|
(11) | EP 0 712 529 B1 |
| (12) | EUROPEAN PATENT SPECIFICATION |
|
|
| (54) |
SYNTHESISING SPEECH BY CONVERTING PHONEMES TO DIGITAL WAVEFORMS SPRACHSYNTHESE DURCH KONVERSION VON PHONEMEN IN DIGITALE WELLENFORMEN SYNTHESE DE LA PAROLE PAR LA CONVERSION DE PHONEMES EN FORMES D'ONDES NUMERIQUES |
|
|
|||||||||||||||||||||||||||||||
| Note: Within nine months from the publication of the mention of the grant of the European patent, any person may give notice to the European Patent Office of opposition to the European patent granted. Notice of opposition shall be filed in a written reasoned statement. It shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention). |
[0001] This invention relates to synthetic speech and more particularly to a method of synthesising a digital waveform from signals representing phonemes.
[0002] There are many circumstances, eg. in telephone systems, where it is convenient to use synthesised speech. In some applications the starting point is an electronic representation of conventional typography, eg. a disk produced by a word processor. Many stages of processing are needed to produce synthesised speech from such a starting point but, as a preliminary part of the processing, it is usual to convert the conventional text into a phonetic text. In this specification the signals representing such a phonetic text will be called "phonemes". Thus this invention addresses the problem of converting the signals representing phonemes into a digital waveform. It will be appreciated that the digital waveforms are commonplace in audio technology and digital-to-analogue converters and loud speakers are well known devices which enable digital waveforms to be converted into acoustic waveforms.
[0003] Many processes for converting phonemes into digital waveforms have been proposed and it is conventional to do this by means of a linked database comprising a large number of entries, each having an access portion defined in phonemes and an output portion containing the digital waveform corresponding to the access phonemes. Clearly all the phonemes should be represented in the access portions but it is also known to incorporate strings of phonemes in addition. However, existing systems only take into account the phoneme strings contained in the access portions and do not further take into account the context of the strings. The document ICASSP 1988, Vol. 1, 11th April 1988 New York, USA, pages 659-662 by Nakajima et al "Automatic Generation of Synthesis Units Based on Context Oriented Clustering" is acknowledge. This document describes a method of synthesising speech using synthesis units (allophones) which are determined and generated automatically using a labelled natural speech database. In the synthesis phase of the method, optimum synthesis units are selected on the basis of a "context matching score" against a given phoneme string. By concatenating these units, a sequence of spectral parameters is obtained. A waveform is obtained from these parameters and, since each unit represents precisely the appropriate coarticulatory phenomena, no smoothing or interpolation techniques are required.
[0004] This invention, which is defined in the claims, uses a linked database to convert strings of phonemes into digital waveform but it also takes into account the context of the selected phoneme string. This invention also comprises a novel form of database which facilitates the taking into account of the context and the invention also includes the method whereby the preferred database strings are selected from alternatives stored therein.
GENERAL DESCRIPTION
[0006] This general description is intended to identify some of the important integers of a preferred embodiment of the invention. Each of these integers will be described in greater detail after this general description.
[0007] The method of the invention converts input signals representing a text expressed in phonemes into a digital waveform which is ultimately converted into an acoustic wave. Before its conversion, the initial digital waveform may be further processed in accordance with methods which will be familiar to persons skilled in the art.
[0008] The phoneme set used in the preferred embodiment conform to the SAMP-PA (Speech Assessment Methologies - Phonetic Alphabet) simple set number 6. It is to be understood that the method of the invention is carried out in electronic equipment and the phonemes are provided in the form of signals so that the method corresponds to the converting of an input waveform into an output waveform.
[0009] The preferred embodiment of the invention converts waveform representing strings of one, two or three phonemes into digital waveform but it always operates on strings of five phonemes so that at least one preceding and at least one following phoneme is taken into account. This has the effect that, when alternative strings of five phonemes are available, the "best" context is selected.
[0010] It has just been explained that this invention makes particular use of a string of five phonemes and this string will hereinafter be called a "context window" and the five phonemes which constitute the "context window" will be identified as P1, P2, P3, P4 and P5 in sequence. It is a key feature of this invention that a "data context window" being five consecutive phonemes from the input signal is matched with an "access context window" being a sequence of five consecutive phonemes contained in the database.
[0011] The prior art includes techniques in which variable length strings are converted into digital waveform. However, the context of the selected strings is not taken into account. Each phoneme comprised in a selected string is, of course, in context with all the other phonemes of the string but the context of the string as a whole is not taken into account. This invention not only takes into account the contexts within the selected string but it also selects a best matching string from the strings available in the database. This specification will now describe important integers of preferred embodiment namely:-
(i) the definition of "best" as used in the selections;
(ii) the configuration of the database which stores the signal representations of the data context windows together with their corresponding digital wave forms;
(iii) the method of selection for (ii) using (i); and
(iv) picking one of the various alternatives provided by (iii).
DEFINITION OF "BEST"
[0012] This invention selects from alternative context windows on the basis of a "best" match between the input context window and the various stored context windows. Since there are many, e. g. 108 or 1010, possible contexts windows (of 5 phonemes each) it is not possible to store all of them, i.e. the database will lack some of the possible context windows. If all possible context windows were stored it would not be necessary to define a "best" match since an exact correspondence would always be available. However, each individual phoneme should be included in the database and it is always possible to achieve an exact match for at least one phoneme, in the preferred embodiment it is always possible to match exactly P3 of the data context window with P3 of the stored context window but, in general, further exact matches may not be possible.
[0013] This invention defines a correlation parameter between two phonemes as follows. Corresponding to each phoneme there is a type-vector which consists of an ordered list of co-efficients. Each of these co-efficients represents a feature of its phoneme, e.g. whether its phoneme is voiced or unvoiced or whether or not its phoneme is a silibant, a plosive or a labil. It is also desirable to include locational features, eg whether or not the phoneme is in a stressed or unstressed syllable. Thus the type vector uniquely characterises its phoneme and two phonemes can be compared by comparing their type-vectors co-efficient by co-efficient; e.g. by using an exclusive-or gate (which is sometimes called an equivalence gate). The number of matchings is one way of defining the correlation parameter. If desired this can be converted to a percentage by dividing by the maximum possible value of the parameter and multiplying by 100.
[0014] (As an alternative, a mis-match parameter can be defined e.g. by counting the numberof discrepancies in the two type vectors. It will be appreciated that selecting an "best" match is equivalent to selecting a lowest mis-match. )
[0015] The primary definition relates to the correlation parameter of a pair of phonemes. The correlation parameter of a string is obtained by summing or averaging the parameters of the corresponding pairs in the two strings. Weighted averages can be utilised where appropriate.
DATABASE
[0016] In the preferred embodiment, the database is based on an extended passage of the selected language, eg English (although the information content of the passage is not important). A suitable passage lasts about two or three minutes and it contains about 1000-1500 phonemes. The precise nature of the extended passage is not particularly important although it must contain every phoneme and it should contain every phoneme in a variety of contexts.
[0017] The extended passage can be stored in two different formats. First the extended passage can be expressed in phonemes to provide the access section of a linked database. More specifically, the phonemes representing the extended passage are divided into context windows each of which contains 5 phonemes. The method of the invention comprises obtaining best matches for the data context windows with the stored context windows just identified.
[0018] The extended passage can also be provided in the form of a digitised wave form. As would be expected, this is achieved by having a reader or reciter speak the extended passage into a microphone so as to make a digital recording using well established technology. Any point in the digital recording can be defined by a parameter, e.g. by the time from the start. Analysing the recording establishes values for the time-parameter corresponding to the break between each pair of phonemes in the equivalent text. This arrangement permits phoneme-to-waveform conversion for any included string by establishing the starting value of the time-parameter corresponding to the first phoneme of the string and the finishing value for the time-parameter corresponding to the last phoneme of the string and retrieving the equivalent portion of database, ie the specified digital waveform. Specifically a conversion for any string of one, two or three phonemes can be achieved.
[0019] The important requirement is to select the best portion of the extended text for the conversion.
[0020] It has already been mentioned that the phoneme version of the extended text is stored in the form of context windows each of five phonemes. This is most suitably achieved by storing the phonemes in a tree which has three hierarchical levels.
[0021] The first level of the hierarchy is defined by phoneme P3 of each window. The effect is that every phoneme gives direct access to a subset of the context windows ie. the totality of context windows is divided into subsets and each subset has the same value of P3.
[0022] The next level of the tree is defined by phonemes P2 and P4 and, since this selection is made from the subsets defined above, the effect is that the totality of context windows is further divided into smaller subsets each of which is defined by having phonemes P2, P3 and P4 in common. (There are approximately half a million subsets but most of them will be empty because the relevant sequence P2, P3, P4 does not occur in the extended text). Empty subsets are not recorded at all so that the database remains of manageable size. Nevertheless it is true that for each triple sequence P2, P3, P4 which occurs in the extended text there will be a subset recorded in the second level of the database under P2, P4 which level will also have been indexed at the first level under P3.
[0023] Finally the second level gives access to a third level which contains subsets having P2, P3 and P4 as exact matches and it contains all the values of P1 and P5 corresponding to these triples. Best matches for data P1 and P5 are selected. This selection completely identifies one of the context windows contained in the extended text and it provides access to time-parameters of said window. Specifically it provides start and finish time-parameters for up to four different strings as follows:-
(a) P3 by itself;
(b) the pair of phonemes P2 + P3;
(c) the pair of phonemes P3 + P4; and
(d) the triple consisting of the phonemes P2 + P3 + P4.
[0024] In the first instance, the database provides beginning and ending values of the time-parameter corresponding to each one of the selected strings (a) - (d). As explained above, the time-parameter defines the relevant portion of a digital wave form so that the equivalent wave form is selected.
[0025] It should be noted that item (d) will be offered if it is contained in the database; in this case items (a), (b), and (c) are all embedded in the selected (d) and they are, therefore, available as alternatives. If item (d) is not contained in the database then, clearly, this option cannot be offered.
[0026] Even if item (d) is missing from the database, then items (b) and/or (c) may still be present in the database. When both of these options are offered they will usually arise from different parts of the database because item (d) is missing. Therefore, depending on the content of the database, the selection will offer (b) alone, or (c) alone, or both (b) and (c). Thus the selection may provide a choice and in any case item (a) is available because it is embedded in the pair.
[0027] Finally, even if (b), (c) and (d) are all absent from the database, item (a) will always be present and thus "best match" will be offered for the single phoneme and this will be the only possibility which is offered.
[0028] It will be apparent that items (b), (c) and (d) imply that strings will overlap. Thus whenever item (c) is selected for any phoneme then item (b) must be available for the next phoneme. If nothing better offered, then the same part of the database will meet the requirements of (c) for the earlier phoneme and (b) for the later but because different correlations are involved better choices may be selected. It will also be apparent that whenever item (d) is available item (c) will be available for the previous phoneme and, in addition, item (b) will be available for the following phoneme. In other words, some of the strings will overlap, ie there will be alternatives for some phonemes such that the Same phoneme occurs in different places in different strings. This aspect of the invention is described in greater detail below.
[0029] It has been emphasised that the preferred embodiment is based on a context window which is five phonemes long. However the full string of five phonemes is never selected. Even if, fortuitously, the input text contains a string of five found in the database only the triple string P2, P3, P4 will be used. This emphasises that the important feature of the invention is the selection of a string from a context and, therefore, the invention selects the "best" context window of five phonemes and only uses a portion thereof in order to ensure that all selected strings are based upon a context.
SELECTION OF "BEST" WINDOW
[0030] The analysis of the text into phonemes contained in the database is carried out phoneme by phoneme, but each phoneme is utilised in its context window. The next part of the description will be based upon the selection procedure for one of the data phonemes it being understood that the same procedure is used for each of the data phonemes.
[0031] The selected data phoneme is not utilised in isolation but as part of its context window. More precisely the selected data phoneme becomes phoneme P3 of a data window with its two predecessors and two successors being selected to provide the five phonemes of the relevant context window. The database described above is searched for this context window; since it is unlikely that the exact window will be located, the search is for the best fitting of the stored context windows.
[0032] The first step of the search involves accessing the tree described above using phoneme P3 as the indexing element. As explained above this gives immediate access to a subset of the stored context windows. More specifically, accessing level one by phoneme P3 gives access to a list of phoneme pairs which correspond to possible values of P2 and P4 of the data context-window. The best pair is selected according to the following four criteria.
[0033] First criterion Fortuitously , it may happen that one pair in the sub-set gives an exact match for data P2 and P4. When this happens that pair is selected and the search immediately proceeds to level 3. This outcome is unlikely because, as explained in greater detail above, the string P2, P3, P4 may not be contained in the extended passage.
[0034] Second criterion. In the absence of a triple match a left pair will be selected if it occurs. The left-hand match is selected when an exact match for P2 is found and, if alternatives offer, the P4 which has the highest correlation parameter will be selected to give access to level 3 of the tree.
[0035] The third criterion is similar to the second except that it is a right-hand pair depending upon an exact match being discovered for P4. In this case access to level 3 is given by the P2 value which provides the highest correlation parameter.
[0036] Criterion four occurs when there is no match for either P2 or P3 in which the case the pair P2, P4 with the highest average correlation parameter is selected as the basis of access to level 3.
[0037] It will be noted that if criterion 1 succeeds, then it will be possible to take as alternatives a left-hand pair, a right-hand pair and a single value in accordance with criterion 2, 3 and 4.
[0038] Even if criterion 1 fails, it is still possible that a left-hand pair will be found by criterion 2 and it is even possible that, simultaneously, a right-hand pair will be found by criterion 3. However because criterion 1 has failed they will be selected from different parts of the database and they will give access to different parts of the tree at level 3.
[0039] Finally criterion 4 will only be accepted when criterion 1, 2 and 3 have all failed and it follows that the phoneme P3 cannot be found in triples or pairings when used in other context windows.
[0040] Thus, when criterion 1 or 4 are utilised there will only be access to one portion of the tree at the third level but it is possible, when criterion 2 and 3 are used that there will be access to two different parts of the third level.
[0041] We have now described how the selection of a context window gives rise to either one or two areas of the third level of the tree. In each case the third level may contain several pairings for phonemes 1 and 5 of the data context window. The pair with the best average correlation parameter is selected as the context window in the access portion of the database. As explained above this context window is converted to digital wave form using the time-parameter.
[0042] To re-emphasise; where criterion 1 is used only one context window is selected but it gives rise to four possibilities, namely time-parameter ranges for: -
(i) the triple P2 + P3 + P4;
(ii) the left-hand pair P2 + P3;
(iii) the right-hand pair P3 + P4, and;
(iv) the single P3 by itself.
[0043] When criterion 2 operates, this provides time-parameter ranges only for the left-hand pair P2 + P3 and for a single P3 by itself. When criterion 3 operates similar considerations apply but the parameter ranges are for the right-hand pair P2 + P3 and for the single P4. If both criterion operate this offers two choices for the single P3 and only the one with the higher correlation parameter for P1 + P5 is selected.
[0044] Finally when criterion 4 operates there only one possibility namely the phoneme P3 by itself.
[0045] The description given above explains how conversions are provided for each phoneme of an input text. Sometimes the method provides a conversion for only a single phoneme and, in this case, no alternatives are offered. In some cases the method provides conversion for strings of two or three adjacent phonemes and, in these circumstances, the conversion provides alternatives for at least one phoneme. In order to complete the selection, it is necessary to reduce the number of alternatives to one. The preferred method of achieving this reduction will now be explained.
[0046] The preferred method of making the reduction is carried out by processing a short segment of input text, eg. a segment which begins and ends with a silence. Provided it is not too long a sentence constitutes a suitable segment. If a sentence is very long, e.g. more than thirty words, it usually contains one or more embedded silences, eg between clauses or other sub-units. In the case of long sentences such sub-units are suitable for use as the segments.
[0047] The processing of a segment to reduce each set of alternatives to one will now be described. As mentioned, no alternative will be offered for some of the phonemes and, therefore, no selection is required for these phonemes. Alternatives will be available for the other phonemes and the selection is made so as to produce a "best" result for the segment as a whole. This may involve making a locally "less good" selection at one point in the segment in order to obtain "better" selection elsewhere in the segment. The criteria of "better" include:-
(i) taking longer strings rather than shorter strings, and
(ii) selecting from strings which overlap rather than from strings which merely abut.
[0048] The rejection of unwanted alternatives produces a position in which each phoneme has one, and only one, conversion. In other words the input text will have been divided into sub-strings of 1, 2 or 3 phonemes matching the database and the beginning and ending values for the selected streams will therefore be established. The output portion of the database takes the form of a digitised waveform and the parameters which have been established define segments of this waveform. Therefore the designated segments are selected and abutted to produce the digital waveform corresponding to the input text. This completes the requirement of the invention.
[0049] Having obtained a digital waveform this can be provided as audible output using conventional digital to analogue conversion techniques and conventional loudspeakers. If desired, the primary digital waveform can be enhanced using techniques known to those skilled in the art.
[0050] The invention will be further described by way of example with reference to the accompanying drawings in which:-
Figure 1 illustrates diagrammatically a speech engine in accordance with the invention; and
Figure 2 shows a speech engine as illustrated in Figure 1 attached to a telephone network.
[0051] As shown in Figure 1 the speech engine according to the invention comprises primary processor 11 which is adapted to accept text in grapnemes and to produce therefrom an equivalent text in phonemes. This text is passed to converter 12 which is operatively associated with a database 13 in accordance with the invention. Converter 12 matches segments of the phoneme text with segments stored in the access portion of database 13. Thus segments of digital waveform are retrieved and these are assembled into extended portions of digital waveform corresponding to extended portions of the original input.
[0052] These extended portions of digital waveform are passed to waveform processor 14 where they are subjected to further processing in order to produce a smooth output. Finally the aigital output is converted into an analogue waveform which is provided at output port 15 for onward transmission.
[0053] As shown in Figure 1 the speech engine is connected to receive its input from an external database 16 which holds texts in conventional orthography. External database 16 is conveniently operated by keyboard 17 to select a text stored in database 16. This text is provided into the primarv processor 11 and it appears at the output port 15 as an analogue waveform.
[0054] Figure 2 shows a speech engine as illustrated in Figure 1 attached to a public access telephone network. As shown in Figure 2, a conventional speech telephone 20 is connected to a station 22 via a switched access network 21. Station 22 includes a speech engine as shown in Figure 1 and the output port 15 is connected to the network so that the information available in the external database 16 can be provided, as an analogue acoustic waveform, to the telephone 20.
[0055] If desired the keypad (used for dialling) of the telephone 20 can be used as the keyboard 17 of the external database 16 (in which case the external database 16 preferably contains instructions which can be read by the speech engine). A simpler technical arrangement provides a human operator at the station 20 and the human operator actuates the keyboard 17 in accordance with instructions received over the network 21. When the operator has selected a portion of text this is read by the speech engine and further participation by the operator is unnecessary. Thus the operator is freed to assist with further enquiries and the use of a speech engine enhances the efficiency of the operation.
1. A method of converting an input signal into an output signal, wherein said input signal
represents a text in phonemes and said output signal is a digital waveform convertible
to an acoustic waveform corresponding to said text, wherein said method makes use
of a two-part database having an access section linked to an output section, characterised
in that said access section defines access windows each of which corresponds to a
string of phonemes and said output section contains digital waveforms corresponding
to the access windows; wherein said method comprises comparing windows of said input
signal with the access windows to select, in each case, the access window which provides
the best match including an exact match for at least one internal phoneme and discarding
at least the first and last phonemes of said best match to identify a shorter string
of phonemes which is an exact match for a portion of said input signal, retrieving
from the output section the digital waveform corresponding to the selected exact match
and thereafter joining together the selected portions of the digital waveform to produce
the output signal.
2. A method according to claim 1, in which the access section is based on an extended
text in phonemes and each access window corresponds to a string of phonemes contained
in said extended text; the output section contains an extended digital waveform corresponding
to the extended phoneme text of the access section; and the portion retrieved from
the output section is the segment of the extended digital waveform which corresponds
to the exact match.
3. A method according to either claim 1 or claim 2, which method includes forming a best
match for a window of five phonemes of said input signal discarding at least the first
and last phonemes of said best match to identify an exact match for a string of one,
two or three phonemes.
4. A method according to claim 3, in which the input section of the database is organised
into three hierarchical levels; namely
(i) a top level containing single phonemes corresponding to the central phoneme of a window;
(ii) a second level which contains the equivalents of the second and fourth phonemes if a window; and
(iii) a lowest level which contains the equivalents of the first and fifth phonemes of a window;
and the matching comprises selecting an exact match for the central phoneme of the input window from the first level of the hierarchy, selecting a best match for phonemes 2 and 4 from the second level of the hierarchy corresponding to the selected portion of the top level of the hierarchy and, finally, selecting from the bottom level of the hierarchy the best match for phonemes 1 and 5 from the portion of the bottom level which corresponds to the selection in the second level of the hierarchy.5. A method according anyone of the preceding claims, in which the digital output is
converted into an analogue signal.
6. A database component for use in speech engine, said database having an access section
containing signals representing phonemes linked to an output section containing digital
waveform, characterised in that said access section is based on an extended text divided
into access windows each of which contains five phonemes and that said output section
contains an extended digital waveform corresponding to the extended phoneme text of
the access section, wherein the access section is organised into three hierarchical
levels; namely
(i) a top level containing single phonemes corresponding to the central phonemes of an access window;
(ii) a second level which contains the equivalents of the second and fourth phonemes of an access window identified in the top level; and
(iii) a lowest level which contain the equivalents of the first and fifth phonemes of an access window identified in the second level.
wherein the linkage between the access and output sections is such that the identification of an access window from levels (i), (ii) and (iii) also identifies the corresponding window of digital waveform.7. A speech engine which comprises a primary processor (11) for converting a text in
graphemes into an equivalent text in phonemes and a converter (12) for converting
said text in phonemes into a digital waveform, characterised in that the converter
(12) includes a database (13) according to claim 6.
8. A telephone network which includes a speech engine according to claim 7, said speech
engine being connected to the network for the transmission of the output of the speech
engine to a remote location.
1. Verfahren zum Umsetzen eines Eingangssignals in ein Ausgangssignal, wobei das Eingangssignal
einen Text in Phonemen repräsentiert und das Ausgangssignal eine digitale Signalform
ist, die in eine akustische Signalform, die dem Text entspricht, umsetzbar ist, wobei
das Verfahren von einer zweiteiligen Datenbank Gebrauch macht, die einen mit einem
Ausgangsabschnitt verbundenen Zugriffsabschnitt besitzt, dadurch gekennzeichnet, daß
der Zugriffsabschnitt Zugriffsfenster definiert, wovon jedes einer Kette von Phonemen
entspricht, und der Ausgangsabschnitt digitale Signalformen enthält, die den Zugriffsfenstern
entsprechen; wobei das Verfahren enthält: Vergleichen von Fenstern des Eingangssignals
mit den Zugriffsfenstern, um in jedem Fall dasjenige Zugriffsfenster zu wählen, das
die beste Anpassung einschließlich einer exakten Anpassung für wenigstens ein internes
Phonem ergibt, und Verwerfen wenigstens des ersten und des letzten Phonems der besten
Anpassung, um eine kürzere Kette von Phonemen zu identifizieren, die eine exakte Anpassung
für einen Abschnitt des Eingangssignals ist, Wiederherstellen der digitalen Signalform,
die der gewählten exakten Anpassung entspricht, aus dem Ausgangsabschnitt und danach
Verbinden der gewählten Abschnitte der digitalen Signalform, um das Ausgangssignal
zu erzeugen.
2. Verfahren nach Anspruch 1, bei dem der Zugriffsabschnitt auf einem erweiterten Text
in Phonemen basiert und jedes Zugriffsfenster einer Kette von in dem erweiterten Text
enthaltenen Phonemen entspricht; der Ausgangsabschnitt eine erweiterte digitale Signalform
enthält, die dem erweiterten Phonem-Text des Zugriffsabschnitts entspricht; und der
vom Ausgangsabschnitt wiedergewonnene Abschnitt dasjenige Segment der erweiterten
digitalen Signalform ist, das der exakten Anpassung entspricht.
3. Verfahren entweder nach Anspruch 1 oder nach Anspruch 2, das enthält: Bilden einer
besten Anpassung für ein Fenster aus fünf Phonemen des Eingangssignals und Verwerfen
wenigstens des ersten und des letzten Phonems der besten Anpassung, um eine exakte
Anpassung für eine Kette aus einem, zwei oder drei Phonemen zu identifizieren.
4. Verfahren nach Anspruch 3, in dem der Eingangsabschnitt der Datenbank in drei hierarchischen
Ebenen organisiert ist; nämlich
(i) einer obersten Ebene, die einzelne Phoneme enthält, die dem zentralen Phonem eines Fensters entsprechen;
(ii) einer zweiten Ebene, die die Äquivalente des zweiten und des vierten Phonems eines Fensters enthält; und
(iii) einer untersten Ebene, die die Äquivalente des ersten und des fünften Phonems eines Fensters enthält;
und die Anpassung enthält: Wählen einer exakten Anpassung für das zentrale Phonem des Eingangsfensters aus der ersten Ebene der Hierarchie, Wählen einer besten Anpassung für die Phoneme 2 und 4 aus der zweiten Ebene der Hierarchie, die dem gewählten Abschnitt der oberen Ebene der Hierarchie entsprechen, und schließlich Wählen der besten Anpassung für die Phoneme 1 und 5 aus der unteren Ebene der Hierarchie aus demjenigen Abschnitt der unteren Ebene, der der Wahl der zweiten Ebene der Hierarchie entspricht.5. Verfahren nach irgendeinem der vorangehenden Ansprüche, bei dem das digitale Ausgangssignal
in ein analoges Signal umgesetzt wird.
6. Datenbankkomponente zur Verwendung in einer Sprachmaschine, wobei die Datenbank einen
Zugriffsabschnitt besitzt, der Phoneme repräsentierende Signale enthält und mit einem
Ausgangsabschnitt verbunden ist, der eine digitale Signalform enthält, dadurch gekennzeichnet,
daß der Zugriffsabschnitt auf einem erweiterten Text basiert, der in Zugriffsfenster
unterteilt ist, wovon jedes fünf Phoneme enthält, und der Ausgangsabschnitt eine erweiterte
digitale Signalform enthält, die dem erweiterten Phonem-Text des Zugriffsabschnitts
entspricht, wobei der Zugriffsabschnitt in drei hierarchischen Ebenen organisiert
ist; nämlich
(i) einer obersten Ebene, die einzelne Phänomene enthält, die den zentralen Phonemen eines Zugriffsfensters entsprechen;
(ii) einer zweiten Ebene, die Äquivalente des zweiten und des vierten Phonems eines Zugriffsfensters, das in der obersten Ebene identifiziert wird, enthält; und
(iii) eine unterste Ebene, die die Äquivalente des ersten und des fünften Phonems eines Zugriffsfensters, das in der zweiten Ebene identifiziert wird, enthält,
wobei die Verbindung zwischen dem Zugriffsabschnitt und dem Ausgangsabschnitt derart ist, daß die Identifizierung eines Zugriffsfensters aus den Ebenen (i), (ii) und (iii) auch das entsprechende Fenster der digitalen Signalform identifiziert.7. Sprachmaschine, die einen primären Prozessor (11) zum Umsetzen eines Texte in Graphemen
in einen äquivalenten Text in Phonemen sowie einen Umsetzer (12) zum Umsetzen des
Texts in Phonemen in eine digitale Signalform enthält, dadurch gekennzeichnet, daß
der Umsetzer (12) eine Datenbank (13) nach Anspruch 6 enthält.
8. Telephonnetz, das eine Sprachmaschine nach Anspruch 7 enthält, wobei die Sprachmaschine
mit dem Netz verbunden ist, um den Ausgang der Sprachmaschine an einen entfernten
Ort zu senden.
1. Procédé de conversion d'un signal d'entrée en un signal de sortie, dans lequel ledit
signal d'entrée représente un texte sous forme de phonèmes et ledit signal de sortie
est une forme d'onde numérique qui peut être convertie en une forme d'onde acoustique
correspondant audit texte, dans lequel ledit procédé fait usage d'une base de données
en deux parties comportant une section d'accès reliée à une section de sortie, caractérisé
en ce que ladite section d'accès définit des fenêtres d'accès, dont chacune correspond
à une chaîne de phonèmes, et ladite section de sortie contient des formes d'onde numériques
correspondant aux fenêtres d'accès, dans lequel ledit procédé comprend la comparaison
des fenêtres dudit signal d'entrée aux fenêtres d'accès afin de sélectionner, dans
chaque cas, la fenêtre d'accès qui fournit la meilleure correspondance comprenant
une correspondance exacte pour au moins un phonème interne, et le rejet d'au moins
le premier et le dernier phonème de ladite meilleure correspondance afin d'identifier
une chaîne plus courte de phonèmes qui constitue une correspondance exacte pour une
partie dudit signal d'entrée, la récupération à partir de la section de sortie de
la forme d'onde numérique correspondant à la correspondance exacte sélectionnée et
ensuite la jonction des parties sélectionnées de la forme d'onde numérique afin de
produire le signal de sortie.
2. Procédé selon la revendication 1, dans lequel la section d'accès est fondée sur un
texte long en phonèmes, et chaque fenêtre d'accès correspond à une chaîne de phonèmes
contenue dans ledit texte long, la section de sortie contient une forme d'onde numérique
longue correspondant au texte long en phonèmes de la section d'accès, et la partie
récupérée à partir de la section de sortie est le segment de la forme d'onde numérique
longue qui correspond à la correspondance exacte.
3. Procédé selon l'une quelconque de la revendication 1 ou 2, lequel procédé comprend
l'élaboration d'une meilleure mise en correspondance pour une fenêtre de cinq phonèmes
dudit signal d'entrée en rejetant au moins les premier et dernier phonèmes de ladite
meilleure correspondance afin d'identifier une correspondance exacte pour une chaîne
de un, deux ou trois phonèmes.
4. Procédé selon la revendication 3, dans lequel la section d'entrée de la base de données
est organisée en trois niveaux hiérarchiques, à savoir
(i) un niveau supérieur contenant des phonèmes isolés correspondant au phonème central d'une fenêtre,
(ii) un second niveau qui contient les équivalents des second et quatrième phonèmes d'une fenêtre, et
(iii) un niveau le plus bas qui contient les équivalents des premier et cinquième phonèmes d'une fenêtre,
et la mise en correspondance comprend la sélection d'une correspondance exacte pour le phonème central de la fenêtre d'entrée à partir du premier niveau de la hiérarchie, la sélection d'une meilleure correspondance pour les phonèmes 2 et 4 à partir du second niveau de la hiérarchie correspondant à la partie sélectionnée du niveau supérieur de la hiérarchie, et enfin la sélection, à partir du niveau inférieur de la hiérarchie, de la meilleure correspondance pour les phonèmes 1 et 5 à partir de la partie du niveau inférieur qui correspond à la sélection au second niveau de la hiérarchie.5. Procédé selon l'une quelconque des revendications précédentes, dans lequel la sortie
numérique est convertie en un signal analogique.
6. Composant de base de données destiné à être utilisé dans un moteur vocal, ladite base
de données comportant une section d'accès contenant des signaux représentant des phonèmes
reliée à une section de sortie contenant une forme d'onde numérique, caractérisé en
ce que ladite section d'accès est fondée sur un texte long divisé en fenêtres d'accès,
dont chacune contient cinq phonèmes et en ce que ladite section de sortie contient
une forme d'onde numérique longue correspondant au texte long en phonèmes de la section
d'accès, dans lequel la section d'accès est organisée en trois niveaux hiérarchiques,
à savoir
(i) un niveau supérieur contenant des phonèmes isolés correspondant aux phonèmes centraux d'une fenêtre d'accès,
(ii) un second niveau qui contient les équivalents des second et quatrième phonèmes d'une fenêtre d'accès identifiée dans le niveau supérieur, et
(iii) un niveau le plus bas qui contient les équivalents des premier et cinquième phonèmes d'une fenêtre d'accès identifiée dans le second niveau,
dans lequel la liaison entre les sections d'accès et de sortie est telle que l'identification d'une fenêtre d'accès depuis les niveaux (i), (ii) et (iii) identifie également la fenêtre correspondante de la forme d'onde numérique.7. Moteur vocal qui comprend un processeur principal (11) destiné à convertir un texte
en graphèmes en un texte équivalent en phonèmes et un convertisseur (12) destiné à
convertir ledit texte en phonèmes en une forme d'onde numérique, caractérisé en ce
que le convertisseur (12) comprend une base de données (13) selon la revendication
6.
8. Réseau téléphonique qui comprend un moteur vocal selon la revendication 7, ledit moteur
vocal étant relié au réseau pour la transmission de la sortie du moteur vocal à un
emplacement à distance.