steps: a) the word is broken down into partial words, b) a grapheme-phoneme conversion of the partial words is carried out, c) the transcriptions of the partial words obtained by the conversion are lined up in succession, with at least one interface between the transcriptions of the partial words, d) the phonemes of the partial words bordering on the at least one interface are determined, e) those graphemes of the partial words are determined which generate the phonemes bordering on the at least one interface, f) the grapheme-phoneme conversion of the particular grapheme is recalculated in the context of the respective interface. Verfahren nach Anspruch 1,
dadurch gekennzeichnet, dass die Graphem-Phonem-Konvertierung der bestimmten Grapheme im Kontext der jeweiligen Schnittstelle mittels eines Neuronalen Netzes neu berechnet werden.Method according to claim 1,
characterized in that the grapheme-phoneme conversion of the specific graphemes in the context of the respective interface is recalculated using a neural network. Verfahren nach Anspruch 1,
dadurch gekennzeichnet, dass die Graphem-Phonem-Konvertierung der bestimmten Grapheme im Kontext der jeweiligen Schnittstelle mittels eines Lexikons neu berechnet werden. Method according to claim 1,
characterized in that the grapheme-phoneme conversion of the specific graphemes are recalculated in the context of the respective interface by means of a lexicon. Verfahren nach einem der vorhergehenden Ansprüche,
dadurch gekennzeichnet, dass für das Wort in einer Datenbank, die phonetische Transkriptionen von Wörtern enthält, nach Teilwörtern des Worts gesucht wird; und- dass für ein in der Datenbank gefundenes Teilwort die in der Datenbank verzeichnete phonetische Transkription gewählt wird.Method according to one of the preceding claims,
characterized in that for the word in a database containing phonetic transcriptions of words, partial words of the word are searched; and that the phonetic transcription recorded in the database is selected for a partial word found in the database. Verfahren nach Anspruch 4,
dadurch gekennzeichnet, dass das Wort neben dem gefundenen Teilwort mindestens einen weiteren Bestandteil aufweist, der nicht in der Datenbank verzeichnet ist; und- dass dieser weitere Bestandteil mittels einer OOV-Behandlung phonetisch transkribiert wird.Method according to claim 4,
characterized in that the word has at least one further component in addition to the found subword which is not recorded in the database; and - that this additional component is transcribed phonetically using OOV treatment. Verfahren nach einem der vorhergehenden Ansprüche,
dadurch gekennzeichnet, dass das Wort in Teilwörter einer gewissen Mindestlänge zerlegt wird.Method according to one of the preceding claims,
characterized in that the word is broken down into partial words of a certain minimum length. Computerprogrammprodukt, das durch einen Computer ausführbar ist und dabei die Schritte nach einem der Ansprüche 1 bis 6 ausführt.Computer program product executable by a computer is and the steps according to one of claims 1 to 6 executes. Computerprogrammprodukt, das auf einem computergeeigneten Medium gespeichert ist und computerlesbare Programmmittel umfaßt, die es einem Computer ermöglichen, das Verfahren nach einem der Ansprüche 1 bis 6 auszuführen.Computer program product that is based on a computer Medium is stored and includes computer-readable program means, which allow a computer to follow the procedure to perform one of claims 1 to 6. Datenträger, auf dem ein Computerprogramm gespeichert ist, das es einem Computer ermöglicht, durch einen Ladeprozess das Verfahren nach einem der Ansprüche 1 bis 6 auszuführen. Data medium on which a computer program is stored, that enables a computer to load through a loading process Execute method according to one of claims 1 to 6. Computersystem mit Mitteln zum Ausführen des Verfahrens nach einem der Ansprüche 1 bis 6.Computer system with means for performing the method according to one of claims 1 to 6. Computersystem zur Graphem-Phonem-Konvertierung eines Worts, das als Ganzes nicht in einem Aussprachelexikon enthalten ist, einer Speichereinrichtung (22, 30) zum Speichern eines Computerprogramms auf einem Speichermedium; einer Verarbeitungseinheit (20) zum Laden des Computerprogramms aus der Speichereinrichtung und zum Ausführen des Computerprogramms; mit Mitteln zum Zerlegen des Worts in Teilwörter; mit Mitteln zum hintereinander Aufreihen der Transkriptionen der Teilwörter, wobei sich mindestens eine Schnittstelle zwischen den Transkriptionen der Teilwörter ergibt; mit Mitteln zum Bestimmen der an die mindestens eine Schnittstelle grenzenden Phoneme der Teilwörter; mit Mitteln zum Bestimmen derjenigen Grapheme der Teilwörter, die die an die mindestens eine Schnittstelle grenzenden Phoneme erzeugen; mit Mitteln zum erneuten Berechnen der Graphem-Phonem-Konvertierung der bestimmten Grapheme im Kontext der jeweiligen Schnittstelle; und mit Mitteln zum anschließenden Schreiben der an der Schnittstelle neu berechneten Phoneme in eine zweite Speichereinrichtung. Computer system for grapheme-phoneme conversion of a word that is not contained as a whole in a pronunciation dictionary, a storage device (22, 30) for storing a computer program on a storage medium; a processing unit (20) for loading the computer program from the storage device and for executing the computer program; with means for dividing the word into partial words; with means for sequencing the transcriptions of the partial words, whereby there is at least one interface between the transcriptions of the partial words; with means for determining the phonemes of the partial words bordering on the at least one interface; with means for determining those graphemes of the partial words which generate the phonemes bordering on the at least one interface; with means for recalculating the grapheme-phoneme conversion of the particular grapheme in the context of the respective interface; and with means for subsequently writing the phonemes recalculated at the interface into a second memory device.
EP01117869A 2000-08-31 2001-07-23 Grapheme-phoneme conversion Expired - Lifetime EP1184839B1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
DE10042944 2000-08-31
DE10042944A DE10042944C2 (en) 2000-08-31 2000-08-31 Grapheme-phoneme conversion

Publications (3)

Publication Number Publication Date
EP1184839A2 true EP1184839A2 (en) 2002-03-06
EP1184839A3 EP1184839A3 (en) 2003-02-05
EP1184839B1 EP1184839B1 (en) 2005-09-28

Family

ID=7654523

Family Applications (1)

Application Number Title Priority Date Filing Date
EP01117869A Expired - Lifetime EP1184839B1 (en) 2000-08-31 2001-07-23 Grapheme-phoneme conversion

Country Status (3)

Country Link
US (1) US7107216B2 (en)
EP (1) EP1184839B1 (en)
DE (2) DE10042944C2 (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1315108C (en) * 2004-03-17 2007-05-09 财团法人工业技术研究院 Method for converting words to phonetic symbols by regrading mistakable grapheme to improve accuracy rate
CN105590623A (en) * 2016-02-24 2016-05-18 百度在线网络技术（北京）有限公司 Letter-to-phoneme conversion model generating method and letter-to-phoneme conversion generating device based on artificial intelligence

Families Citing this family (178)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8645137B2 (en) 2000-03-16 2014-02-04 Apple Inc. Fast, language-independent method for user authentication by voice
DE10042942C2 (en) * 2000-08-31 2003-05-08 Siemens Ag Speech synthesis method
ITFI20010199A1 (en) 2001-10-22 2003-04-22 Riccardo Vieri SYSTEM AND METHOD TO TRANSFORM TEXTUAL COMMUNICATIONS INTO VOICE AND SEND THEM WITH AN INTERNET CONNECTION TO ANY TELEPHONE SYSTEM
US7353164B1 (en) * 2002-09-13 2008-04-01 Apple Inc. Representation of orthography in a continuous vector space
US7047193B1 (en) 2002-09-13 2006-05-16 Apple Computer, Inc. Unsupervised data-driven pronunciation modeling
US8285537B2 (en) * 2003-01-31 2012-10-09 Comverse, Inc. Recognition of proper nouns using native-language pronunciation
JP4001283B2 (en) * 2003-02-12 2007-10-31 インターナショナル・ビジネス・マシーンズ・コーポレーション Morphological analyzer and natural language processor
US8032377B2 (en) * 2003-04-30 2011-10-04 Loquendo S.P.A. Grapheme to phoneme alignment method and relative rule-set generating system
US7280963B1 (en) * 2003-09-12 2007-10-09 Nuance Communications, Inc. Method for learning linguistically valid word pronunciations from acoustic data
US20050108013A1 (en) * 2003-11-13 2005-05-19 International Business Machines Corporation Phonetic coverage interactive tool
TWI233589B (en) * 2004-03-05 2005-06-01 Ind Tech Res Inst Method for text-to-pronunciation conversion capable of increasing the accuracy by re-scoring graphemes likely to be tagged erroneously
JP4328698B2 (en) * 2004-09-15 2009-09-09 キヤノン株式会社 Fragment set creation method and apparatus
TWI250509B (en) * 2004-10-05 2006-03-01 Inventec Corp Speech-synthesizing system and method thereof
US20060259301A1 (en) * 2005-05-12 2006-11-16 Nokia Corporation High quality thai text-to-phoneme converter
US8677377B2 (en) 2005-09-08 2014-03-18 Apple Inc. Method and apparatus for building an intelligent automated assistant
US7633076B2 (en) 2005-09-30 2009-12-15 Apple Inc. Automated response to and sensing of user activity in portable devices
TWI340330B (en) * 2005-11-14 2011-04-11 Ind Tech Res Inst Method for text-to-pronunciation conversion
US9318108B2 (en) 2010-01-18 2016-04-19 Apple Inc. Intelligent automated assistant
US8135590B2 (en) 2007-01-11 2012-03-13 Microsoft Corporation Position-dependent phonetic models for reliable pronunciation identification
US8977255B2 (en) 2007-04-03 2015-03-10 Apple Inc. Method and system for operating a multi-function portable electronic device using voice-activation
US9053089B2 (en) 2007-10-02 2015-06-09 Apple Inc. Part-of-speech tagging using latent analogy
US8620662B2 (en) 2007-11-20 2013-12-31 Apple Inc. Context-aware unit selection
US7991615B2 (en) * 2007-12-07 2011-08-02 Microsoft Corporation Grapheme-to-phoneme conversion using acoustic data
US10002189B2 (en) 2007-12-20 2018-06-19 Apple Inc. Method and apparatus for searching using an active ontology
US9330720B2 (en) 2008-01-03 2016-05-03 Apple Inc. Methods and apparatus for altering audio output signals
US8065143B2 (en) 2008-02-22 2011-11-22 Apple Inc. Providing text input using speech data and non-speech data
US8996376B2 (en) 2008-04-05 2015-03-31 Apple Inc. Intelligent text-to-speech conversion
US10496753B2 (en) 2010-01-18 2019-12-03 Apple Inc. Automatically adapting user interfaces for hands-free interaction
US8464150B2 (en) 2008-06-07 2013-06-11 Apple Inc. Automatic language identification for dynamic text processing
US20100030549A1 (en) 2008-07-31 2010-02-04 Lee Michael M Mobile device having human language translation capability with positional feedback
US8768702B2 (en) 2008-09-05 2014-07-01 Apple Inc. Multi-tiered voice feedback in an electronic device
US8898568B2 (en) 2008-09-09 2014-11-25 Apple Inc. Audio user interface
US8712776B2 (en) 2008-09-29 2014-04-29 Apple Inc. Systems and methods for selective text to speech synthesis
US8583418B2 (en) 2008-09-29 2013-11-12 Apple Inc. Systems and methods of detecting language and natural language strings for text to speech synthesis
US8676904B2 (en) 2008-10-02 2014-03-18 Apple Inc. Electronic devices with voice command and contextual data processing capabilities
US9959870B2 (en) 2008-12-11 2018-05-01 Apple Inc. Speech recognition involving a mobile device
US8862252B2 (en) 2009-01-30 2014-10-14 Apple Inc. Audio user interface for displayless electronic device
US8788256B2 (en) * 2009-02-17 2014-07-22 Sony Computer Entertainment Inc. Multiple language voice recognition
US8380507B2 (en) 2009-03-09 2013-02-19 Apple Inc. Systems and methods for determining the language to use for speech generated by a text to speech engine
US10241644B2 (en) 2011-06-03 2019-03-26 Apple Inc. Actionable reminder entries
US10540976B2 (en) 2009-06-05 2020-01-21 Apple Inc. Contextual voice commands
US10241752B2 (en) 2011-09-30 2019-03-26 Apple Inc. Interface for a virtual digital assistant
US9858925B2 (en) 2009-06-05 2018-01-02 Apple Inc. Using context information to facilitate processing of commands in a virtual assistant
US10706373B2 (en) 2011-06-03 2020-07-07 Apple Inc. Performing actions associated with task items that represent tasks to perform
US9431006B2 (en) 2009-07-02 2016-08-30 Apple Inc. Methods and apparatuses for automatic speech recognition
US8682649B2 (en) 2009-11-12 2014-03-25 Apple Inc. Sentiment prediction from textual data
US8600743B2 (en) 2010-01-06 2013-12-03 Apple Inc. Noise profile determination for voice-related feature
US8381107B2 (en) 2010-01-13 2013-02-19 Apple Inc. Adaptive audio feedback system and method
US8311838B2 (en) 2010-01-13 2012-11-13 Apple Inc. Devices and methods for identifying a prompt corresponding to a voice input in a sequence of prompts
US10705794B2 (en) 2010-01-18 2020-07-07 Apple Inc. Automatically adapting user interfaces for hands-free interaction
US10276170B2 (en) 2010-01-18 2019-04-30 Apple Inc. Intelligent automated assistant
US10679605B2 (en) 2010-01-18 2020-06-09 Apple Inc. Hands-free list-reading by intelligent automated assistant
US10553209B2 (en) 2010-01-18 2020-02-04 Apple Inc. Systems and methods for hands-free notification summaries
WO2011089450A2 (en) 2010-01-25 2011-07-28 Andrew Peter Nelson Jerram Apparatuses, methods and systems for a digital conversation management platform
US8682667B2 (en) 2010-02-25 2014-03-25 Apple Inc. User profiling for selecting user specific voice input processing information
US8713021B2 (en) 2010-07-07 2014-04-29 Apple Inc. Unsupervised document clustering using latent semantic density analysis
US8719006B2 (en) 2010-08-27 2014-05-06 Apple Inc. Combined statistical and rule-based part-of-speech tagging for text-to-speech synthesis
US8719014B2 (en) 2010-09-27 2014-05-06 Apple Inc. Electronic device with text error correction based on voice recognition data
US10515147B2 (en) 2010-12-22 2019-12-24 Apple Inc. Using statistical language models for contextual lookup
US10762293B2 (en) 2010-12-22 2020-09-01 Apple Inc. Using parts-of-speech tagging and named entity recognition for spelling correction
US8781836B2 (en) 2011-02-22 2014-07-15 Apple Inc. Hearing assistance system for providing consistent human speech
US9262612B2 (en) 2011-03-21 2016-02-16 Apple Inc. Device access using voice authentication
US10057736B2 (en) 2011-06-03 2018-08-21 Apple Inc. Active transport based notifications
US10672399B2 (en) 2011-06-03 2020-06-02 Apple Inc. Switching between text data and audio data based on a mapping
US8812294B2 (en) 2011-06-21 2014-08-19 Apple Inc. Translating phrases from one language into another using an order-based set of declarative rules
US8706472B2 (en) 2011-08-11 2014-04-22 Apple Inc. Method for disambiguating multiple readings in language conversion
US8994660B2 (en) 2011-08-29 2015-03-31 Apple Inc. Text correction processing
US8762156B2 (en) 2011-09-28 2014-06-24 Apple Inc. Speech recognition repair using contextual information
US10134385B2 (en) 2012-03-02 2018-11-20 Apple Inc. Systems and methods for name pronunciation
US9483461B2 (en) 2012-03-06 2016-11-01 Apple Inc. Handling speech synthesis of content for multiple languages
US9280610B2 (en) 2012-05-14 2016-03-08 Apple Inc. Crowd sourcing information to fulfill user requests
US8775442B2 (en) 2012-05-15 2014-07-08 Apple Inc. Semantic search using a single-source semantic model
US10417037B2 (en) 2012-05-15 2019-09-17 Apple Inc. Systems and methods for integrating third party services with a digital assistant
WO2013185109A2 (en) 2012-06-08 2013-12-12 Apple Inc. Systems and methods for recognizing textual identifiers within a plurality of words
US9721563B2 (en) 2012-06-08 2017-08-01 Apple Inc. Name recognition system
US9495129B2 (en) 2012-06-29 2016-11-15 Apple Inc. Device, method, and user interface for voice-activated navigation and browsing of a document
US9576574B2 (en) 2012-09-10 2017-02-21 Apple Inc. Context-sensitive handling of interruptions by intelligent digital assistant
US9547647B2 (en) 2012-09-19 2017-01-17 Apple Inc. Voice-based media searching
US8935167B2 (en) 2012-09-25 2015-01-13 Apple Inc. Exemplar-based latent perceptual modeling for automatic speech recognition
EP2954514B1 (en) 2013-02-07 2021-03-31 Apple Inc. Voice trigger for a digital assistant
US10652394B2 (en) 2013-03-14 2020-05-12 Apple Inc. System and method for processing voicemail
US9733821B2 (en) 2013-03-14 2017-08-15 Apple Inc. Voice control to diagnose inadvertent activation of accessibility features
US10642574B2 (en) 2013-03-14 2020-05-05 Apple Inc. Device, method, and graphical user interface for outputting captions
US9368114B2 (en) 2013-03-14 2016-06-14 Apple Inc. Context-sensitive handling of interruptions
US10572476B2 (en) 2013-03-14 2020-02-25 Apple Inc. Refining a search based on schedule items
US9977779B2 (en) 2013-03-14 2018-05-22 Apple Inc. Automatic supplementation of word correction dictionaries
US10748529B1 (en) 2013-03-15 2020-08-18 Apple Inc. Voice activated device for use with a voice-based digital assistant
AU2014233517B2 (en) 2013-03-15 2017-05-25 Apple Inc. Training an at least partial voice command system
CN105144133B (en) 2013-03-15 2020-11-20 苹果公司 Context-sensitive handling of interrupts
WO2014144579A1 (en) 2013-03-15 2014-09-18 Apple Inc. System and method for updating an adaptive speech recognition model
WO2014144395A2 (en) 2013-03-15 2014-09-18 Apple Inc. User training by intelligent digital assistant
WO2014197336A1 (en) 2013-06-07 2014-12-11 Apple Inc. System and method for detecting errors in interactions with a voice-based digital assistant
WO2014197334A2 (en) 2013-06-07 2014-12-11 Apple Inc. System and method for user-specified pronunciation of words for speech synthesis and recognition
US9582608B2 (en) 2013-06-07 2017-02-28 Apple Inc. Unified ranking with entropy-weighted information for phrase-based semantic auto-completion
WO2014197335A1 (en) 2013-06-08 2014-12-11 Apple Inc. Interpreting and acting upon commands that involve sharing information with remote devices
DE112014002747T5 (en) 2013-06-09 2016-03-03 Apple Inc. Apparatus, method and graphical user interface for enabling conversation persistence over two or more instances of a digital assistant
US10176167B2 (en) 2013-06-09 2019-01-08 Apple Inc. System and method for inferring user intent from speech inputs
KR101809808B1 (en) 2013-06-13 2017-12-15 애플 인크. System and method for emergency calls initiated by voice command
AU2014306221B2 (en) 2013-08-06 2017-04-06 Apple Inc. Auto-activating smart responses based on activities from remote devices
US10296160B2 (en) 2013-12-06 2019-05-21 Apple Inc. Method for extracting salient dialog usage from live data
US9620105B2 (en) 2014-05-15 2017-04-11 Apple Inc. Analyzing audio input for efficient speech and music recognition
US10592095B2 (en) 2014-05-23 2020-03-17 Apple Inc. Instantaneous speaking of content on touch devices
US9502031B2 (en) 2014-05-27 2016-11-22 Apple Inc. Method for supporting dynamic grammars in WFST-based ASR
US9633004B2 (en) 2014-05-30 2017-04-25 Apple Inc. Better resolution when referencing to concepts
US9430463B2 (en) 2014-05-30 2016-08-30 Apple Inc. Exemplar-based natural language processing
AU2015266863B2 (en) 2014-05-30 2018-03-15 Apple Inc. Multi-command single utterance input method
US9734193B2 (en) 2014-05-30 2017-08-15 Apple Inc. Determining domain salience ranking from ambiguous words in natural speech
US10078631B2 (en) 2014-05-30 2018-09-18 Apple Inc. Entropy-guided text prediction using combined word and character n-gram language models
US10289433B2 (en) 2014-05-30 2019-05-14 Apple Inc. Domain specific language for encoding assistant dialog
US9715875B2 (en) 2014-05-30 2017-07-25 Apple Inc. Reducing the need for manual start/end-pointing and trigger phrases
US9785630B2 (en) 2014-05-30 2017-10-10 Apple Inc. Text prediction using combined word N-gram and unigram language models
US9842101B2 (en) 2014-05-30 2017-12-12 Apple Inc. Predictive conversion of language input
US10170123B2 (en) 2014-05-30 2019-01-01 Apple Inc. Intelligent assistant for home automation
US9760559B2 (en) 2014-05-30 2017-09-12 Apple Inc. Predictive text input
US9338493B2 (en) 2014-06-30 2016-05-10 Apple Inc. Intelligent automated assistant for TV user interactions
US10659851B2 (en) 2014-06-30 2020-05-19 Apple Inc. Real-time digital assistant knowledge updates
US10446141B2 (en) 2014-08-28 2019-10-15 Apple Inc. Automatic speech recognition based on user feedback
US9818400B2 (en) 2014-09-11 2017-11-14 Apple Inc. Method and apparatus for discovering trending terms in speech requests
US10789041B2 (en) 2014-09-12 2020-09-29 Apple Inc. Dynamic thresholds for always listening speech trigger
US9646609B2 (en) 2014-09-30 2017-05-09 Apple Inc. Caching apparatus for serving phonetic pronunciations
US9886432B2 (en) 2014-09-30 2018-02-06 Apple Inc. Parsimonious handling of word inflection via categorical stem + suffix N-gram language models
US10127911B2 (en) 2014-09-30 2018-11-13 Apple Inc. Speaker identification and unsupervised speaker adaptation techniques
US10074360B2 (en) 2014-09-30 2018-09-11 Apple Inc. Providing an indication of the suitability of speech recognition
US9668121B2 (en) 2014-09-30 2017-05-30 Apple Inc. Social reminders
US10552013B2 (en) 2014-12-02 2020-02-04 Apple Inc. Data detection
US9711141B2 (en) 2014-12-09 2017-07-18 Apple Inc. Disambiguating heteronyms in speech synthesis
US9865280B2 (en) 2015-03-06 2018-01-09 Apple Inc. Structured dictation using intelligent automated assistants
US9886953B2 (en) 2015-03-08 2018-02-06 Apple Inc. Virtual assistant activation
US10567477B2 (en) 2015-03-08 2020-02-18 Apple Inc. Virtual assistant continuity
US9721566B2 (en) 2015-03-08 2017-08-01 Apple Inc. Competing devices responding to voice triggers
US9899019B2 (en) 2015-03-18 2018-02-20 Apple Inc. Systems and methods for structured stem and suffix language models
US9842105B2 (en) 2015-04-16 2017-12-12 Apple Inc. Parsimonious continuous-space phrase representations for natural language processing
US10083688B2 (en) 2015-05-27 2018-09-25 Apple Inc. Device voice control for selecting a displayed affordance
US10127220B2 (en) 2015-06-04 2018-11-13 Apple Inc. Language identification from short strings
US10101822B2 (en) 2015-06-05 2018-10-16 Apple Inc. Language input correction
US9578173B2 (en) 2015-06-05 2017-02-21 Apple Inc. Virtual assistant aided communication with 3rd party service in a communication session
US10186254B2 (en) 2015-06-07 2019-01-22 Apple Inc. Context-based endpoint detection
US11025565B2 (en) 2015-06-07 2021-06-01 Apple Inc. Personalized prediction of responses for instant messaging
US10255907B2 (en) 2015-06-07 2019-04-09 Apple Inc. Automatic accent detection using acoustic models
US10671428B2 (en) 2015-09-08 2020-06-02 Apple Inc. Distributed personal assistant
US10747498B2 (en) 2015-09-08 2020-08-18 Apple Inc. Zero latency digital assistant
US9697820B2 (en) 2015-09-24 2017-07-04 Apple Inc. Unit-selection text-to-speech synthesis using concatenation-sensitive neural networks
US11010550B2 (en) 2015-09-29 2021-05-18 Apple Inc. Unified language modeling framework for word prediction, auto-completion and auto-correction
US10366158B2 (en) 2015-09-29 2019-07-30 Apple Inc. Efficient word encoding for recurrent neural network language models
US11587559B2 (en) 2015-09-30 2023-02-21 Apple Inc. Intelligent device identification
US10691473B2 (en) 2015-11-06 2020-06-23 Apple Inc. Intelligent automated assistant in a messaging environment
US10049668B2 (en) 2015-12-02 2018-08-14 Apple Inc. Applying neural network language models to weighted finite state transducers for automatic speech recognition
US9910836B2 (en) * 2015-12-21 2018-03-06 Verisign, Inc. Construction of phonetic representation of a string of characters
US10102189B2 (en) * 2015-12-21 2018-10-16 Verisign, Inc. Construction of a phonetic representation of a generated string of characters
US10102203B2 (en) * 2015-12-21 2018-10-16 Verisign, Inc. Method for writing a foreign language in a pseudo language phonetically resembling native language of the speaker
US9947311B2 (en) 2015-12-21 2018-04-17 Verisign, Inc. Systems and methods for automatic phonetization of domain names
US10223066B2 (en) 2015-12-23 2019-03-05 Apple Inc. Proactive assistance based on dialog communication between devices
US10446143B2 (en) 2016-03-14 2019-10-15 Apple Inc. Identification of voice inputs providing credentials
US9934775B2 (en) 2016-05-26 2018-04-03 Apple Inc. Unit-selection text-to-speech synthesis based on predicted concatenation parameters
US9972304B2 (en) 2016-06-03 2018-05-15 Apple Inc. Privacy preserving distributed evaluation framework for embedded personalized systems
US10249300B2 (en) 2016-06-06 2019-04-02 Apple Inc. Intelligent list reading
US10049663B2 (en) 2016-06-08 2018-08-14 Apple, Inc. Intelligent automated assistant for media exploration
DK179309B1 (en) 2016-06-09 2018-04-23 Apple Inc Intelligent automated assistant in a home environment
US10490187B2 (en) 2016-06-10 2019-11-26 Apple Inc. Digital assistant providing automated status report
US10192552B2 (en) 2016-06-10 2019-01-29 Apple Inc. Digital assistant providing whispered speech
US10509862B2 (en) 2016-06-10 2019-12-17 Apple Inc. Dynamic phrase expansion of language input
US10067938B2 (en) 2016-06-10 2018-09-04 Apple Inc. Multilingual word prediction
US10586535B2 (en) 2016-06-10 2020-03-10 Apple Inc. Intelligent digital assistant in a multi-tasking environment
DK179343B1 (en) 2016-06-11 2018-05-14 Apple Inc Intelligent task discovery
DK201670540A1 (en) 2016-06-11 2018-01-08 Apple Inc Application integration with a digital assistant
DK179415B1 (en) 2016-06-11 2018-06-14 Apple Inc Intelligent device arbitration and control
DK179049B1 (en) 2016-06-11 2017-09-18 Apple Inc Data driven natural language event detection and classification
US10043516B2 (en) 2016-09-23 2018-08-07 Apple Inc. Intelligent automated assistant
US10593346B2 (en) 2016-12-22 2020-03-17 Apple Inc. Rank-reduced token representation for automatic speech recognition
DK201770439A1 (en) 2017-05-11 2018-12-13 Apple Inc. Offline personal assistant
DK179496B1 (en) 2017-05-12 2019-01-15 Apple Inc. USER-SPECIFIC Acoustic Models
DK179745B1 (en) 2017-05-12 2019-05-01 Apple Inc. SYNCHRONIZATION AND TASK DELEGATION OF A DIGITAL ASSISTANT
DK201770432A1 (en) 2017-05-15 2018-12-21 Apple Inc. Hierarchical belief states for digital assistants
DK201770431A1 (en) 2017-05-15 2018-12-20 Apple Inc. Optimizing dialogue policy decisions for digital assistants using implicit feedback
DK179549B1 (en) 2017-05-16 2019-02-12 Apple Inc. Far-field extension for digital assistant services
US11195513B2 (en) * 2017-09-27 2021-12-07 International Business Machines Corporation Generating phonemes of loan words using two converters
CN112487797B (en) * 2020-11-26 2024-04-05 北京有竹居网络技术有限公司 Data generation method and device, readable medium and electronic equipment
CN113707131B (en) * 2021-08-30 2024-04-16 中国科学技术大学 Speech recognition method, device, equipment and storage medium

Family Cites Families (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
ES2139066T3 (en) * 1993-03-26 2000-02-01 British Telecomm CONVERSION OF TEXT TO A WAVE FORM.
US5651095A (en) * 1993-10-04 1997-07-22 British Telecommunications Public Limited Company Speech synthesis using word parser with knowledge base having dictionary of morphemes with binding properties and combining rules to identify input word class
AU3734395A (en) * 1994-10-03 1996-04-26 Helfgott & Karas, P.C. A database accessing system
DE4440598C1 (en) * 1994-11-14 1996-05-23 Siemens Ag World Wide Web hypertext information highway navigator controlled by spoken word
DE19500494C2 (en) * 1995-01-10 1997-01-23 Siemens Ag Feature extraction method for a speech signal
DE19636739C1 (en) * 1996-09-10 1997-07-03 Siemens Ag Multi-lingual hidden Markov model application for speech recognition system
DE19719381C1 (en) * 1997-05-07 1998-01-22 Siemens Ag Computer based speech recognition method
US5913194A (en) * 1997-07-14 1999-06-15 Motorola, Inc. Method, device and system for using statistical information to reduce computation and memory requirements of a neural network based speech synthesis system
US6108627A (en) * 1997-10-31 2000-08-22 Nortel Networks Corporation Automatic transcription tool
US6076060A (en) * 1998-05-01 2000-06-13 Compaq Computer Corporation Computer method and apparatus for translating text to sound
US6411932B1 (en) * 1998-06-12 2002-06-25 Texas Instruments Incorporated Rule-based learning of word pronunciations from training corpora
US6188984B1 (en) * 1998-11-17 2001-02-13 Fonix Corporation Method and system for syllable parsing
US6208968B1 (en) * 1998-12-16 2001-03-27 Compaq Computer Corporation Computer method and apparatus for text-to-speech synthesizer dictionary reduction

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
HAIN H: "A Hybride Approach for Grapheme-to-Phoneme Conversion Based on a Combination of Partial String Matching and a Neural Network" INT. CONFERENCE ON SPOKEN LANGUAGE PROCESSING (ICSLP 2000), ISCA, Bd. III, 16. - 20. Oktober 2000, Seiten 291-294, XP002224174 Peking, China *
HAIN H: "Automation of the Training Procedures for Neural Networks Performing Multi-Lingual Grapheme-to-Phoneme Conversion" PROC. EUROSPEECH '99, Bd. 5, 6. - 9. September 1999, Seiten 2087-2090, XP002223264 Budapest, Ungarn *
HAIN H: "Ein hybrider Ansatz zur Graphem-Phonem-Konvertierung unter Verwendung eines Lexikons und eines neuronalen Netzes" ELEKTRONISCHE SPRACHVERARBEITUNG, ELFTE KONFERENZ, TAGUNGSBAND, 4. - 6. September 2000, Seiten 160-167, XP002223265 Cottbus *
KIM B, LEE W, LEE G, LEE J H: "Unlimited Vocabulary Grapheme to Phoneme Conversion for Korean TTS" 36TH ANNUAL MEETING IF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS & 17TH INT. CONFERENCE ON COMPUTATIONAL LINGUISTICS (COLING-ACL '98), 10. - 14. August 1998, Seiten 675-679, XP002224173 Montréal, Kanada *
MANNELL R, CLARK J E: "Text-to-Speech Rule and Dictionary Development" SPEECH COMMUNICATION, Bd. 6, Nr. 4, Dezember 1987 (1987-12), Seiten 317-324, AMSTERDAM, THE NETHERLANDS *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1315108C (en) * 2004-03-17 2007-05-09 财团法人工业技术研究院 Method for converting words to phonetic symbols by regrading mistakable grapheme to improve accuracy rate
CN105590623A (en) * 2016-02-24 2016-05-18 百度在线网络技术（北京）有限公司 Letter-to-phoneme conversion model generating method and letter-to-phoneme conversion generating device based on artificial intelligence
CN105590623B (en) * 2016-02-24 2019-07-30 百度在线网络技术（北京）有限公司 Letter phoneme transformation model generation method and device based on artificial intelligence

Also Published As

Publication number Publication date
EP1184839B1 (en) 2005-09-28
US20020046025A1 (en) 2002-04-18
EP1184839A3 (en) 2003-02-05
DE10042944A1 (en) 2002-03-21
DE10042944C2 (en) 2003-03-13
US7107216B2 (en) 2006-09-12
DE50107556D1 (en) 2005-11-03

Similar Documents

Publication Publication Date Title
EP1184839B1 (en) Grapheme-phoneme conversion
DE69937176T2 (en) Segmentation method to extend the active vocabulary of speech recognizers
DE60201262T2 (en) HIERARCHICAL LANGUAGE MODELS
DE3242866C2 (en)
DE69832393T2 (en) LANGUAGE RECOGNITION SYSTEM FOR THE DETECTION OF CONTINUOUS AND ISOLATED LANGUAGE
DE602005002706T2 (en) Method and system for the implementation of text-to-speech
DE60035001T2 (en) Speech synthesis with prosody patterns
DE69834553T2 (en) ADVANCED VOICE RECOGNITION SYSTEM WITH AN AUDIO FEEDBACK
DE69917415T2 (en) Speech synthesis with prosody patterns
DE60216069T2 (en) LANGUAGE-TO-LANGUAGE GENERATION SYSTEM AND METHOD
EP0886853B1 (en) Microsegment-based speech-synthesis process
DE69827988T2 (en) Speech models for speech recognition
DE69829235T2 (en) Registration for speech recognition
EP1466317B1 (en) Operating method for an automated language recognizer intended for the speaker-independent language recognition of words in different languages and automated language recognizer
EP1282112B1 (en) Method of supporting proofreading of a recognized text in a speech to text system with playback speed adapted to confidence of recognition
DE60004420T2 (en) Recognition of areas of overlapping elements for a concatenative speech synthesis system
DE2212472A1 (en) Procedure and arrangement for the speech synthesis of printed message texts
EP1892700A1 (en) Method for speech recognition and speech reproduction
DE69917960T2 (en) Phoneme-based speech synthesis
DE19942178C1 (en) Method of preparing database for automatic speech processing enables very simple generation of database contg. grapheme-phoneme association
EP1273003B1 (en) Method and device for the determination of prosodic markers
DE69738116T2 (en) Localization of a pattern in a signal
DE69727046T2 (en) METHOD, DEVICE AND SYSTEM FOR GENERATING SEGMENT PERIODS IN A TEXT-TO-LANGUAGE SYSTEM
EP1264301A1 (en) Method for recognition of verbal utterances by a non-mother tongue speaker in a speech processing system
DE10040063A1 (en) Procedure for assigning phonemes

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

AK Designated contracting states

Kind code of ref document: A2

Designated state(s): AT BE CH CY DE DK ES FI FR GB GR IE IT LI LU MC NL PT SE TR

AX Request for extension of the european patent

Free format text: AL;LT;LV;MK;RO;SI

PUAL Search report despatched

Free format text: ORIGINAL CODE: 0009013

AK Designated contracting states

Designated state(s): AT BE CH CY DE DK ES FI FR GB GR IE IT LI LU MC NL PT SE TR

AX Request for extension of the european patent

Extension state: AL LT LV MK RO SI

17P Request for examination filed

Effective date: 20030107

17Q First examination report despatched

Effective date: 20030718

AKX Designation fees paid

Designated state(s): DE FR GB IT

GRAP Despatch of communication of intention to grant a patent

Free format text: ORIGINAL CODE: EPIDOSNIGR1

GRAP Despatch of communication of intention to grant a patent

Free format text: ORIGINAL CODE: EPIDOSNIGR1

GRAS Grant fee paid

Free format text: ORIGINAL CODE: EPIDOSNIGR3

GRAA (expected) grant

Free format text: ORIGINAL CODE: 0009210

AK Designated contracting states

Kind code of ref document: B1

Designated state(s): DE FR GB IT

REG Reference to a national code

Ref country code: GB

Ref legal event code: FG4D

Free format text: NOT ENGLISH

GBT Gb: translation of ep patent filed (gb section 77(6)(a)/1977)

Effective date: 20050928

REF Corresponds to:

Ref document number: 50107556

Country of ref document: DE

Date of ref document: 20051103

Kind code of ref document: P

ET Fr: translation filed
PLBE No opposition filed within time limit

Free format text: ORIGINAL CODE: 0009261

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: NO OPPOSITION FILED WITHIN TIME LIMIT

26N No opposition filed

Effective date: 20060629

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: FR

Payment date: 20130724

Year of fee payment: 13

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: IT

Payment date: 20130726

Year of fee payment: 13

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: DE

Payment date: 20140919

Year of fee payment: 14

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: GB

Payment date: 20140709

Year of fee payment: 14

REG Reference to a national code

Ref country code: FR

Ref legal event code: ST

Effective date: 20150331

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: IT

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20140723

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: FR

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20140731

REG Reference to a national code

Ref country code: DE

Ref legal event code: R119

Ref document number: 50107556

Country of ref document: DE

GBPC Gb: european patent ceased through non-payment of renewal fee

Effective date: 20150723

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: DE

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20160202

Ref country code: GB

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20150723