CN102033859A

CN102033859A - Method and system for compressing dictionary and processing words, text-to-speed system and electronic equipment

Info

Publication number: CN102033859A
Application number: CN2009101763682A
Authority: CN
Inventors: 亓超; 金浩; 康恒
Original assignee: Canon Inc
Current assignee: Canon Inc
Priority date: 2009-09-28
Filing date: 2009-09-28
Publication date: 2011-04-27
Anticipated expiration: 2029-09-28
Also published as: CN102033859B

Abstract

The invention provides a method for compressing electronic data in a dictionary by using a computer. The method comprises the following steps of: inputting the dictionary to be compressed, wherein words and pronunciations are stored in the dictionary to be compressed; for each Chinese character, determining the pronunciation which occurs most frequently as the default pronunciation, and forming a first character pronunciation mapping table based on the default pronunciation; for each multi-pronunciation Chinese character, determining the other pronunciations except the default pronunciation as the non-default pronunciations, and forming a second character pronunciation mapping table based on the non-default pronunciations; and according to information of an index or a position of a combination of each Chinese character and the pronunciation thereof in the first character pronunciation mapping table or the second character pronunciation mapping table, compressing the combination into a 16-bit code so as to form the compressed dictionary which comprises the information in a 16-bit coding form, wherein the 16-bit code comprises the information of the index of the Chinese character in the first character pronunciation mapping table or the second character pronunciation mapping table and the information of the index of the pronunciation in the first character pronunciation mapping table or the second character pronunciation mapping table.

Description

Dictionary compression and word treatment method and system, text-to-speech system, electronic equipment

Technical field

Present invention relates in general to be used for method and system, word treatment method and system, text-to-speech system and the electronic equipment of compression of electronic dictionary.

Background technology

Use the electronic equipment of electronic dictionary to be widely used in a plurality of fields.Electronic dictionary usually adopts the data structure that speech and the associating information relevant with this speech are stored.For Chinese electronic dictionary (hereinafter being also referred to as " dictionary "), the information relevant with Chinese word can comprise the lexical or textual analysis information and the pronunciation information of this speech.

In the Chinese electronic dictionary of majority, pronunciation information is a pith of word information.Usually, Chinese pronunciations information can directly be stored in the electronic dictionary.Yet,, in PDA(Personal Digital Assistant), wish electronic dictionary is minimized taking of storer if electronic dictionary is installed in the equipment of storage condition difference.

Usually, each Chinese word in the electronic dictionary has the pronunciation information of himself, and for example, the pronunciation of speech " Country family " is " guo2jia1 ", wherein the tone in the numeral Chinese pronunciations system such as " 2 " and " 1 ".Obviously, the pronunciation character string usually needs to preserve more byte than speech itself.

Chinese patent CN02159546.1 discloses a kind of method that is used for the pronunciation of 2 bytes (16 bit) storage Chinese character.Therefore, the required memory space of storage pronunciation information is reduced for the 2*N byte, and wherein N represents the quantity of the Chinese character that electronic dictionary comprises.To use Big5 word collection that 2 bytes encode to a Chinese character as example, the memory space that the combination of Chinese character and its pronunciation will take will be 4 bytes (32 bit).

Yet,, still need considerable memory space to store pronunciation information for comprising thousands of electronic dictionaries as the speech of the combination of Chinese character.For example, suppose total total about 70000 speech in electronic dictionary, it then need preserve pronunciation information more than the 273k byte by forming more than 140000 Chinese characters.

In fact, Duo Shuo Chinese character only has a pronunciation.For example, Chinese character “ Wind " pronunciation be " feng1 ".Some Chinese characters have a plurality of pronunciations, and it is called as " multitone Chinese character ".For example, the pronunciation of Chinese character " poor " comprises " cha4 ", " cha1 ", " chai1 ", " ci1 ".In some cases, the tone of some Chinese characters in speech will become gently and read.For example, the pronunciation of Chinese character " Hot " is " re4 ", Chinese character “ Downtown " pronunciation be " nao4 ", but speech “ Hot Downtown " pronunciation be " re4nao5 " (numeral " 5 " expression gently read).That is to say that though a Chinese character only has an acquiescence pronunciation as its conventional pronunciation, the tone of this pronunciation may become gently in some cases to be read.

Chinese patent application CN200310114889.8 discloses the method for a kind of utilization bytes store pronunciation information still less, and this method may further comprise the steps: distinguish the multitone Chinese character from other Chinese characters; There is the acquiescence pronunciation table that is used for all Chinese characters; There is the non-acquiescence pronunciation table that is used for the multitone Chinese character; Produce supplementary for the speech that comprises the multitone Chinese character, described supplementary is represented the non-acquiescence pronunciation of these multitone Chinese characters; And speech is stored in the electronic dictionary with described supplementary relatedly.According to disclosed method in Chinese patent application CN200310114889.8, the multitone Chinese character that only has non-acquiescence pronunciation need be stored with the supplementary of the non-acquiescence pronunciation of indication, and pronounce to show by means of Chinese character being stored acquiescence pronunciation table and non-acquiescence, the storage of the acquiescence pronunciation of Chinese character becomes no longer necessary.The method can the memory space that the pronunciation of Chinese character is shared be reduced to average each speech and be less than 1 byte.

Yet, according to disclosed method among the Chinese patent application CN200310114889.8, for the combination of Chinese character and its pronunciation, on average need be more than 2 bytes.Because to the demand of electronic equipment more cheaply, need Chinese character and its pronunciation to occupy than still less electronic dictionary of memory block all in the past.

Summary of the invention

The present inventor notices the following inherent law or the fact of Chinese language: with Big5 word collection as an example, Big5 word collection has 13060 unsimplified Hanzis at most, it has 1295 valid utterances, wherein have only 943 Chinese characters to have a plurality of pronunciations, and the maximum pronunciation countings that Chinese character had are 6; On the other hand, Big5 word collection uses 2 bytes (16 bit) that a Chinese character is encoded, and still concentrates at the Big5 word to also have a large amount of codings not represent Chinese character.

Consider the technical matters that in aforesaid prior art, exists and the above-mentioned inherent law or the fact of Chinese language, still less byte new method and system that Chinese character or Chinese word are stored with its pronunciation of a kind of use is provided.

The universal character set that such as GB2312 other are used for Chinese character also has similar characteristic.Therefore, though use Big5 word collection that principle of the present invention is described in some embodiment or example, the method that is used to present Chinese character is not limited to Big5 word collection, and can use any other method that presents Chinese character.

According to an aspect of the present invention, provide a kind of method of utilizing the electronic data in the computing machine compression dictionary, comprising: input step, import dictionary to be compressed, store speech and its pronunciation with the electronic data form in the described dictionary to be compressed; The first word sound mapping table forms step, for each Chinese character in the dictionary to be compressed, a pronunciation is defined as the acquiescence pronunciation, and forms the first word sound mapping table based on the acquiescence pronunciation; The second word sound mapping table forms step, for each the multitone Chinese character in the dictionary to be compressed, will be defined as non-acquiescence pronunciation except that the residue pronunciation the acquiescence pronunciation, and forms the second word sound mapping table based on non-acquiescence pronunciation; And compression step, according to being combined in the index in the first word sound mapping table or the second word sound mapping table or the information of position about each Chinese character in each speech in the dictionary to be compressed and its pronunciation, should make up boil down to 16 bits of encoded, the compression dictionary that comprises information with formation with 16 bits of encoded forms, wherein, described 16 bits of encoded comprise about this Chinese character in the first word sound mapping table or the second word sound mapping table index or the information of position, and about the pronunciation corresponding with this Chinese character in the first word sound mapping table or the second word sound mapping table index or the information of position.

According to another aspect of the present invention, provide a kind of system that utilizes the electronic data in the computing machine compression dictionary, comprising: input media, import dictionary to be compressed, store speech and its pronunciation with the electronic data form in the described dictionary to be compressed; The first word sound mapping table forms device, for each Chinese character in the dictionary to be compressed, a pronunciation is defined as the acquiescence pronunciation, and forms the first word sound mapping table based on the acquiescence pronunciation; The second word sound mapping table forms device, for each the multitone Chinese character in the dictionary to be compressed, will be defined as non-acquiescence pronunciation except that the residue pronunciation the acquiescence pronunciation, and forms the second word sound mapping table based on non-acquiescence pronunciation; And compression set, according to being combined in the index in the first word sound mapping table or the second word sound mapping table or the information of position about each Chinese character in each speech in the dictionary to be compressed and its pronunciation, should make up boil down to 16 bits of encoded, the compression dictionary that comprises information with formation with 16 bits of encoded forms, wherein, described 16 bits of encoded comprise about this Chinese character in the first word sound mapping table or the second word sound mapping table index or the information of position, and about the pronunciation corresponding with this Chinese character in the first word sound mapping table or the second word sound mapping table index or the information of position.

According to a further aspect of the invention, a kind of word treatment method that is used for electronic equipment is provided, described electronic equipment comprises the compression dictionary, each Chinese character in each speech in the described compression dictionary and the combination of its pronunciation have utilized the method according to this invention to be compressed into 16 bits of encoded, described method comprises: word sound mapping table obtaining step, obtain the first word sound mapping table and the second word sound mapping table; And decompression step, utilize the described first word sound mapping table or the second word sound mapping table that 16 bits of encoded corresponding with the Chinese character in the speech in the described compression dictionary and the combination of its pronunciation are decompressed.

According to a further aspect of the invention, a kind of speech disposal system that is used for electronic equipment is provided, described electronic equipment comprises the compression dictionary, each Chinese character in each speech in the described compression dictionary and the combination of its pronunciation have utilized the method according to this invention to be compressed into 16 bits of encoded, described system comprises: word sound mapping table deriving means, obtain the first word sound mapping table and the second word sound mapping table; And decompressing device, utilize the described first word sound mapping table or the second word sound mapping table that 16 bits of encoded corresponding with the Chinese character in the speech in the described compression dictionary and the combination of its pronunciation are decompressed.

According to a further aspect of the invention, provide a kind of electronic equipment, described electronic equipment comprises: the compression dictionary, and each Chinese character in each speech in the described compression dictionary and the combination of its pronunciation have utilized the method according to this invention to be compressed into 16 bits of encoded; And according to speech disposal system of the present invention.

According to a further aspect of the invention, provide a kind of compression dictionary that utilizes with the text-to-speech system of text-converted for voice, described text-to-speech system comprises: the compression dictionary, and each Chinese character in each speech in the described compression dictionary and the combination of its pronunciation have utilized the method according to this invention to be compressed into 16 bits of encoded; Text input device is used for input text; Text processing apparatus is used for according to the compression dictionary described text dividing being become speech and being institute's predicate phonetic notation; And the speech production device, be used for producing voice based on the result of described text processing apparatus.

According to a further aspect of the invention, provide a kind of electronic equipment, comprising: according to text-to-speech system of the present invention; Be used for and at least one of screen, keyboard and the mouse of described text input device interface; And be used for and at least one of loudspeaker, earphone and the display of described speech production device interface.

Benefit from according to said method of the present invention and system, can realize that storage space that Chinese character and its pronunciation take is than still less electronic dictionary all in the past.Particularly, for example, Big5 word collection uses 2 bytes (16 bit) to compress a Chinese character.According to the present invention, the combination of Chinese character and its pronunciation only takies 2 bytes (16 bit), this with equate according to the shared storage space of the Chinese character of Big5 word collection itself.That is to say that except 16 bits that are used for Chinese character itself traditionally, pronunciation information will not take any extra or additional storage space.

In addition, all will bring inconvenience even have more 1 bit, and efficient is reduced implementing storage than 2 bytes (16 bit).Therefore, to make the combination of Chinese character and its pronunciation be 16 bits just reducing aspect the storage space that will use is very useful.

From following description with reference to accompanying drawing, other property features of the present invention and advantage will become clear.

Description of drawings

The accompanying drawing of incorporating an instructions and a formation instructions part into illustrates embodiments of the invention, and is used from explanation principle of the present invention with description one.

Fig. 1 is the block diagram that the hardware configuration of the computer system 1000 that can implement embodiments of the invention is shown.

Fig. 2 is that diagram is according to the block diagram of utilizing the system 2000 that computing machine compresses the electronic data in the dictionary of the present invention.

Fig. 3 is the block diagram of diagram according to the speech disposal system of using in comprising the electronic equipment that compresses dictionary 3000 of the present invention.

Fig. 4 is the block diagram of diagram according to text-to-speech system 4000 of the present invention.

Fig. 5 is used for implementing the process flow diagram that utilizes the method for the electronic data that computing machine compresses dictionary according to of the present invention.

Fig. 6 illustrates the process flow diagram that is used to form the preferred process of the first word sound mapping table and the second word sound mapping table according to of the present invention.

Fig. 7 is the process flow diagram that is used for implementing according to the exemplary compression process of the step S540 of Fig. 5 of the present invention.

Fig. 8 is that diagram utilizes compression dictionary according to the present invention to handle the process flow diagram of the method for speech.

Fig. 9 is that diagram utilizes compression dictionary according to the present invention to come the process flow diagram of method of the speech of process user input.

Figure 10 is the process flow diagram that diagram text-to-speech system 4000 utilizes the process that compression dictionary according to the present invention carries out.

Embodiment

Describe embodiments of the invention in detail hereinafter with reference to accompanying drawing.

Note that similar reference number and letter refer among the figure similarly project, thereby in case in a width of cloth figure, defined a project, needn't after figure in this project is discussed again.

In this manual, " dictionary " is meant Chinese electronic dictionary, and " word " is meant Chinese character.

In this manual, the pronunciation of the acquiescence of Chinese character can be one of pronunciation of this Chinese character.For example, acquiescence pronunciation can be the pronunciation of frequent use on statistics of this Chinese character.For example, acquiescence pronunciation can be a priori to find people's pronunciation of frequent use in daily life, perhaps can be for the maximum pronunciation of this Chinese character occurrence number in all pronunciations corresponding with this Chinese character in the dictionary.Typically, a Chinese character only has an acquiescence pronunciation, and in some cases, the tone of this acquiescence pronunciation can become gently to be read.

In addition, the non-acquiescence pronunciation of multitone Chinese character is meant other pronunciations except its acquiescence pronunciation.The multitone Chinese character can have one or more non-acquiescence pronunciations.

In this manual, the Chinese character in the combination of speech and its pronunciation and this speech and the combination of its pronunciation have identical meanings.That is to say that the combination of speech and its pronunciation means a plurality of combinations, each combination all is made of a Chinese character and its pronunciation in this speech.

In this manual, the speech of compression is meant wherein all Chinese characters and all compressed speech of its pronunciation, and the word of compression is meant that this word is compressed into the Chinese character of 16 bits of encoded with its pronunciation according to the present invention.

As shown in Figure 1, computer system comprises computing machine 1110.Computing machine 1110 comprises processing unit 1120, system storage 1130, fixed non-volatile memory interface 1140, mobile non-volatile memory interface 1150, user's input interface 1160, network interface 1170, video interface 1190 and the output peripheral interface 1195 that connects via system bus 1121.

System storage 1130 comprises ROM (ROM (read-only memory)) 1131 and RAM (random access memory) 1132.BIOS (Basic Input or Output System (BIOS)) 1133 resides in the ROM 1131.Operating system 1134, application program 1135, other program module 1136 and some routine data 1137 reside in the RAM 1132.

Fixed non-volatile memory 1141 such as hard disk is connected to fixed non-volatile memory interface 1140.Fixed non-volatile memory 1141 for example can storage operating system 1144, application program 1145, other program module 1146 and some routine data 1147.

Mobile nonvolatile memory such as floppy disk 1151 and CD-ROM drive 1155 is connected to mobile non-volatile memory interface 1150.For example, floppy disk can be inserted in the floppy disk 1151, and CD (CD) can be inserted in the CD-ROM drive 1155.

Input equipment such as mouse 1161 and keyboard 1162 is connected to user's input interface 1160.

Computing machine 1110 can be connected to remote computer 1180 by network interface 1170.For example, network interface 1170 can be connected to remote computer 1180 by LAN (Local Area Network) 1171.Perhaps, network interface 1170 can be connected to modulator-demodular unit (modulator-demodulator) 1172, and modulator-demodular unit 1172 is connected to remote computer 1180 via wide area network 1173.

Remote computer 1180 can comprise the storer 1181 such as hard disk, and it can store remote application 1185.

Video interface 1190 is connected to monitor 1191.

Output peripheral interface 1195 is connected to printer 1196 and loudspeaker 1197.

Computer system shown in Figure 1 only is illustrative and never means any restriction to invention, its application, or uses.

Computer system shown in Figure 1 can be incorporated in any embodiment, can be used as stand-alone computer, also can be used as the disposal system in the electronic equipment, can remove one or more unnecessary assemblies, also can add one or more additional assemblies to it.For example, when computer system 1000 is used for electronic dictionary equipment, electronic learning machine, personal digital assistant, mobile phone, video camera or multi-function peripheral, for example can not comprise floppy disk 1151 and CD-ROM drive 1155.

Fig. 2 is that diagram utilizes computing machine to compress the block diagram of the system 2000 of the electronic data in the dictionary.As shown in Figure 2, system 2000 comprises: input media 2100, import dictionary to be compressed, and store speech and its pronunciation with the electronic data form in the described dictionary to be compressed; The first word sound mapping table forms device 2200, for each Chinese character in the dictionary to be compressed, a pronunciation is defined as the acquiescence pronunciation, and forms the first word sound mapping table based on the acquiescence pronunciation; The second word sound mapping table forms device 2300, for each the multitone Chinese character in the dictionary to be compressed, will be defined as non-acquiescence pronunciation except that the residue pronunciation the acquiescence pronunciation, and forms the second word sound mapping table based on non-acquiescence pronunciation; And compression set 2400, according to being combined in the index in the first word sound mapping table or the second word sound mapping table or the information of position about each Chinese character in each speech in the dictionary to be compressed and its pronunciation, should make up boil down to 16 bits of encoded, the compression dictionary that comprises information with formation with 16 bits of encoded forms, wherein, described 16 bits of encoded comprise about this Chinese character in the first word sound mapping table or the second word sound mapping table index or the information of position, and about the pronunciation corresponding with this Chinese character in the first word sound mapping table or the second word sound mapping table index or the information of position.

Fig. 3 is shown in the block diagram that comprises the speech disposal system of using in the electronic equipment that compresses dictionary 3000.As shown in Figure 3, speech disposal system 3000 comprises: word sound mapping table deriving means 3300, obtain the first word sound mapping table and the second word sound mapping table; And decompressing device 3400, utilize the described first word sound mapping table or the second word sound mapping table that 16 bits of encoded corresponding with the Chinese character in the speech in the described compression dictionary and the combination of its pronunciation are decompressed.

Speech disposal system 3000 can further comprise: speech input media 3100, import speech by the user; Searcher 3200, utilize the compressed format of the first word sound mapping table and the second word sound mapping table speech of search input in the compression dictionary, the compressed format of the speech of described input is made up of a plurality of 16 bits of encoded, each 16 bits of encoded is represented the combination of a Chinese character and its pronunciation, wherein said decompressing device 3400 utilizes the first word sound mapping table or the second word sound mapping table, and each 16 bits of encoded in the compressed format of institute's predicate that will search out from the compression dictionary de-compress into the combination of this Chinese character and its pronunciation; And output unit 3500, the pronunciation of the speech that obtains by described decompressing device 3400 by text and/or voice output.

Described decompressing device 3400 can further comprise: first determines device 3410, determines that according to the predetermined bit in described 16 bits of encoded it still is the second word sound mapping table that this combination is arranged in the first word sound mapping table; And second determine device 3430, determine that described first device 3410 definite these combinations are arranged under the situation of the first word sound mapping table, according to the 14 predetermined bits in the remaining bits determine about this Chinese character in the first word sound mapping table index or the information of position, and determine that according to a remaining bit whether this pronunciation is for gently reading; Perhaps, determine that described first device 3410 definite these combinations are arranged under the situation of the second word sound mapping table, according to the 12 predetermined bits in the remaining bits determine about this Chinese character in the second word sound mapping table index or the information of position, and determine information about this position of pronunciation in the second word sound mapping table of this Chinese character according to remaining 3 bits.

Fig. 4 is the block diagram of diagram text-to-speech system 4000, described text-to-speech system 4000 is used for utilizing the compression dictionary and text-converted is become voice, described text-to-speech system 4000 comprises: the compression dictionary, and each Chinese character in each speech in the described compression dictionary and the combination of its pronunciation have utilized the method according to this invention to be compressed into 16 bits of encoded; Text input device 4100 is used for input text; Text processing apparatus 4300 is used for according to the compression dictionary described text dividing being become speech and being institute's predicate phonetic notation; And speech production device 4500, be used for producing voice based on the result of described text processing apparatus 4300.

Described text processing apparatus 4300 can preferably include: text is cut speech device 4310, and being used for text dividing is speech, and speech disposal system 3000.

More than device is the exemplary and/or preferred module that is used to implement the following process that will describe.The module that is used to implement each step is not below at large described.Yet,, just can be useful on the functional module or the device of the correspondence of implementing same process as long as the step of certain process of execution is arranged.The technical scheme that all combinations by step described below and the device corresponding with these steps limit all is included in the disclosure of the specification, as long as these technical schemes that their constitute are complete and applicable.

In addition, the said system that is made of each device can be incorporated in the hardware device such as electronic equipment as functional module.Except these functional modules, these electronic equipments certainly have other hardware or component software.

With reference to the method for Fig. 5-7 description according to the Chinese electronic dictionary of compression of the present invention.

In step S510, import dictionary to be compressed.Form with electronic data in the dictionary to be compressed stores speech and its pronunciation.

In step S520, form the first word sound mapping table (be also referred to as acquiescence pronunciation table, will be described in more detail below).

In step S530, form the second word sound mapping table (be also referred to as non-acquiescence pronunciation table, will be described in more detail below).

In step S540, based on the described first word sound mapping table or the second word sound mapping table, each Chinese character of each speech in the dictionary to be compressed and each combination of its pronunciation are compressed into 16 bits of encoded.

Fig. 6 is the process flow diagram that the preferred process that is used to form the first word sound mapping table (step S520) and the second word sound mapping table (step S530) is shown.Though note that step S520 and step S530 is two steps of separating in the process flow diagram of Fig. 5, these two steps can serial implement also can walk abreast enforcement.Process flow diagram shown in Figure 6 has provided the wherein example of step S520 and step S530 executed in parallel.The flow process of Fig. 6 left side branch (i.e. the flow process that is made of step S610, S620, S630, S640, S660 and S670) is corresponding to step S520, and the flow process of Fig. 6 right branch (i.e. the flow process that is made of step S610, S620, S630, S650, S660 and S670) is corresponding to step S530.

At first, first initial word sound mapping table and the second initial word sound mapping table for example are empty.

In step S610, obtain a Chinese character in the dictionary, and in step S620, consider a pronunciation of this Chinese character.

Determine that in step S630 whether the pronunciation considered is the acquiescence pronunciation of the Chinese character that obtains in step S610 in step S620.This acquiescence pronunciation can be any one pronunciation of this Chinese character.For example first pronunciation of handling for this Chinese character in the processing procedure can be pronounced as acquiescence.Be used to judge that whether pronunciation is that the criterion of the acquiescence pronunciation of this Chinese character also can be for example based on the statistical information of obtaining in advance.As mentioned above, the acquiescence of Chinese character pronunciation can be for the pronunciation of frequent use on this Chinese character statistics.For example, acquiescence pronunciation can be a priori to find people's pronunciation of frequent use in daily life, perhaps can be for the maximum pronunciation of this Chinese character occurrence number in all pronunciations corresponding with this Chinese character in the dictionary.Typically, a Chinese character only has an acquiescence pronunciation, and in some cases, the tone of this acquiescence pronunciation can become gently to be read.

If determining the pronunciation of considering in step S620 in step S630 is the acquiescence pronunciation (being "Yes" among the step S630) of this Chinese character, then process advances to step S640.In step S640, upgrade this first word sound mapping table by add the information relevant to the first word sound mapping table, and process advances to step S660 with the combination of this Chinese character and the pronunciation of this acquiescence.

Otherwise, not the acquiescence pronunciation (being non-acquiescence pronunciation) (being "No" among the step S630) of this Chinese character if in step S630, determine the pronunciation of in step S620, considering, then process advances to step S650.In step S650, upgrade this second word sound mapping table by add the information relevant to the second word sound mapping table, and process advances to step S660 with the combination of this Chinese character and this non-acquiescence pronunciation.

In step S660, determine whether still have another pronunciation not consider for current Chinese character.If still have another pronunciation not consider (being "Yes" among the step S660) for current Chinese character, then process advances to step S620 to consider the pronunciation that next is not considered for current Chinese character.If do not have other pronunciations to consider (being "No" among the step S660) for current Chinese character, then process advances to step S670 has any Chinese character of not considering to determine whether still to remain in dictionary.If still residue has the Chinese character of not considering (being "Yes" among the step S670) in dictionary, then process advances to step S610 to obtain next Chinese character.If in dictionary, do not remain the Chinese character of not considering (being "No" among the step S670), then finish formation, and the first word sound mapping table and the second word sound mapping table can be used for other processing to the first word sound mapping table and the second word sound mapping table.

Table 1 illustrates an illustrative examples of the first word sound mapping table (acquiescence pronunciation table).Note that table 1 is an example wherein using Big5 word collection.Yet, also can use other word collection.

Table 1: the first word sound mapping table

Table 2 illustrates an illustrative examples of the second word sound mapping table (non-acquiescence pronunciation table).Note that table 2 is examples wherein using Big5 word collection.Yet, also can use other word collection.

Table 2: the second word sound mapping table

As can be seen from Table 1, the first word sound mapping table can comprise following information.For a Chinese character, the first word sound mapping table can comprise: the acquiescence pronunciation (16 bit) of index (T1_INDEX) (14 bit), this Chinese character itself (16 bits for example have form according to the coding of universal character set such as Big5 or GB2312) and this Chinese character of this Chinese character in the first word sound mapping table.In the first word sound mapping table, because each Chinese character only has an acquiescence pronunciation, clauses and subclauses of the Chinese character in the first word sound mapping table are corresponding to the combination of this Chinese character and the pronunciation of its acquiescence.Although not shown in table 1, the acquiescence pronunciation can be rendered as conventional tone or gently read.In 16 bits of encoded, can have the pronunciation of indication acquiescence and whether get the information of gently reading.Though this information may not present in the first word sound mapping table, this information can be considered in the first word sound mapping table for an index that pronounces.

As can be seen from Table 2, the second word sound mapping table can comprise following information.For a Chinese character, the second word sound mapping table can comprise: the one or more non-acquiescence pronunciation (each 16 bit) of index (T2_INDEX) (12 bit), this Chinese character itself (16 bits for example have form according to the coding of universal character set such as Big5 or GB2312) and this Chinese character of this Chinese character in the second word sound mapping table.In the second word sound mapping table, because each Chinese character has one or more non-acquiescence pronunciations, therefore the clauses and subclauses of the non-acquiescence pronunciation of the Chinese character in the second word sound mapping table are corresponding to all non-acquiescence pronunciations, each non-acquiescence pronunciation has its own column index, and the non-acquiescence pronunciation of this Chinese character in the second word sound mapping table can be counted as the combination of pronouncing corresponding to this Chinese character and this non-acquiescence.

In above-mentioned this mode, every kind of combination of Chinese character and its pronunciation is registered in the first word sound mapping table or the second word sound mapping table.

Note that the first word sound mapping table shown in table 1 and the table 2 and the second word sound mapping table only are used for the example illustrative purposes, and be not intended to limit the form of the first word sound mapping table and the second word sound mapping table.For example, Chinese character in the first word sound mapping table and the second word sound mapping table and/or pronunciation can be taked other forms but not the form shown in table 1 and the table 2.

The index of the Chinese character in the first word sound mapping table and the second word sound mapping table can at random design, as long as a Chinese character is corresponding to an index in table.Preferably, this index increases progressively from 0 open numbering and with 1, and making needs less bit with each index binarization.Notice that the index of Chinese character in the first word sound mapping table can be different with the index of this Chinese character in the second word sound mapping table.Because the Chinese character much less in the second word sound mapping table (because only have polyphone to have non-acquiescence pronunciation thereby be registered in the second word sound mapping table), the index of the Chinese character in the second word sound mapping table takies still less bit than the index in the first word sound mapping table.

Get back to Fig. 5, in step S540, each combination of each Chinese character in each speech in the dictionary to be compressed and its pronunciation is compressed to 16 bits of encoded based on the information that is combined in index in the first word sound mapping table or the second word sound mapping table or position about this, so that form the compression dictionary that comprises the information with 16 bits of encoded forms.

Fig. 7 is the process flow diagram that is used for the exemplary compression process of implementation step S540.

In step S710, obtain the speech in the dictionary to be compressed.In step S720, obtain the Chinese character in this speech and the combination of its pronunciation.In step S730, judge whether this pronunciation is the acquiescence pronunciation.If judge this pronunciation in step S730 is acquiescence pronunciation (being "Yes" among the step S730), and then process advances to step S740, and in step S740, according to the first word sound mapping table, this combination is compressed to 16 bits of encoded.If judge this pronunciation in step S730 is not acquiescence pronunciation (be "No" among the step S730, be non-acquiescence pronunciation), and then process advances to step S750, and in step S750, according to the second word sound mapping table, this combination is compressed to 16 bits of encoded.Compressed in step S740 or S750 after the combination of this Chinese character and its pronunciation, process advances to step S760 to determine whether unpressed Chinese character is arranged in this speech.If determine to still have in this speech unpressed Chinese character (being "Yes" among the step S760) in step S760, then process advances to step S720 to obtain next Chinese character and its pronunciation in this speech.If determine there is not unpressed Chinese character in this speech (being "No" among the step S760) in step S760, then process advances to step S770 to determine in this dictionary whether any unpressed speech being arranged.If determine that in step S770 unpressed speech is arranged in the dictionary (being "Yes" among the step S770), then process advances to step S710 to obtain the next speech in the dictionary.If in step S770, determine there is not unpressed speech in the dictionary (being "No" among the step S770), then finished the compression of dictionary, and obtained the compression dictionary.

Particularly, 16 above-mentioned bits of encoded comprise: about Chinese character in the first word sound mapping table or the second word sound mapping table index or the information of position, and about the pronunciation corresponding with this Chinese character in the first word sound mapping table or the second word sound mapping table index or the information of position.

For example, described 16 bits of encoded can have following exemplary information.Particularly, on behalf of the combination of this Chinese character and its pronunciation, 1 bit in described 16 bits of encoded can be in the first word sound mapping table still in the second word sound mapping table.Under described combination is in situation in the first word sound mapping table, 14 bits in the remaining bits can be represented about the index of this Chinese character in the first word sound mapping table or the information of position (T1_INDEX that example is as shown in table 1), and remaining 1 bit is that gently reading of gently reading indicates as the tone of representing this pronunciation.Though not shown in the table 1, the bit that representative is gently read is actually the index of acquiescence pronunciation in the first word sound mapping table.Otherwise, under this combination is in situation in the second word sound mapping table, 12 bits in the remaining bits can be represented about the index of this Chinese character in the second word sound mapping table or the information of position (T2_INDEX that example is as shown in table 2), remaining 3 bits representative is about the information (for example, indicating this non-acquiescence pronunciation to be positioned at the information of which row of the second word sound mapping table) of the position of the pronunciation corresponding with this Chinese character in the second word sound mapping table.

For example, 16 bits of encoded " 0011101101100000 " for the combination of representing one of a Chinese character and its pronunciation, first bit " 0 " represents that this pronunciation is the acquiescence pronunciation, 14 bits " 01110110110000 " then represent that (we can know from table 1 index (T1_INDEX) of this Chinese character in the first word sound mapping table, T1_INDEX " 01110110110000 " represents Chinese character " shadow "), last 1 bit " 0 " represents that this pronunciation is not gently to read.

Lift another example, 16 bits of encoded " 1000001101011000 " for another combination of representing one of Chinese character and its pronunciation, first bit " 1 " represents that this pronunciation is non-acquiescence pronunciation, 12 bits " 000001101011 " then represent that (we can know from table 2 index (T2_INDEX) of this Chinese character in the second word sound mapping table, T2_INDEX " 000001101011 " expression Chinese character " falls "), last 3 bits " 000 " are represented this position of pronunciation in the second word sound mapping table, that is pronunciation " dao4 ".

Get Big5 word collection as an illustrative examples, the first word sound mapping table can cover all Chinese characters of Big5 word centralized definition.By this way, each index in the first word sound mapping table can have one-to-one relationship or mapping with each Chinese character that the Big5 word is concentrated.There is a large amount of applicable mapping methods.As example, following formula provides a kind of simple method that is used for the Big5 code conversion is become the index of each Chinese character.

If L＜128 then T1_INDEX=(H-0xA4) * 0xA0+L-0x40;

Otherwise T1_INDEX=(H-0xA4) * 0xA0+L-0x60,

Wherein, H is first byte of the Big5 coding of this Chinese character, and L is second byte of the Big5 coding of this Chinese character, and T1_INDEX is the index of this Chinese character in the first word sound mapping table.Can be similarly from this index to the conversion of Big5 coding, as follows:

H＝T1_INDEX/160+164；

If T1_INDEX%160≤64 then L=(T1_INDEX%160)+64

Otherwise L=(T1_INDEX%160)+96

Wherein, T1_INDEX/160 means by with the integral part of T1_INDEX divided by 160 merchants that obtained, and T1_INDEX%160 mean by with T1_INDEX divided by 160 remainders that obtained.

Note that owing to concentrating to comprise 16239 Chinese characters, therefore in the first word sound mapping table, have 16239 clauses and subclauses corresponding just enough with these 16239 Chinese characters at the Big5 word.That is to say that above-mentioned formula never limits the concrete mapping method between Big5 coding and the index (T1_INDEX).Preferably, index is since 0, and increases progressively with 1.That is to say that for example under the situation of using Big5 word collection, the index in the first word sound mapping table was from 0 to 16238 (it can be represented by 14 bits).

About the index in the second word sound mapping table, can between multitone Chinese character and index (T2_INDEX), set up similar corresponding relation or mapping.Because only 943 Chinese characters to be arranged are polyphones, therefore the index of the Chinese character in the second word sound mapping table takies still less bit (being 12 bits) than the index of the Chinese character in the first word sound mapping table under the situation of Big5 word collection.

Though note that above with Big5 word collection as example, the method that is used to compress dictionary according to the present invention also can be applied to the Chinese character represented by other Chinese Character Sets.

Now, for the method that is used to compress dictionary according to the present invention, will a concrete example be described.

In the process of compression dictionary, speech and its pronunciation " inverted image (dao4ying3) " are transfused to, and will be compressed.

At first, Chinese character " falls " and is acquired with its pronunciation, and be compressed into 16 bits of encoded 1000001101011000 (0x86B0), wherein, first bit " 1 " expression pronunciation " dao4 " is the non-acquiescence pronunciation that Chinese character " falls ", 12 bits " 000001101011 " expression Chinese character then " falls " index (T2_INDEX) in the second word sound mapping table, last 3 bits " 000 " are represented this position of pronunciation in the second word sound mapping table, that is, pronunciation " dao4 " is first non-acquiescence pronunciation that this Chinese character " falls ".Then, Chinese character " shadow " is acquired with its pronunciation, and be compressed into 16 bits of encoded 0011101101100000 (0x3B60), wherein first bit " 0 " expression pronunciation " ying3 " is the acquiescence pronunciation of this Chinese character " shadow ", the index (T1_INDEX) of 14 bits " 01110110110000 " expression Chinese character " shadow " in the first word sound mapping table then, last bit " 0 " represents that this pronunciation is not gently to read.Then, two 16 bits of encoded of the speech of compression-promptly corresponding with these two Chinese characters and its pronunciation-be stored in the dictionary with compressed format, described two 16 bits of encoded are represented Chinese word " inverted image " and its pronunciation " dao4ying3 ".

Method according to the Chinese electronic dictionary of compression of the present invention has more than been described.Step in the above method can be carried out on computers by computer program.

Utilize described compression dictionary to handle the method for speech now with reference to Fig. 8 and Fig. 9 description.

Fig. 8 is that diagram utilizes the compression dictionary to handle the process flow diagram of the method for speech.Described compression dictionary is according to the compressed dictionary of one of above-mentioned method of the present invention.That is to say that by utilizing one of above-described the method according to this invention, each Chinese character in each speech in the described compression dictionary and each of its pronunciation combination have been compressed to 16 bits.

In step S810, obtain the first word sound mapping table and the second word sound mapping table.For example, the first word sound mapping table and the second word sound mapping table can be written into from server or memory device etc.The first word sound mapping table and the second word sound mapping table can be one of the above-mentioned types.

In step S820, decompress by using the 16 corresponding bits of encoded of Chinese character and the combination of its pronunciation in the speech in the first word sound mapping table or the second word sound mapping table pair and the compression dictionary.

Therefore, represent Chinese character and 16 bits of encoded of the combination of its pronunciation in the speech decompressed, and can present this combination.

Fig. 9 is that diagram utilizes the compression dictionary to come the process flow diagram of method of the speech of process user input.Described compression dictionary is according to the compressed dictionary of one of above-mentioned method of the present invention.That is to say that by utilizing one of above-described the method according to this invention, each Chinese character of each speech in the described compression dictionary and each of its pronunciation combination have been compressed to 16 bits of encoded.In method shown in Figure 9, the speech of user input is processed, each Chinese character of this speech of and the combination of this speech and its pronunciation-promptly and each combination of its pronunciation-be output.

In step S910, import a speech by the user.A speech can be made of a plurality of Chinese characters.

In step S920, obtain the first word sound mapping table and the second word sound mapping table.For example, the first word sound mapping table and the second word sound mapping table can be written into from server or memory device etc.The first word sound mapping table and the second word sound mapping table can be one of the above-mentioned types.

In step S930, by the first word sound mapping table and the second word sound mapping table, the compressed format of the speech of search input in the compression dictionary.The compressed format of the speech of input is meant by using the above-mentioned compressed speech of method (for example seeing the description to step S540).That is to say that the described compressed format of the speech of input is made of a plurality of 16 bits of encoded, each described 16 bits of encoded is represented the Chinese character in this speech and the combination of its pronunciation.By execution in step S930, can determine the pronunciation of the speech imported.

Below be to be used for implementing by the first word sound mapping table and the second word sound mapping table and in a kind of illustrative methods of the step S930 of the compressed format of the speech of compression dictionary search input.

At first, at the first word sound mapping table and the second word sound mapping table each Chinese character in this speech of search in the two.The first word sound mapping table (acquiescence pronunciation mapping table) comprises the clauses and subclauses of all Chinese characters, and existence is comprised 16 bits of encoded of the index (T1_INDEX) of each Chinese character in this speech in the first word sound mapping table.Yet, therefore some Chinese characters (for example if not the multitone Chinese character) can not appear in the second word sound mapping table (non-acquiescence pronunciation mapping table), are not that each Chinese character in the speech all necessarily has 16 bits of encoded that are included in the index (T2_INDEX) in the second word sound mapping table.

Then, carry out matching process with one of each combination of the speech (constituting) of judging which compression in the compression dictionary and 16 bits of encoded of these Chinese characters coupling by a plurality of 16 bits of encoded.For example, if a speech is made of two Chinese characters, and 16 bits of encoded (c2_1) that first Chinese character has 16 bits of encoded (c1_1) of the hypothesis that comprises T1_INDEX (T1_INDEX_1) and comprises the hypothesis of T2_INDEX (T2_INDEX_1), second 16 bits of encoded (c2_2) that Chinese character has 16 bits of encoded (c1_2) of the hypothesis that comprises T1_INDEX (T1_INDEX_2) and comprises the hypothesis of T2_INDEX (T2_INDEX_2), then comprise (c1_1 in the search compression dictionary, c1_2), (c2_1, c1_2), (c1_1, c2_2) and (c2_1, c2_2) speech of one of compression.First bit that comprises 16 bits of encoded (for example c1_1 or c1_2) of the hypothesis of T1_INDEX for example can be " 0 ", and it is used to indicate it to be based on 16 bits of encoded of the first word sound mapping table.Similarly, first bit that comprises 16 bits of encoded (for example c2_1 or c2_2) of the hypothesis of T2_INDEX for example can be " 1 ", and it is used to indicate it to be based on 16 bits of encoded of the second word sound mapping table.Can determine described 14 bit index (T1_INDEX) and described 12 bit index (T2_INDEX) for specific Chinese character is unique.Other undetermined bits can be any possible bits, and can not consider in matching process.

After in step S930, in the compression dictionary, finding the compressed format of comprising of this speech of a plurality of 16 bits of encoded, in step S940, each combination that in described a plurality of 16 bits of encoded each is extracted and shortens Chinese character and its pronunciation into is so that by utilizing the first word sound mapping table and/or the second word sound mapping table to obtain the combination of this speech and its pronunciation.

In step S950, the combination of the pronunciation of this speech or this speech and its pronunciation is output.The pronunciation of this speech can be exported by text and/or voice.

According to one embodiment of present invention, step S820 or the S940 that is used to decompress can comprise following step:

-the first determining step determines that according to a predetermined bit in described 16 bits of encoded (for example first bit) it still is the second word sound mapping table that the combination of Chinese character and pronunciation is arranged in the first word sound mapping table; And

-the second determining step, in first determining step, determine under the situation that this combination is arranged in the first word sound mapping table, according to the 14 predetermined bits in the remaining bits determine about this Chinese character in the first word sound mapping table index or the information of position, and determine according to a remaining bit whether this pronunciation is gently to read, perhaps, in first determining step, determine under the situation that this combination is arranged in the second word sound mapping table, according to the 12 predetermined bits in the remaining bits determine about this Chinese character in the second word sound mapping table index or the information of position, and determine information about this position of pronunciation in the second word sound mapping table of this Chinese character according to remaining 3 bits.

Now, for utilizing compression dictionary according to the present invention to handle the method for the speech of input, a concrete example will be described.

The user imports speech " inverted image ", and expects the output of its pronunciation.

At first, Chinese character " falls " the obtained 000101011111001X of being of 16 bits of encoded (C1_1) with the combination of its acquiescence pronunciation, wherein, first bit " 0 " represents that this pronunciation is in the first word sound mapping table, the the 2nd to the 15th bit " 00101011111001 " is represented the index (T1_INDEX_1) of this Chinese character in the first word sound mapping table, and last bit " X " is got any possible values.In addition, Chinese character " falls " the obtained 1000001101011XXX of being of 16 bits of encoded (C2_1) of combination of acquiescence non-with it pronunciation, wherein, first bit " 1 " represents that this pronunciation is in the second word sound mapping table, the the 2nd to the 13rd bit " 000001101011 " is represented the index (T2_INDEX_1) of this Chinese character in the second word sound mapping table, and last 3 bits " XXX " are got any possible values.In addition, Chinese character " shadow " is obtained with 16 bits of encoded (C1_2) of the combination of its acquiescence pronunciation to be 001110110110000X, wherein, first bit " 0 " represents that this pronunciation is in the first word sound mapping table, the the 2nd to the 15th bit " 01110110110000 " is represented the index (T1_INDEX_2) of this Chinese character in the first word sound mapping table, and last bit " X " is got any possible values.Because Chinese character " shadow " is not the multitone Chinese character, and do not have the pronunciation of non-acquiescence, do not exist about 16 bits of encoded (C2_2) of the index (T2_INDEX_2) of this Chinese character in the second word sound mapping table.

Then, comprise in the search compression dictionary (C1_1, C1_2) or (C2_1, the speech of compression C1_2).

As a result, found and comprised (C2_1, the speech of C1_2) compression corresponding in the compression dictionary with " inverted image dao4ying3 ".

Then, utilize the speech of the compression that the first word sound mapping table and/or the second word sound mapping table will find to decompress, and the pronunciation " dao4ying3 " that obtains and export this speech.

With reference to Fig. 8 and Fig. 9 the method for utilizing the compression dictionary to handle speech has been described.Step in the said method can be carried out on computers by computer program.

The method of above-mentioned compression Chinese electronic dictionary can be implemented as a system (for example above-mentioned system 2000) by computer program, and the method for utilizing the compression dictionary to handle speech also can be implemented as a system (for example above-mentioned system 3000) by computer program.In this case, each step is corresponding to a functional module (device), and each system can be contained in the other system by the computer program implementation and operation, perhaps can be installed in the electronic equipment, as functional module.

For example, the text-to-speech system 4000 that is used for input text is converted to voice can comprise above-mentioned system 3000.

Figure 10 illustrates by text-to-speech system 4000 to utilize process flow diagram according to the performed process of the compression dictionary of above-mentioned any compression method compression.Each step that note that the process flow diagram among Figure 10 is corresponding to each device in the system among Fig. 4 4000.

In step S1010, for example by user input text.This step S1010 is corresponding to the text input device among Fig. 4 4100.

In step S1020, be speech with text cutting according to the compression dictionary.With text dividing is that the process of speech can be carried out by any prior art.This step S1020 cuts speech device 4310 corresponding to the text among Fig. 4.

Then, process advances to step S910, and the combination of each speech and its pronunciation is processed according to process flow diagram shown in Figure 9.By the processing according to process flow diagram shown in Figure 9, the pronunciation of speech can be determined, but and mark with phonetic symbols.Then, in step S950, export result, and process advances to the step S1030 of the process flow diagram shown in Figure 10 as the pronunciation of each speech in the text.According to this process of the process flow diagram among Fig. 9 corresponding to the speech disposal system 3000 among Fig. 3.

Above-mentioned steps S1020 with according to the combination of the process of the process flow diagram of Fig. 9 corresponding to the text processing apparatus among Fig. 4 4300.

In step S1030, generate voice based on result.This step S1030 is corresponding to the speech production device 4500 among Fig. 4.

According to the said method of in text-to-speech system 4000, carrying out, can utilize speech disposal system 3000 of the present invention to realize the conversion of literary composition language.

Further, one or more in the said system can be installed in the electronic equipment such as electronic dictionary equipment, electronic learning machine, personal digital assistant, mobile phone, video camera or multi-function peripheral.

For example, described electronic equipment can comprise according to compression dictionary of the present invention, and above-mentioned speech disposal system 3000 or text-to-speech system 4000.Preferably, described electronic equipment has and is used for and at least one of screen, keyboard and the mouse of its input media interface, and has and be used for and at least one of loudspeaker, earphone and the display device of its output unit interface.

According to the present invention, the memory requirement of electronic dictionary being stored pronunciation information reduces greatly.Particularly, by using the present invention, can reduce the size of the pronunciation information that is used for a speech.

At the basic electronic dictionary that is used for NLP (natural language processing) module, total total about 70000 speech, and the longest speech is made of four Chinese characters, and wherein 33525 speech do not comprise any Chinese character with a plurality of pronunciations, and 28014 speech only comprise a Chinese character with a plurality of pronunciations.

By being applied in the method that proposes among the Chinese patent application CN200310114889.8, the storage extra memory space that pronunciation information spent greater than the 27k byte less than the 43k byte.On the other hand, by utilizing the method according to this invention, extra memory space reduces to 0, and this is because store pronunciation without any need for extra bit.

Can implement method and system of the present invention by many modes.For example, can implement method and system of the present invention by software, hardware, firmware or its any combination.The order of above-mentioned method step only is illustrative, and method step of the present invention is not limited to above specifically described order, unless otherwise offer some clarification on.In addition, in certain embodiments, the present invention can also be implemented as the program that is recorded in the recording medium, and it comprises the machine readable instructions that is used to realize the method according to this invention.Thereby the present invention also covers the recording medium that storage is used to realize the program of the method according to this invention.

Though by the example detail display specific embodiments more of the present invention, it will be appreciated by those skilled in the art that above-mentioned example only is intended that exemplary and does not limit the scope of the invention.It should be appreciated by those skilled in the art that the foregoing description to be modified and do not depart from the scope and spirit of the present invention.Scope of the present invention is to limit by appended claim.

Claims

1. method of utilizing the electronic data in the computing machine compression dictionary comprises:

Input step is imported dictionary to be compressed, stores speech and its pronunciation with the electronic data form in the described dictionary to be compressed;

The first word sound mapping table forms step, for each Chinese character in the dictionary to be compressed, a pronunciation is defined as the acquiescence pronunciation, and forms the first word sound mapping table based on the acquiescence pronunciation;

The second word sound mapping table forms step, for each the multitone Chinese character in the dictionary to be compressed, will be defined as non-acquiescence pronunciation except that the residue pronunciation the acquiescence pronunciation, and forms the second word sound mapping table based on non-acquiescence pronunciation; And

Compression step, according to being combined in the index in the first word sound mapping table or the second word sound mapping table or the information of position about each Chinese character in each speech in the dictionary to be compressed and its pronunciation, should make up boil down to 16 bits of encoded, the compression dictionary that comprises information with formation with 16 bits of encoded forms

Wherein, described 16 bits of encoded comprise about this Chinese character in the first word sound mapping table or the second word sound mapping table index or the information of position, and about the pronunciation corresponding with this Chinese character in the first word sound mapping table or the second word sound mapping table index or the information of position.

2. the method for claim 1, wherein

Maximum pronunciations is confirmed as the acquiescence pronunciation for this Chinese character occurrence number in the middle of all pronunciations corresponding with this Chinese character.

3. method as claimed in claim 2, wherein

In the middle of all pronunciations corresponding with described Chinese character, maximum pronunciations is confirmed as the acquiescence pronunciation for this Chinese character occurrence number in dictionary to be compressed.

4. the method for claim 1, wherein

A bit in described 16 bits of encoded represents that it still is the second word sound mapping table that described combination is arranged in the first word sound mapping table;

Be arranged in this combination under the situation of the first word sound mapping table, 14 bits in the remaining bits represent about this Chinese character in the first word sound mapping table index or the information of position, remaining bit is as gently reading sign;

Be arranged in this combination under the situation of the second word sound mapping table, 12 bits in the remaining bits represent about this Chinese character in the second word sound mapping table index or the information of position, and remaining 3 bits are represented the information about the position of the pronunciation corresponding with this Chinese character in the second word sound mapping table.

5. system that utilizes the electronic data in the computing machine compression dictionary comprises:

Input media is imported dictionary to be compressed, stores speech and its pronunciation with the electronic data form in the described dictionary to be compressed;

The first word sound mapping table forms device, for each Chinese character in the dictionary to be compressed, a pronunciation is defined as the acquiescence pronunciation, and forms the first word sound mapping table based on the acquiescence pronunciation;

The second word sound mapping table forms device, for each the multitone Chinese character in the dictionary to be compressed, will be defined as non-acquiescence pronunciation except that the residue pronunciation the acquiescence pronunciation, and forms the second word sound mapping table based on non-acquiescence pronunciation; And

Compression set, according to being combined in the index in the first word sound mapping table or the second word sound mapping table or the information of position about each Chinese character in each speech in the dictionary to be compressed and its pronunciation, should make up boil down to 16 bits of encoded, the compression dictionary that comprises information with formation with 16 bits of encoded forms

6. system as claimed in claim 5, wherein

7. system as claimed in claim 6, wherein

8. system as claimed in claim 5, wherein

9. word treatment method that is used for electronic equipment, described electronic equipment comprises the compression dictionary, each Chinese character in each speech in the described compression dictionary and the combination of its pronunciation have utilized according to each described method among the claim 1-4 and have been compressed into 16 bits of encoded, and described method comprises:

Word sound mapping table obtaining step obtains the first word sound mapping table and the second word sound mapping table; And

Decompression step utilizes the described first word sound mapping table or the second word sound mapping table that 16 bits of encoded corresponding with the Chinese character in the speech in the described compression dictionary and the combination of its pronunciation are decompressed.

10. word treatment method as claimed in claim 9 also comprises:

The speech input step is imported speech by the user;

Search step, utilize the compressed format of the first word sound mapping table and the second word sound mapping table speech of search input in the compression dictionary, the compressed format of the speech of described input is made up of a plurality of 16 bits of encoded, each 16 bits of encoded is represented the combination of a Chinese character and its pronunciation, wherein in described decompression step, utilize the first word sound mapping table or the second word sound mapping table, each 16 bits of encoded in the compressed format of institute's predicate that will search out from the compression dictionary de-compress into the combination of this Chinese character and its pronunciation; And

The output step, the pronunciation of the speech that in described decompression step, obtains by text and/or voice output.

11. as claim 9 or 10 described word treatment methods, wherein, described decompression step comprises:

First determining step determines that according to the predetermined bit in described 16 bits of encoded it still is the second word sound mapping table that this combination is arranged in the first word sound mapping table; And

Second determining step, determine that in described first determining step this combination is arranged under the situation of the first word sound mapping table, according to the 14 predetermined bits in the remaining bits determine about this Chinese character in the first word sound mapping table index or the information of position, and determine that according to a remaining bit whether this pronunciation is for gently reading; Perhaps, determine that in described first determining step this combination is arranged under the situation of the second word sound mapping table, according to the 12 predetermined bits in the remaining bits determine about this Chinese character in the second word sound mapping table index or the information of position, and determine information about this position of pronunciation in the second word sound mapping table of this Chinese character according to remaining 3 bits.

12. speech disposal system that is used for electronic equipment, described electronic equipment comprises the compression dictionary, each Chinese character in each speech in the described compression dictionary and the combination of its pronunciation have utilized according to each described method among the claim 1-4 and have been compressed into 16 bits of encoded, and described system comprises:

Word sound mapping table deriving means obtains the first word sound mapping table and the second word sound mapping table; And

Decompressing device utilizes the described first word sound mapping table or the second word sound mapping table that 16 bits of encoded corresponding with the Chinese character in the speech in the described compression dictionary and the combination of its pronunciation are decompressed.

13. speech disposal system as claimed in claim 12 also comprises:

The speech input media is imported speech by the user;

Searcher, utilize the compressed format of the first word sound mapping table and the second word sound mapping table speech of search input in the compression dictionary, the compressed format of the speech of described input is made up of a plurality of 16 bits of encoded, each 16 bits of encoded is represented the combination of a Chinese character and its pronunciation, wherein said decompressing device utilizes the first word sound mapping table or the second word sound mapping table, and each 16 bits of encoded in the compressed format of institute's predicate that will search out from the compression dictionary de-compress into the combination of this Chinese character and its pronunciation; And

Output unit, the pronunciation of the speech that obtains by described decompressing device by text and/or voice output.

14. as claim 12 or 13 described speech disposal systems, wherein, described decompressing device comprises:

First determines device, determines that according to the predetermined bit in described 16 bits of encoded it still is the second word sound mapping table that this combination is arranged in the first word sound mapping table; And

Second determines device, determine that at described first definite device this combination is arranged under the situation of the first word sound mapping table, according to the 14 predetermined bits in the remaining bits determine about this Chinese character in the first word sound mapping table index or the information of position, and determine that according to a remaining bit whether this pronunciation is for gently reading; Perhaps, determine that at described first definite device this combination is arranged under the situation of the second word sound mapping table, according to the 12 predetermined bits in the remaining bits determine about this Chinese character in the second word sound mapping table index or the information of position, and determine information about this position of pronunciation in the second word sound mapping table of this Chinese character according to remaining 3 bits.

15. an electronic equipment, described electronic equipment comprises:

The compression dictionary, each Chinese character in each speech in the described compression dictionary and the combination of its pronunciation have utilized according to each described method among the claim 1-4 and have been compressed into 16 bits of encoded; And

According to each described speech disposal system among the claim 12-14.

16. electronic equipment as claimed in claim 15, wherein, described electronic equipment is one of in electronic dictionary equipment, electronic learning machine, personal digital assistant, mobile phone, video camera and the multi-function peripheral.

17. electronic equipment as claimed in claim 15 also comprises being used for and at least one of screen, keyboard and the mouse of institute predicate input media interface.

18. electronic equipment according to claim 15 also comprises being used for and at least one of loudspeaker, earphone and the display device of described output unit interface.

19. one kind is utilized the compression dictionary with the text-to-speech system of text-converted for voice, described text-to-speech system comprises:

The compression dictionary, each Chinese character in each speech in the described compression dictionary and the combination of its pronunciation have utilized according to each described method among the claim 1-4 and have been compressed into 16 bits of encoded;

Text input device is used for input text;

Text processing apparatus is used for according to the compression dictionary described text dividing being become speech and being institute's predicate phonetic notation; And

The speech production device is used for producing voice based on the result of described text processing apparatus.

20. text-to-speech system according to claim 19, wherein said text processing apparatus comprises:

Text is cut the speech device, and being used for text dividing is speech; And

According to each described speech disposal system among the claim 12-14.

21. an electronic equipment comprises:

According to claim 19 or 20 described text-to-speech systems;

Be used for and at least one of screen, keyboard and the mouse of described text input device interface; And

Be used for and at least one of loudspeaker, earphone and the display of described speech production device interface.

22. electronic equipment according to claim 21, wherein, described electronic equipment is one of in electronic dictionary equipment, electronic learning machine, personal digital assistant, mobile phone, video camera and the multi-function peripheral.