WO2000055842A2

WO2000055842A2 - Speech synthesis

Info

Publication number: WO2000055842A2
Application number: PCT/GB2000/000854
Authority: WO
Inventors: Stephen Minnis
Original assignee: British Telecommunications Public Limited Company
Priority date: 1999-03-15
Filing date: 2000-03-08
Publication date: 2000-09-21
Also published as: EP1163663A2; AU2931600A; CA2366952A1; US6996529B1; WO2000055842A3

Abstract

Conventional methods of predicting phrase boundaries occasionally result in the output of text-to-speech conversion apparatus sounding unnatural. Text-to-speech conversion apparatus described herein uses pattern-matching to predict the position of phrase boundaries in its spoken output. The apparatus analyses text input to the apparatus to identify groups of words (known as 'chunks') which are unlikely to contain internal phrase boundaries. Both the chunks and individual words are labelled with their syntactic characteristics. The apparatus has access to a database of sentences which also contains such syntactic labels, together with indications of where a human reader would insert minor and major phrase boundaries. The parts of the database which have the most similar syntactic characteristics are found and phrase boundaries are predicted based on the phrase boundaries found in those parts. Other characteristics are also used in the pattern-matching process.

Description

SPEECH SYNTHESIS

The present invention relates to a method and apparatus for converting text to speech.

Although text-to-speech conversion apparatus has improved markedly over recent years, the sound of such apparatus reading a piece of text is still distinguishable from the sound of a human reading the same text. One reason for this is that text-to- speech converters occasionally apply phrasing that differs from that which would be applied by a human reader. This makes speech synthesised from text more onerous to listen to than speech read by a human.

The development of methods for predicting the phrasing for an input sentence has, thus far, largely mirrored developments in language processing. Initially, automatic language processing was not available, so early text-to-speech converters relied on punctuation for predicting phrasing. It was found that punctuation only represented the most significant boundaries between phrases, and often did not indicate how the boundary was to be conveyed acoustically. Hence, although this method was simple and reasonably effective, there was still room for improvement. Thereafter, as automatic language processing developed, lexicons which indicated the part-of- speech associated with each word in the input text were used. Associating part-of- speech tags with words in the text increased the complexity of the apparatus without offering a concomitant improvement in the prediction of phrasing. More recently, the possibility of using rules to predict phrase boundaries from the length and syntactic structure of the sentence has been discussed (Bachenko J and Fitzpatrick E: 'A computational grammar of discourse-neutral prosodic phrasing in English', Computational Linguistics, vol. 1 6, No. 3, pp1 55-170 ( 1 990)). Others have proposed deriving statistical parameters from a database of sentences which have natural prosodic phrase boundaries marked (Wang, M. and Hirschberg J: 'Predicting intonational boundaries automatically from text: the ATIS domain', Proc. of the DARPA Speech and Natural Language Workshop, pp 378-383 (February 1 991 )) . These recent approaches to the prediction of phrasing still do not provide entirely satisfactory results.

According to a first aspect of the present invention, there is provided a method of converting text to speech comprising the steps of: receiving an input word sequence in the form of text; comparing said input word sequence with each one of a plurality of reference word sequences provided with phrasing information; identifying one or more reference word sequences which most closely match said input word sequence; and predicting phrasing for a synthesised spoken version of the input text on the basis of the phrasing information included with said one or more most closely matching reference word sequences.

By predicting phrasing on the basis of one or more closely matching reference word sequences, sentences are given a more natural-sounding phrasing than has hitherto been the case.

Preferably, the method involves the matching of syntactic characteristics of words or groups of words. It could instead involve the matching of the words themselves, but that would require a large amount of storage and processing power. Alternatively, the method could compare the role of the words in the sentence - i.e. it could identify words or groups of words as the subject, verb or object of a sentence etc. and then look for one or more reference sentences with a similar pattern of subject, verb, object etc.

Preferably, the method further comprises the step of identifying clusters of words in the input text which are unlikely to include prosodic phrase boundaries. In this case, the reference sentences are further provided with information identifying such clusters of words within them. The comparison step then comprises a plurality of per-cluster comparisons. By limiting the possible locations of phrase boundary sites to locations between clusters of words, the amount of processing required is lower than would be required were every inter-word location to be considered. Nevertheless, other embodiments are possible in which a per-word comparison is used.

Measures of similarity between the input clusters and reference clusters which might be used include:

a) measures of similarity in the syntactic characteristics of the input cluster and the reference cluster;

b) measures of similarity in the syntactic characteristics of the words in the input cluster and the words in the reference cluster; and

c) measures of similarity in the number of words or syllables in the input cluster and the reference cluster.

d) measures of similarity in the role (e.g. subject, verb, object) of the input cluster and the reference cluster;

e) measures of similarity in the role of the words in the input cluster and the reference cluster;

f) measures of similarity in word grouping information, such as the start and end of sentences and paragraphs; and

g) measures of similarity in whether new or previously information is being presented in the cluster.

One or a weighted combination of the above measures might be used. Other possible inter-cluster similarity measures will occur to those skilled in the art. In some embodiments, the comparison comprises measuring the similarity in the positions of prosodic boundaries previously predicted for the input sentence and the positions of the prosodic boundaries in the reference sequences. In a preferred embodiment a weighted combination of all the above measures is used.

According to a second aspect of the present invention, there is provided a text to speech conversion apparatus comprising: a word sequence store storing a plurality of reference word sequences which are provided with prosodic boundary information; a program store storing a program; a processor in communication with said program store and the word sequence store; means for receiving an input word sequence in the form of text; wherein said program controls said processor to: compare said input word sequence with each one of a plurality of said reference word sequences; identify one or more reference word sequences which most closely match said input word sequence; and derive prosodic boundary information for the input text on the basis of the prosodic boundary information included with said one or more most closely matching reference word sequences.

According to a third aspect of the present invention, there is provided a program storage device readable by a computer, said device embodying computer readable code executable by the computer to perform a method according to the first aspect of the present invention.

According to a fourth aspect of the present invention, there is provided a signal embodying computer executable code for loading into a computer for the performance of the method according to the first aspect of the present invention. There now follows, by way of example only, a description of specific embodiments of the present invention. The description is given with reference to the accompanying drawings in which:

Figure 1 shows the hardware used in providing a first embodiment of the present invention;

Figures 2A and 2B show the top-level design of a text-to-speech conversion program which controls the operation of the hardware shown in Figure 1 ;

Figures 3A & 3B show the text analysis process of Figure 2A in more detail;

Figure 4 is a diagram showing part of a syntactic classification of words; and

Figure 5 is a flow chart illustrating the prosodic structure assignment process of Figure 2B.

Figure 1 shows a hardware configuration of a personal computer operable to provide a first embodiment of the present invention. The computer has a central processing unit 10 which is connected by data lines to a Random Access Memory (RAM) 1 2, a hard disc 14, a CD-ROM drive 1 6, input/output peripherals 1 8,20,22 and two interface cards 24,28. The input/output peripherals include a visual display unit 1 8, a keyboard 20 and a mouse 22. The interface cards comprise a sound card 24 which connects the computer to a loudspeaker 26 and a network card 28 which connects the computer to the Internet 30.

The computer is controlled by conventional operating system software which is transferred from the hard disc 14 to the RAM 1 2 when the computer is switched on. A CD-ROM 32 carries: a) software which the computer can execute to provide the user with a text-to- speech facility; and b) five databases used in the text-to-speech conversion process.

To use the software, the user loads the CD-ROM 32 into the CD-ROM drive 1 6 and then, using the keyboard 20 and the mouse 22, causes the computer to copy the software and databases from the CD-ROM 32 to the hard disc 14. The user can then select a text-representing file (such as an e-mail loaded into the computer from the Internet 30) and run the text-to-speech program to cause the computer to produce a spoken version of the e-mail via the loudspeaker 26. On running the text-to-speech program both the program itself and the databases are loaded into the RAM 12.

The text-to-speech program then controls the computer to carry out the functions illustrated in Figures 2A and 2B. As will be described in more detail below, the computer first carries out text analysis process 42 on the e-mail (shown as text 40) which the user has indicated he wishes to be converted to speech. The text analysis process 42 uses a lexicon 44 (the first of the five databases stored on the CD-ROM 32) to generate word grouping data 46, syntactic information 48 and phonetic transcription data 49 concerning the text-file 40. The output data 46,48,49 is stored in the RAM 1 2.

After completion of the text analysis program 42, the program controls the computer to carry out the prosodic structure prediction process 50. The process 50 operates on the syntactic data 48 and word grouping data 46 stored in RAM 1 2 to produce phrase boundary data 54. The phrase boundary data 54 is also stored in RAM 1 2. The prosodic structure prediction process 50 uses the prosodic structure corpus 52 (which is the second of the five databases stored on the CD-ROM 32). The process will be described in more detail (with reference to Figures 4 and 5) below.

Once the phrase boundary data 54 has been generated, the program controls the computer to carry out prosody prediction process (Figure 2B, 56) to generate performance data 58 which includes data on the pitch, amplitude and duration of phonemes to be used in generating the output speech 72. A description of the prosody prediction process 56 is given in Edgington M et al: Overview of current text-to-speech techniques part 2 - prosody and speech synthesis', BT Technology Journal, Volume 14, No. 1 , pp 84-99 (January 1 996). The disclosure of that paper (hereinafter referred to as part 2 of the BTTJ article) is hereby incorporated herein by reference.

Thereafter, the computer performs a speech sound generation process 62 to convert the phonetic transcription data 49 to a raw speech waveform 66. The process 62 involves the concatenation of segments of speech waveforms stored in a speech waveform database 64 (the speech waveform database is the third of the five databases stored on the CD-ROM 32). Suitable methods for carrying out the speech sound generation process 62 are disclosed in the applicant's European patent no. 0 71 2 529 and European patent application no. 95302474.9. Further details of such methods can be found in part 2 of the BTTJ article.

Thereafter, the computer carries out a prosody and speech combination process 70 to manipulate the raw speech waveform data 66 in accordance with the performance data 58 to produce speech data 72. Again, those skilled in the art will be able to write suitable software to carry out combination process 70. Part 2 of the BTTJ article describes the process 70 in more detail. The program then controls the computer to forward the speech data 72 to the sound card 24 where it is converted to an analogue electrical signal which is used to drive loudspeaker 26 to produce a spoken version of the text file 40.

The text analysis process 42 is illustrated in more detail in Figures 3A and 3B. The program first controls the computer to execute a segmentation and normalisation process (Figure 3A, 80). The normalisation aspect of the process 80 involves the expansion of numerals, abbreviations, and amounts of money into the form of words, thereby generating an expanded text file 88. For example, '£100' in the text file 40 is expanded to 'one hundred pounds' in the expanded text file 88. These operations are done with the aid of an abbreviations database 82, which is the fourth of the five databases stored on the CD-ROM 32. The segmentation aspect of the process 80 involves the addition of start-of-sentence, end-of-sentence, start-of-paragraph and end-of-paragraph markers to the text, thereby producing the word grouping data (Figure 2A:46) which comprises sentence markers 86 and paragraph markers 87. The segmentation and normalisation process 80 is conventional, a fuller description of it can be found in Edgington M et al: Overview of current text-to-speech techniques part 1 - 'text and linguistic analysis' , BT Technology Journal, Volume 14, No. 1 , pp 68-83 (January 1 996) . The disclosure of that paper (hereinafter referred to as part 1 of the BTTJ article) is hereby incorporated herein by reference.

The computer is then controlled by the program to run a pronunciation and tagging process 90 which converts the expanded text file 88 to an unresolved phonetic transcription file 92 and adds tags 93 to words indicating their syntactic characteristics (or a plurality of possible syntactic characteristics). The process 90 makes use of the lexicon 44 which outputs possible word tags 93 and corresponding phonetic transcriptions of input words. The phonetic transcription 92 is unresolved to the extent that some words (e.g. 'live') are pronounced differently when playing different roles in a sentence. Again, the pronunciation process is conventional - more details are to be found in part 1 of the BTTJ article.

The program then causes the computer to run a conventional parsing process 94. A more detailed description of the parsing process can be found in part 1 of the BTTJ article.

The parsing process 94 begins with a stochastic tagging procedure which resolves the syntactic characteristic associated with each one of the words for which the pronunciation and tagging process 90 has given a plurality of possible syntactic characteristics. The unresolved word tags data 93 is thereby turned into word tags data 95. Once that has been done, the correct pronunciation of the word is identified to form phonetic transcription data 97. In a conventional manner, the parsing process 94 then assigns syntactic labels 96 to groups of words. To give an example, if the sentence 'Similarly Britain became popular after a rumour got about that Mrs Thatcher had declared open house.' were to be input to the text- to-speech synthesiser, then the output from the parsing process 94 would be:

SENTSTART <ADV Similarly RR ADV > ,_, (NR Britain NP1 NR) [VG became WD VG] <ADJ popular JJ ADJ > [pp after ICS (NR a_AT1 rumour NN1 NR) pp] [VG got VVD about RP VG] that CST (NR Mrs NNSB1 Thatcher NP1 NR) [VG had VHD declared VVN VG] (NR open JJ house NNL1 NR) SENTEND ._.

Where SENTSTART and SENTEND represent the sentence markers 86, RR, NP1 etc. represent the word tag data 95, and < ADV ADV > , (NR NR) etc. represent the syntactic groups 96. The meanings of the word tags used in this description will be understood by those skilled in the art - a subset of the word tags used is given in Table 1 below, a full list can be found in Garside, R., Leech, 0. and Sampson, G . eds 'The Computation Analysis of English : A Corpus based Approach', Longman ( 1 987) .

Table 1

Next, in chunking process 98, the program controls the computer to label 'chunks' in the input sentence. In the present embodiment, the syntactic groups shown in Table 2 below are identified as chunks.

TAG Description Example

IVG Infinite verb group [IVG to TO be VBO IVG] VG (non infinite) verb group [VG was VBDZ beaten_WN VG] com comment phrase < com Well UH com > vpp verb with preposistional [vpp ofJO |J [VG handling_WG VG] particle vpp]

PP preposistional phrase [pp ^"m il (NR practice_NN 1 NR) pp]

NR noun phrase (non referent) (NR Dinamo NPI Kiev_NP1 NR)

R noun phrase (referent) (R it_PPH 1 R)

WH wh-word phrase (WH which_DDQ WH)

QNT quantifier phrase < QNT much_DA1 QNT >

ADV adverb phrase < ADV still_RR ADV >

WHADV wh-adverb phrase < WHADV when_RRQ WHADV >

ADJ adjective phrase < ADJ prone JJ ADJ >

Table 2

The process then divides the input sentence into elements. Chunks are regarded as elements, as are sentence markers, paragraph markers, punctuation marks and words which do not fall inside chunks. Each chunk has a marker applied to it which identifies it as a chunk. These markers constitute chunk markers 99.

The output from the chunking process for the above example sentence is shown in Table 3 below, each line of that table representing an element, and 'phrasetag' representing a chunk marker.

SENTSTART phrasetag( < ADV) Similarly RR

phrasetagf(NR) Britain_NP1 phrasetag([VG) became VVD phrasetag( < ADJ) popular JJ phrasetagtpp afterJCS (NR a_AT1 rumour_NN 1 NR) pp] phrasetag[VG got_WD about_RP VG] that_CST phrasetag(NR Mrs NNSB1 Thatcher NP1 NR) phrasetag[VG had_ VHD declared_WN VG] phrasetag(NR oper 1_JJ house_NNL1 NR)

SENTEND

Table 3

The computer then carries out classification process 100 under control of the program. The classification process 100 uses a classification of words and pronunciation database 100A. The classification database 100A is the fifth of the five databases stored on the CD-ROM 32.

The classification database is divided into classes which broadly correspond to parts- of-speech. For example, verbs, adverbs and adjectives are classes of words. Punctuation is also treated as a class of words. The classification is hierarchical, so many of the classes of words are themselves divided into sub-classes. The subclasses contain a number of word categories which correspond to the word tags 95 applied to words in the input text 40 by the parsing process 94. Some of the sub- classes contain only one member, so they are not divided further. Part of the classification (the part relating to verbs, prepositions and punctuation) used in the present embodiment is given in Table 4 below.

Table 4

It will be seen that the left-hand column of Table 4 contains the classes, the central column contains the sub-classes and the right-hand column contains the word categories. Figure 4 shows part of the classification of verbs. The class of words 'verbs' includes four sub-classes, one of which contains only the word category 'RP' . The other sub-classes ('beverbs', 'doverbs', and 'past') each contain a plurality of word categories. For example, the sub-class 'doverbs' contains the word categories corresponding to the word tags VDO, VDG, VDN, and VDZ.

In carrying out the classification process 100 the computer first identifies a core word contained within each chunk in the input text 40. The core word in a prepositional chunk (i.e. one labelled 'pp' or 'vpp') is the first preposition within the chunk. The core word in a chunk labelled 'WH' or 'WHADV is the first word in the chunk. In all other types of chunk, the core word is the last word in the chunk. The computer then uses the classification of words 100A to label each chunk with the class, sub-class and word category of the core word.

Each non-chunk word is similarly labelled on the basis of the classification of words 100A, as is each piece of punctuation.

The classifications 101 for the elements generated by the classification process 100 are stored in RAM 1 2.

Returning again to the example sentence, after classification of the elements of the input sentence would be as shown in Table 5 below

CLASS = [sentstart ] phrasetag( < ADV) CLASS = [adv ] Similarly RR

CLASS = [punct minpunct ] phrasetagf (NR) CLASS = [nonreferent proper ] Britain NPI phrasetag([VG) CLASS = [vg past ] became VVD phrasetag( < ADJ) CLASS = [adj ] popular JJ phrasetag([pp) CLASS = [pp icspp after ] afterJCS phrasetag ([pp) CLASS = [pp icspp after ] afterJCS

< < SUBCAT phrasetagf (NR) CLASS = [nonreferent ] a_AT1 rumour_NN 1 > > phrasetag(.VG) CLASS = [vg verbpart] got_WD about RP

CLASS = I [lex coords cst ] that CST phrasetag((NR) CLASS = [nonreferent proper place titular] Mrs_NNSB1 Thatcher NP1 phrasetag([VG) CLASS = [vg past ] had VHD declared_WN phrasetag(NR CLASS = [nonreferent locative ] open JJ house_NNL1 NR)

CLASS = [punct majpunct ]

CLASS = [sentend ]

Table 5 It will be seen that each element is labelled with a class and also a sub-class where there are a number of word categories within the sub-class.

Returning to Figure 2A, as stated above, the syntactic information 48 and word grouping data 46 are stored in the RAM 1 2 by the text analysis process 42. The syntactic information 48 comprises word tags 95, syntactic groups 96, chunk markers 99 and element classifications 101 . The word grouping data comprises the sentence markers 86 and paragraph markers 87.

Similar processing is carried out in forming the prosodic structure corpus 52 stored on the CD-ROM 32. Therefore, each of the reference sentences within the corpus is divided into elements and has similar syntactic information relating to each of the elements contained within it. Furthermore, the corpus contains data indicating where a human would insert prosodic boundaries when reading each of the example sentences. The type of the boundary is also indicated.

An example of the beginning of a sentence that might be found in the corpus 52 is given in Table 6 below. In Table 6, the absence of a boundary is shown by the label 'sfNONE' after an element, the presence of a boundary is shown by 'sfMINOR' or 'sfMAJOR' depending on the strength of the boundary. The start of the example sentence is "As ever , | the American public | and the world 's press | are hungry for drama... "

CLASS = [sentstart ] sfNONE phrasetag( < ADV) CLASS = [adv ] As_RG ever RR sfNONE

CLASS = [punct minpunct ] ,_, sfMINOR phrasetagUNR) CLASS = [nonreferent ] the AT American JJ public N sfMINOR

CLASS = [lex coords cc ] and CC sfNONE phrasetagf(NR) CLASS = [nonreferent ] the AT world_NN 1 's_$ press JMN sfMINOR phrasetag([VG) CLASS = [vg beverbs ] are VBR sfNONE phrasetag( < ADJ) CLASS = [adj ] hungry JJ sfNONE phrasetag([pp) CLASS = [pp ifpp for ] forJF < < SUBCAT phrase tag((NR) CLASS = [nonreferent ] drama NN 1 sfNONE > >

Table 6

The prosodic structure prediction process 50 involves the computer in finding the sequence of elements in the corpus which best matches a search sequence taken from the input sentence. The degree of matching is found in terms of syntactic characteristics of corresponding elements, length of the elements in words and a comparison of boundaries in the reference sentence and those already predicted for the input sentence. The process 50 will now be described in more detail with reference to Figure 5.

Figure 5 shows that the process 50 begins with the calculation of measures of similarity between each element of the input sentence and each element of the corpus 52. This part of the program is presented in the form of pseudo-code below:

FOR each element(eι) of the input sentence: FOR each element(er) of the corpus: calculate degree of syntactic match between elements e, and e_r ( = A) calculate no. of words match between elements βι and er ( = B) calculate syntactic match between words in elements e. and er ( = C) match(eι,er) = w1 *A + w2 * B + w3 * C NEXT e_r NEXT e,

where βι increments from 1 to the number of elements in the input sentence, and er increments from 1 to the number of elements in the corpus.

In order to calculate the degree of syntactic match between elements, the program controls the computer to find: a) whether the core words of the two elements are in the same class; and

b) where the two elements are both chunks whether both chunks have the same phrasetag (as seen in Table 2).

A match in both cases might, for example, be given a score of 2, a score of 1 being given for a match in one case, and a score of 0 being given otherwise.

In order to calculate the degree of syntactic match between words in the elements, the program controls the computer to find to what level of the hierarchical classification the corresponding words in the elements are syntactically similar. A match of word categories might be given a score of 5, a match of sub-classes a score of 2 and a match of classes a score of 1 . For example, if the reference sentence has [VG is_VBZ argued VVN VG] and the input sentence has [VG was_VBDZ beaten_WN VG] then 'is VBZ' only matches 'was_VBDZ' to the extent that both are classified as verbs. Therefore a score of 1 would be given on the basis of the first word. With regard to the second word, 'beaten VVN' and 'argued VVN' fall into identical word categories and hence would be given a score of 5. The two scores are then added to give a total score of 6.

The third component of each element similarity measure is the negative magnitude of the difference in the number of words in the reference element, er, and the number of words in the element of the input sentence, eι. For example, if an element of the input sentence has one word and an element of the reference sequence has three words, then the third component is -2.

A weighted addition is then performed on the three components to yield an element similarity measure (match(eι,er) in the above pseudo-code).

Those skilled in the art will thus appreciate that the table calculation step 102 results in the generation of a table giving element similarity measures between every element in the corpus 52 and every element in the input sentence. Then, in step 103, a subject element counter (m) is initialised to 1 . The value of the counter indicates which of the elements of the input sentence is currently subject to a determination of whether it is to be followed by a boundary. Thereafter, the program controls the computer to execute an outermost loop of instructions (steps 104 to 125) repeatedly. Each iteration of the outermost loop of instructions corresponds to a consideration of a different subject element of the input sentence. It will be seen that each execution of the final instruction (step 1 25) in the outermost loop results in the next iteration of the outermost loop looking at the element in the input sentence which immediately follows the input sentence element considered in the previous iteration. Step 1 24 ensures that the outermost loop of instructions ends once the last element in the input sentence has been considered.

The outermost loop of instructions (steps 104 to 125) begins with the setting of a best match value to zero (step 104) . Also, a current reference element count (e_r) is initialised to 1 (step 106).

Within the outermost loop of instructions (steps 104 to 1 25), the program controls the computer to repeat some or all of an intermediate loop of instructions (steps 108 to 1 21 ) as many times as there are elements in the prosodic structure corpus 52. Each iteration of the intermediate loop of instructions (steps 108 to 1 21 ) therefore corresponds to a particular subject element in the input sentence (determined by the current iteration of the outermost loop) and a particular reference element in the corpus 52 (determined by the current iteration of the intermediate loop). Steps 1 20 and 1 21 ensure that the intermediate loop of instructions (steps 108 to 1 21 ) is carried out for every element in the corpus 52 and ends once the final element in the corpus has been considered.

The intermediate loop of instructions (steps 108 to 1 21 ) starts by defining (step 108) a search sequence around the subject element of the input sentence.

The start and end of the search sequence are given by the expressions: srch seq start = mind , m - no of elements before) srch seq end = max( no of input sentence elements, m + no_of_elements_after)

In the preferred embodiment, no of elements before is chosen to be 10, and no_of_elements_after is chosen to be 4. It will be realised that the search sequence therefore includes the current element m, up to 10 elements before it and up to 4 elements after it.

In step 1 10 a sequence similarity measure is reset to zero. In step 1 12 a measure of the similarity between the search sequence and a sequence of reference elements is calculated. The reference sequence has the current reference element (i.e. that set in the previous execution of step 121 ) as it core element. The reference sequence contains this core element as well as the four elements that precede it and the ten elements that follow it (i.e. the reference sequence is of the same length as the search sequence). The calculation of the sequence similarity measure involves carrying out first and second innermost loops of instructions. Pseudo-code for the first innermost loop of instructions is given below:

FOR current position in_srch_seq ( = p) = srch seq start to srch_seq_end s.s.m = s.s.m + weight(p) * match(srch_element_p, corres ref element) NEXT

Where s.s.m is an abbreviation for sequence similarity measure.

In carrying out the steps represented by the above pseudo-code, in effect, the subject element of the input sentence (set in step 103 or 1 25) is aligned with the core reference element. Once those elements are aligned, the element similarity measure between each element of the search sequence and the corresponding element in the reference sequence is found. A weighted addition of those element similarity measures is then carried out to obtain a first component of a sequence similarity measure. The measures of the degree of matching are found in the values obtained in step 102. The weight applied to each of the constituent element matching measures generally increases with proximity to the subject element of the input sentence. Those skilled in the art will be able to find suitable values for the weights by trial and error.

The second innermost loop of instructions then supplements the sequence similarity measure by taking into account the extent to which the boundaries (if any) already predicted for the input sentence match the boundaries present in the reference sequence. Only the part of the search sequence before the subject element is considered since no boundaries have yet been predicted for the subject element or the elements which follow it. Pseudo-code for the second innermost loop of instructions is given below:

FOR current position in srch seq ( = q) = srch seq start to m-1 s.s.m = s.s.m + weight(q) * bdymatch(srch_element_q, corres_ref_element) NEXT

The boundary matching measure between two elements (expressed in the form bdymatch(element x, element y) in the above pseudo-code) is set to two if both the input sentence and the reference sentence have a boundary of the same type after the qth element, one if they have boundaries of different types, zero if neither has a boundary, minus one if one has a minor boundary and the other has none, and minus two if one has a strong boundary and the other has none. A weighted addition of the boundary matching measures is applied, those inter-element boundaries close to the current element being given a higher weight. The weights are chosen so as to penalise heavily sentences whose boundaries do not match.

It will be realised that the carrying out of the first and second innermost loop of instructions results in the generation of a sequence similarity measure for the subject element of the input sentence and the reference element of the corpus 52. If the sequence similarity measure is the highest yet found for the subject element of the input sentence, then the best match value is updated to equal that measure (step 1 16) and the number of the associated element is recorded (step 1 18). Once the final element has been compared, the computer ascertains whether the core element in the best matching sequence has a boundary after it. If it does, a boundary of a similar type is placed into the input sentence at that position (step 1 22).

Thereafter a check is made to see whether the current element is now the final element (step 1 24). If it is, then the prosodic structure prediction process 50 ends (step 1 26). The boundaries which are placed in the input sentence by the above prosodic boundary prediction process (Figure 5) constitute the phrase boundary data (Figure 2A : 54) . The remainder of the text-to-speech conversion process has already been described above with reference to Figure 2B.

In a preferred embodiment of the present invention, boundaries are predicted on the basis of the ten best matching sequences in the prosodic structure corpus. If the majority of those ten sequences feature a boundary after the current element then a boundary is placed after the corresponding element in the input sentence.

In the above-described embodiment pattern matching was carried out which compared an input sentence with sequences in the corpus that included sequences bridging consecutive sentences. Alternative embodiments can be envisaged, where only reference sequences which lie entirely within a sentence are considered. A further constraint can be placed on the pattern matching by only considering reference sequences that have an identical position in the reference sentence to the position of the search sequence in the input sentence. Other search algorithms will occur to those skilled in the art.

The description of the above embodiments describes a text-to-speech program being loaded into the computer from a CD-ROM . It is to be understood that the program could also be loaded into the computer via a computer network such as the Internet.

Claims

1 . A method of converting text to speech comprising the steps of:

receiving an input word sequence in the form of text; comparing said input word sequence with each one of a plurality of reference word sequences provided with prosodic boundary information; identifying one or more reference word sequences which most closely match said input word sequence; and predicting prosodic boundaries for a synthesised spoken version of the input text on the basis of the prosodic boundary information included with said one or more most closely matching reference word sequences.

2. A method according to claim 1 further comprising the step of: identifying clusters of words in the input text which are unlikely to include prosodic phrase boundaries; wherein: said plurality of reference sentences are further provided with information identifying such clusters of words therein; and said comparison step comprises a plurality of per-cluster comparisons.

3. A method according to claim 2 wherein said per-cluster comparison comprises quantifying the degree of similarity between the syntactic characteristics of the clusters.

4. A method according to claim 2 wherein said per-cluster comparison comprises quantifying the degree of similarity between the syntactic characteristics of the words within the clusters.

5. A method according to claim 2 wherein said per-cluster comparison comprises measuring the difference in the number of words in the clusters being compared.

6. A method according to claim 1 wherein said comparison comprises measuring the similarity in the positions of prosodic boundaries previously predicted for the input sentence and the positions of the prosodic boundaries in the reference sentences.

7. A text to speech conversion apparatus comprising:

a word sequence store storing a plurality of reference word sequences which are provided with prosodic boundary information;

a program store storing a program;

a processor in communication with said program store and said store;

means for receiving an input word sequence in the form of text;

wherein said program is executable to control said processor to:

compare said input word sequence with each one of a plurality of said reference word sequences; identify one or more reference word sequences which most closely match said input word sequence; and derive prosodic boundary information for the input text on the basis of the prosodic boundary information included with said one or more most closely matching reference word sequences.

8. A text to speech conversion apparatus comprising:

a word sequence store storing a plurality of reference word sequences which are provided with prosodic boundary information; means arranged in operation to receive an input word sequence in the form of text; means arranged in operation to compare said input text with each one of a plurality of said reference word sequences; means arranged in operation to identify one or more reference word sequences which most closely match said input word sequence; and means arranged in operation to predict prosodic boundaries for the input text on the basis of the prosodic boundary information included with said one or more most closely matching reference word sequences.

9. A program storage device readable by a computer, said device embodying computer readable code executable by the computer to perform method steps according to any one of claims 1 to 6.

10. A signal embodying computer executable code for loading into a computer for the performance of the method according to any one of claims 1 to 6.