CN1121028C - Solid pronouncing and literal learning machine and the making method of special composite pronouncing and literal tape - Google Patents

Abstract

The present invention provides a solid speech and literal learning machine and a fabricating method of a special composite speech and literal tape thereof. The solid speech and literal learning machine comprises an earphone, a microphone, a speech and literal magnetic head signal jack, a telephone line interface, a demodulator, a front end circuit of the demodulator, an analog multiplexer, an analog signal processor, a speech coder/decoder, a digital signal processor, a main controller, a speech and prompting tone memory, a literal and literal pool memory, a displayer, a computer communication interface, etc. The solid speech and literal learning machine enables a user to see various kinds of playback in foreign languages or Chinese literals on a display screen while hearing the speech after the solid speech and literal learning machine enters speech and literal composite signals recorded on a cassette magnetic tape. The present invention can exchange digital information with a computer. The present invention can be used as a telephone answering machine when not used as the learning machine.

Description

The solid pronouncing and literal learning machine

Technical field

The present invention relates to pronunciation-learning system, particularly a kind of solid pronouncing and literal learning machine.

Background technology

With sound-track engraving apparatus foreign language studying often for owing to fast rewind, forward (FWD) makes tape around unrest even twine extremely and puzzle one's brains, so present popular language playback device just occurred.There is audio record circuit language repeater inside; wherein the last one section voice playing of tape are being deposited in protection all the time; the length of this section voice does not wait from several seconds to a few minutes according to language repeater class difference; when pressing " re-reading " key; tape just stops to play, and the voice in the sound circuit come out with regard to repeat playing, the part that allows the user catch not hear; continue to play tape again by " re-reading " key again, thereby greatly facilitate foreign language learner.But language repeater be can't see corresponding character when hearing voice, must seek on books.Turn over book again, it is very inconvenient to operate language repeater again.(patent No. is reference paper 1: 92229942.0) disclose a kind of language learner that can see the foreign language literal of the interval interior correspondence of multiple language, but the literal that this learning machine shows when pressing " re-reading " key probably is disorderly.Because pressing that " re-reading " is strong is at random fully, if be not exactly in a certain character-coded beginning when pressing " re-reading " key, literal is play at the beginning will be disorderly, demonstration may be indescribable mess code.If just in time be that the broadcast of literal can normally begin in a certain character-coded beginning when pressing, but can not guarantee normal termination.Because end point is limited by the language repeater memory size only, generally can exactly not drop on the place that character code just in time finishes, so re-reading second time the time literal show also can be disorderly.(patent No. is reference paper 2: 85100922.A) disclose a kind of foreign language video and audio recorder, its core technology writes down voice respectively with two tracks of magnetic tape cassette and literal is play, but limit by technical conditions, scheme is perfect inadequately, to the letter-by-letter input of literal also be recorded to coding on the tape simultaneously with keyboard, doing like this cannot be synchronous with voice recording, even have and the corresponding voice synchronous typing of institute's input alphabet, single alphabetical sound also can not be formed with the continuant of foreign language " charm ".Not too suitable in the reference paper 3 (patent No. is 86207130.U) with the LED display text, and rewinding repeatedly, forward (FWD) inevitably when playing with tape, every designed method for synchronous that all deposits in from the display random access memory zero-address of this patent also can become nonsensical in this case.Above-mentioned three patents all have a common defective be that volume is big, power consumption is many, not portable.Reference paper 4 (patent No. 90100160.0) is disclosed to be a kind of method for making of language and characters composite tape, but because this method is all wanted switch sound-track engraving apparatus motor when every typing, this makes and can occur ear-piercing " click is rattled away " sound at every interval when playing the tape that so produces continuously, so be not suitable for using in the foreign language academic environment.

Summary of the invention

The objective of the invention is to overcome above-mentioned deficiency, a kind of any mechanical component that do not comprise is provided, form fully, resemble playing card hand held solid of a size sound style learning machine by large scale integrated circuit and electronic devices and components.Use it can make the user when hearing speech play, on LCD, can see corresponding foreign language or Chinese written language.For reducing cost, it is the voice of strict synchronism and the normal stereo cassette audio tape of letter signal that main external memory medium adopts on two tracks record respectively to have with sentence, section.Transcribe quality in order to increase external memory type and raising, the dataphone interface is equipped with, can easily the confidential language and characters complex data of working out in advance in the computer of asking of satisfied basis be imported this sound literary composition machine, also can deposit the language and characters complex data in this sound literary composition machine in computer.For increasing the service life, also be furnished with telephone line interface, when not doing the use of sound literary composition machine, can be used as answering machine and forever use.

The objective of the invention is to be achieved by the following scheme, the solid pronouncing and literal learning machine by earphone, microphone, voice magnetic head signal socket, literal magnetic head signal socket, telephone line interface, detuner and front end circuit, simulate Port Multiplier and analog signal processing circuit, audio coder ﹠ decoder (codec), digital signal processor, master controller, voice and prompt tone flash memory, LCD, dataphone interface, push-button array, battery and charging circuit etc. constitute.Said earphone, microphone and voice magnetic head signal socket, telephone line interface are connected with simulation Port Multiplier and analog signal processing circuit respectively, these signals are switched and nurse one's health by the latter and deliver to the back level behind the suitable level and continue to handle, telephone line interface also is the output channel of voice signal simultaneously, it is connected to simulation Port Multiplier and analog signal processing circuit with earphone, and the voice signal that the back level is sent here is by simulation Port Multiplier and analog signal processing circuit switching and be amplified to suitable level and power; Simulation Port Multiplier and analog signal processing circuit are connected by bidirectional lines with audio coder ﹠ decoder (codec) (CODEC), the voice signal that prime can be sent here carries out becoming after AD changes the mathematics time series and is sent to the back level, also back grade sequence digit time that send here can be carried out become analog voice signal after AD changes and deliver to prime; Digital signal processor (DSP) is connected with master controller (MCU), Voice decoder and voice and prompt tone flash memory with bidirectional lines, make the digital signal processor deposit voice and prompt tone flash memory in after digitized voice signal compresses fetching from prime under the control of master controller, also can under the control of master controller, from voice and prompt tone people storer, take out and be sent to prime after the digital voice data that has compressed decompresses; Literal and character library flash memory are connected with master controller is two-way, so that the lteral data that receives is deposited in literal and the character library flash memory by master controller, or with the literal taking-up that will show in literal and the character library flash memory; Literal magnetic head signal socket is connected with detuner and front end circuit, and simultaneously detuner and front end circuit are connected with master controller, is sent to master controller after the literal code signal demodulation of literal magnetic head signal socket being read by detuner and front end circuit; LCD, dataphone interface and the push-button array in addition that are connected with master controller simultaneously.

The present invention by dump on the external memory medium in inner voice and prompt tone flash memory speech data and the literal code data in literal and the character library flash memory keeping with sentence, the section be the strict corresponding relation of unit.

The said telephone line interface of the present invention is connected with simulation Port Multiplier and analog signal processing circuit, thereby when not doing the study use, this learning machine can be used as the circuit answer phone and uses.

The present invention goes into computer by built-in parallel dataphone interface with the complex data unloading that has recorded in the internal memory, also the normal data in the computer can be imported internal memory, and then the exterior storage medium that uses can be all computer software storage mediums such as disk, CD.

Advantage of the present invention is: can hear voice simultaneously and see literal when foreign language studying, can be that unit does any redirect and repetition with sentence, section.And language and characters keeps corresponding relation automatically in this process; Literal is play sentence by sentence after can also establishing behind the literal voice or first voice earlier, is convenient to the user and reads etc. from reading, carry on the back to read or translate.Secondly, the volume of learning machine has only the playing card size, and very light in weight can be contained in coat pocket and carry everywhere.Be built-in with the 300-500mAh rechargeable battery, but continuous working is more than 10 hours, for outdoor application provides convenience.Learning machine also is provided with the dataphone interface except that the tool function of language learning, other expanded functions such as telephone line interface, as can be used as interview machine, voice monitoring machine, answering machine etc., improved the cost performance of this sound literary composition machine greatly.

Description of drawings

Below in conjunction with accompanying drawing the present invention is illustrated:

Fig. 1 is a circuit function block scheme of the present invention;

Fig. 2 is a process flow diagram of making wave file;

Fig. 3 is the segment table format chart of record segmentation markers position;

Fig. 4 is the sentence sheet format figure of record subordinate sentence mark position;

Fig. 5 is a process flow diagram of recording sound literary composition data from the language and characters composite tape;

Fig. 6 is an embodiment of the invention front-end circuit schematic diagram;

Fig. 7 is the digital signal processor part schematic diagram of the embodiment of the invention;

Fig. 8 is the master controller part schematic diagram of the embodiment of the invention;

Fig. 9 is the liquid crystal display-driving part schematic diagram of the embodiment of the invention;

Figure 10 is the liquid crystal display displays part schematic diagram of the embodiment of the invention;

Figure 11 is the part schematic diagram of the embodiment of the invention.

Embodiment

As shown in Figure 1: the present invention is by earphone 1, microphone 2, voice magnetic head signal socket 3, literal magnetic head signal socket 4, telephone line interface 5, detuner and front end circuit 6, simulation Port Multiplier and analog signal processing circuit 7, Voice decoder (CODEC) 8, digital signal processor (DSP) 9, master controller (MCU) 10, voice and prompt tone flash memory 11, literal and character library flash memory 12, LCD 13, dataphone interface 14, push-button array 15, formations such as battery and charging circuit 16.Simulation Port Multiplier and analog signal processing circuit 7 are that the signal of microphone 12, voice magnetic head signal socket 3 and telephone line interface 5 switches and nurses one's health and delivers to the back level behind the suitable level and continue to handle, and also the voice signal that late-class circuit can be sent here is delivered to earphone 1 sounding or through telephone line interface 5 prompt tone is sent to phone or calling party's one end after being amplified to suitable level and power.The effect of Voice decoder 8 is the voice signal that prime is sent here to be carried out becoming after AD changes sequence digit time (just digitized voice signal) be sent to the back level, also back grade sequence digit time that send here can be carried out become analog voice signal after DA changes and deliver to prime.Digital signal processor 9 deposits voice and prompt tone flash memory 11 in after the voice signal of fetching numeral under the control of primary controller 10 from prime compresses, be sent to prime also can take out the digital speech digital solution compression of having compressed from voice and prompt tone flash memory 11 under the control of master controller 10 after.When being used for learning machine and using, all deposit the digital voice data that has compressed in voice and the prompting flash memory 11 in.When being used for answering machine and using, voice and prompt tone flash memory 11 are divided into two districts, deposit the prompt tone of prerecording to the calling party for one, and another deposits the calling party's message that has compressed.Master controller 10 carries out work from each parts of commands coordinate that receive the user by strong array 15 by user's requirement, except that 9 work of control figure signal processor, also want the literal code digital signal sent here after the literal code signal demodulation of receiving demodulation device and front end circuit 6 with literal magnetic head signal socket 4 modulation, it is stored in literal and the character library flash memory 12, and set up address search relation one to one with being stored in speech data in voice and the prompt tone flash memory 11, deliver to LCD 13 and show synchronously so that in the speech play process, corresponding lteral data can be obtained corresponding character pattern data through the character library conversion.Dataphone interface 14 deposits speech data in two parts flash memory and literal complex data in two flash memories of this sound literary composition machine under the coordination control of master controller 10 and computer.

Dataphone interface 14 among Fig. 1 adopts the parallel communication mode, link to each other with the printer interface of PC, but the driver of PC side must rewrite.Limit by the historical reasons of PC printer adapter design, 8 position datawires can only unidirectionally use.Thereby 4 condition lines of dataphone interface 14 usefulness of this sound literary composition machine are as the data line that transmits data to PC.Like this, what transmit that data use to this sound literary composition machine from PC is 8 position datawires, and what transmit that data use to PC from sound literary composition machine is 4 position datawires.The latter is than the former doppio lunghezza di tempo, but still ratio is faster with serial line interface.

Fig. 6 to Figure 11 provided with Fig. 1 in each module have the specific embodiments of better corresponding relation.J102 is a receiver J-Horner 1; J103 is a microphone socket 2; Half is a voice magnetic head signal socket 3 on the J104; J101 and U101 constitute telephone line interface 5; 1,2,3 and 5 all is to link to each other with simulation Port Multiplier and analog signal processing circuit 7 that a part by U102 and U104 constitutes, and the latter can determine which signal condition is sent to the back level to suitable level, also the signal of back level can be delivered to prime.If what the back level was sent is the normal speech signal, is amplified to suitable power through power amplifier U102 and delivers to the earphone pronunciation; If the back level is a telephone speech signal, then, after handling, U101 delivers on the telephone wire again through the 5 pin input of U101.U101 is special-purpose telephone line interface IC, comprises polarity transformation circuits, simulation circuit for answering, speech treatment circuit, the anti-side-tone circuit necessary circuit of line interface of expecting someone's call in the base.Its 1 pin directly is connected with telephone wire with 2 pin, and when not having ringing signal, U101 is equivalent to disconnect with telephone wire; When ringing signal is arranged, if Nobody answers in 3 to 5, then internal simulation circuit for answering action, externally present off hook state, sound circuit and telephone wire are connected in the U101, the prompt tone that 5 pin are sent here is delivered to U104 with calling party's message through 7 pin then and is handled toward the calling party of the telephone wire other end earlier.U104 is an audio coder ﹠ decoder (codec) 8, it is connected via Data In-Line CODEC-DIN, DOL Data Output Line CODEC-DOUT, bit synchronization line CODEC-BCLK and frame synchronization line CODEC-SYNC with digital signal processor 9U201, effect is that the voice signal with prime carries out becoming after the AD conversion digit time sequence and is sent to the back level, or sequence digit time of will the back level sending here is carried out becoming analog voice signal after the DA conversion and delivered to prime.

U201 is a digital signal processor 9 among Fig. 7, it and master controller 10U301 are through Data In-Line MWDIN, DOL Data Output Line MWDOUT, (oblique line of signal name front represents that this signal is that low level is effective to condition line/MWRDY, be equivalent to the upward line of signal name in the circuit diagram, request of data line/MWRQST, clock line MWCLK links to each other with chip select line/MWCS, simultaneously with voice suggestion sound flash memory 11U202 through data/address line D[0:7] link to each other with some state of a control lines, can make the U201 deposit U202 in after digitized voice signal compresses fetching from prime under the control of U301 like this, also can under the control of U301, from U202, take out and be sent to prime after the digital voice data that has compressed decompresses.U301 is a master controller 10 among Fig. 8, it is except with digital signal processor 9U201 is connected, and also is connected with literal and character library flash memory 12U302, drive part U401, detuner and the front end circuit 6U103 of LCD 13 and the dataphone interface of being made up of U303 and J301 14.The method of attachment of U301 and U302 is identical with the method for attachment of U201 and U202; What be connected with U401 is realized by chip select line/LCD-CS, data line LCD-SDA and clock line LCD-SCL; With being connected of U103 be data line DEM-DAT and clock line DEM-BCK.U103 is special-purpose detuner, and its input pin AIN (1 pin) links to each other with the literal magnetic head signal socket 4 of following half formation of J104.The effect of U103 is to obtain the literal code digital signal after the literal code simulating signal demodulation that 1 pin is received to be sent to U301 through 7 pin, 8 pin; U401 is the character mode lcd display driver, and it can receive the Word message that U301 sends here and be presented at (being the J501 among Figure 11) on the LCDs.Based on these connections, just can be with the literal code digital signal behind the U103 receiving demodulation under software systems control, be stored among the U302, and be stored in speech data among the U202 and set up address search relation one to one, deliver to U401 so that in the speech play process, corresponding word data can be obtained corresponding character pattern data after the character library conversion, and then on LCDs J501, show synchronously.The connection of master controller U301 and computer is connected on the J301 through U303, the printer interface of J301 and computer directly is connected with parallel cable, transmit 8 bit data bus of data to J301 from PC with the D0-D7 conduct, / STB ,/SLCTIN and BUSY be as the handshake line ,/ACK, PE, SLCT and/ERR is as the 4 bit data general lines that transmit data from J301 to PC.Take this, this sound literary composition machine can deposit speech data in two parts flash memory and literal in computer by the corresponding retrieval relation in address under the coordination control of U301 and PC, and also the language and characters complex data that computer can be met the demands deposits in two flash memories of this sound literary composition machine.Because U401 itself has the key scanning function, so the push-button array of being made up of SW401-SW409 15 directly is connected on the U401, user's input of control commands thus operates this sound literary composition machine.Cell voltage can be transformed to suitable voltage based on the power circuit of U601 and be sent to everywhere, and can when voltage is lower than certain predetermined value, report to the police by warp/BAT-LOW line notice master controller 10.D601-D603 and Z601 form simple battery charger.Need to prove that Fig. 1 is a functional block diagram, during actual design, multiple scheme can be arranged, wherein the selection of digital signal processor (F claims DSP) is the key of whole design.State of development by present technology, DSP can have three kinds of selection schemes: the one, select the DSP in the telegraphone, the voice compression algorithm and the software that comprises DTMF tone generation and telegraphone functions such as detection and the monitoring of process sound of standard wherein have been solidified with, make the Control Software of master controller (calling MCU in the following text) fairly simple like this, but the general power consumption of this class DSP is bigger.Provide the company of this DSP a lot, as SIEMENS, ISD, PHILPS, LUCENT, DSP Grounp, Digital VoiceSystem.Inc, or the like.The 2nd, select solid recorder special IC (calling ASIC in the following text) for use, this ASIC generally is integrated into audio coder ﹠ decoder (codec) (calling CODEC in the following text) even voice flash memory (flash memory is hereinafter to be referred as FLASH) in the encapsulation, circuit structure is simplified greatly, and power consumption is less.But this class ASIC does not possess the telegraphone function, and this part function realizes in MCU, if do not need telephony interface, adopts this IC proper.Provide the company of this IC still not many at present, have only several families such as OKI, INVOX.Technology, MACRONIX.The 3rd, the selection general dsp is not wanted MCU, whole softwares is lumped together be placed among the DSP, and all external devices also all hang on the DSP.Provide the company of this general dsp also a lot, as TI, SIEMENS, NATIONAL, ANALOG, MOTOROLA or the like.Speech Signal Compression is less demanding to DSP, selects low-grade 20MIPS fixed DSP to get final product.

Several more typical embodiment are: make SDP as the ISD-T360SB that adopts ISD company, can select for use the MSM7704 of OKI company to make CODEC, select for use the FLASH of SAMSUNG or TOSHIBA to make voice and prompt tone flash memory and literal and character library flash memory, and select for use the MCS-51 series monolithic to make MCU; If adopt the PSB2167 of SIEMENS company to make DSP, can select for use the MC14LC5480 of MOTOROLA company to make CODEC, the FLASH of still optional SAMSUNG and TOSHIBA, but also can select the serial FLASH of ATMEL for use.Because the control program of PSB2167 is built-in, thereby MCU can save, and all peripheral hardwares of MCU completely change and are suspended on the PSB2167; If adopt the MSM9841 of OKI company to make DSP, because its inside has included CODEC, so the audio coder ﹠ decoder (codec) among Fig. 1 can save, but voice and prompt tone flash memory will and be suspended on the MCU with literal and character library flash memory together, and MCU still can select MCS-51 series monolithic or other similar functions single-chip microcomputers for use; If adopt the IVM1764 of INVOX TECHNOLOGY company to make DSP, then because wherein integrated FLASH and CODEC and partial simulation imput output circuit, audio coder ﹠ decoder (codec) among Fig. 1, voice and prompt tone flash memory and master controller all can save, and circuit is changed between greatly.If employing and general dsp or further to semiconductor factory customization ASIC (special IC), circuit structure can be simplified further.More circuit module such as CODEC, two FLASH, lcd driver, computer interface, keystroke interface etc. all can be integrated on the dsp chip, design the product that volume is smaller and more exquisite, cost performance is higher.

The method for making of the storage file of language and characters composite tape and use computer medium is as follows:

1, on multimedia computer with the content of raw tone tape with the whole typing computers of monophony form, save as a wav file, edit this file then and want the place of subordinate sentence, segmentation and last full end of tape all to add suitable mark, edit a text that contains the corresponding literal of voice tape full content again as the original audio files of recording composite band at each.Wherein the foreign language literal enters a new line in each place corresponding with original audio files subordinate sentence position, and adds the Chinese translation of this row correspondence below, also accounts for delegation.Like this, each sentence of original audio files is all corresponding to text delegation foreign language literal and back to back delegation Chinese written language.Though each of voice section not with the corresponding relation of a certain position of text, but since the position of segmentation inevitable also be first position in this section, so only need first fullstop in segment number and this section has been connected just, the text that obtains like this is called the original character file.The making of complex data is that carry out on the basis with original audio files and original character file all in composite tape and the computer medium.

2, with above-mentioned two source documents of custom-designed software processes produce with the language and characters composite tape that will make on two tracks desire the same two-channel simulating signal files of recording signal, form is the same with the two-channel wav file of Windows, is called wave file.Processing procedure is: with the original audio files of software scans, just partly write the no acoustic wave form that equates with mark lengths to segmentation markers whenever, partly write the modulation waveform of this label coding respectively as the segmentation on the word rail, subordinate sentence mark to the word rail of wave file to the track of wave file; Every voice signal in the original audio files is directly copied to the track part of wave file, simultaneously with the foreign language of this correspondence and Chinese written language by taking out in the original character file, be converted to the word rail part that the Wave data of having modulated writes wave file, finish and the position of Chinese written language end also all will add mark at every foreign language literal.Remembered and corresponding this track waveform unclosed time period still at this literal waveform, partly charge to no acoustic wave form at the word rail.So repeatedly sentence by sentence to the last one handle, partly add full end-of-tape marker at the word rail, just form final wave file.The coding of various marks can not be obscured mutually with any literal code, can be represented by the escape sequence of escape character switch.The literal modulation waveform length of each can not be than longer with the speech waveform of sentence, and the bright read rate of general English is 10 foreign language characters of per second, is no more than 20 characters the soonest.And the FSK modulation system that adopts 2400 baud rates on tape, write down the modulation the literal code signal time, per second can write down 200 foreign language characters or 100 Chinese written languages at least.The encoding amount ratio of Chinese translation and English original text on average has only 63%, and the production process of waveform literal is seen Fig. 2.

3, on computers with custom-designed wave file, produce two binary data files that the FLASH storer needs in the sound literary composition machine with software processes formation identical with this sound literary composition machine recording algorithm.This binary file can be directly imported sound literary composition machine into by computer control and is divided and be stored in two FLASH storeies, and in fact this binary file also can store the civilian machine Standard File Format of sound on the computer medium.

4, on computers with custom-designed have put the identical software processes of algorithm with the machine sowing of this sound literary composition and form binary file and carry out the sound literary composition and want emulation, can a certain step begin to revise and emulation again from the front as dissatisfied, till satisfaction.

5, play wave file on computers, output to the recording arrangement that connects, make satisfactory language and characters composite tape.

To the binary file of making, can directly import sound literary composition machine into by the dataphone interface of computer printer interface and sound literary composition machine and use.And the language and characters composite tape to making also will import sound literary composition machine through magnetic tape sound reproducer, and to need suitable control algolithm could be that numerical data deposits two FLASH storeies in the analog signal conversion on the audio tape to sound literary composition machine during this.Specific practice is as follows: two sound channels of recorder audio output socket are linked to each other with literal magnetic head plug signal socket with the voice magnetic head socket of this sound literary composition machine respectively.After sound-track engraving apparatus began to play, sound literary composition machine linked to each other by voice magnetic head signal socket.After sound-track engraving apparatus begins to play, the analog voice signal that sound literary composition machine receives by voice magnetic head signal socket under the control of DSP, be converted to the digital voice data that compressed continuously order deposit voice and prompt tone flash memory in, the simulation literal code signal of the modulation that receives by literal magnetic head signal socket is then directly sent into MCU after detuner and front end circuit demodulation, determine its location mode in literal and character library flash memory according to the coded data that receives.To begin successively from the address of appointment to the literal code data M CU that receives, order deposits literal and character library flash memory in; To all segmentations, subordinate sentence mark and the last end mark that receives, MCU will set up a segment table and a sentence table respectively in literal and character library flash memory, taking this data in two FLASH storeies is that unit sets up strict corresponding relationship with the syntagma.Here " section " actual finger is " text ", also examines the needs that the chamber answering machine is used simultaneously according to adding up and leave enough surpluses in a large number, and the present invention supposes that hop count (or text number) is no more than 256 in 60 minutes tapes of a dish, and a sentence number is no more than 8192.Therefore, segment table has 256 search terms in this sound literary composition machine, and every 2 bytes are deposited first numbering in the sentence table in this section.Segment table is totally 512 bytes, and form is seen Fig. 3.The sentence table has 8192 search terms, every 8 bytes, the start address of speech data in voice and prompt tone flash memory that the highest 3 bytes are these correspondences, 3 bytes are this corresponding lteral data start addresses in literal and character library flash memory afterwards, following again 1 byte is that this wants sentence foreign language alphabetic character number, and last 1 byte is this Chinese written language number of words.The sentence table is the 64K byte altogether, and form is seen Fig. 4.

Below in conjunction with Fig. 5 simply describe this sound literary composition machine when the typing composite tape signal MCU in the way of structure segment table and sentence table.At first establish three pointers respectively for literal and character library flash memory: the current literal of segment pointer, sentence pointer and literal and character library flash memory deposits address pointer (being called for short the literal pointer) in.Segment pointer and sentence pointer are predisposed to the start address of segment table and sentence table respectively during beginning, and the literal pointer is changed to predetermined literal and deposits start address.Whenever receive a segmentation markers then, just the address that writes literal and character library flash memory stage casing pointed with 2 byte unsigned integer formats numbered in current sentence, and segment pointer is added 2.Subordinate sentence mark of every reception, just ask for the current address of voice and prompt tone flash memory to DSP, it is write the address of sentence pointed in literal and the character library flash memory with 3 byte unsigned integer formats, and the sentence pointer added 3, simultaneously the currency of literal pointer is also write the address of the sentence pointed of having revised in literal and the character library flash memory with 3 byte unsigned integer formats, again the sentence pointer is added 3 afterwards.Then will be with the inferior address that deposits literal and character library flash memory Chinese words pointed in of the foreign language literal code that receives one by one, whenever deposit a character code in the literal pointer is added 1, receive behind the foreign language literal end mark foreign language characters number of this reception is write the address of pointed with the byte unsigned integer format, a sentence pointer adds 1 simultaneously.The Chinese written language coding that then continues to receive one by one again deposits the address of literal and character library flash memory Chinese words pointed successively in, whenever deposit an encode Chinese characters for computer in the literal pointer is added 2, receive behind the Chinese written language end mark Chinese number of words of this reception is write the address of a pointed with the byte unsigned integer format, again the sentence pointer is added 1.Like this, sentence search terms of every processing, the sentence pointer increases by 8, just in time is the address of a sentence search terms.Handle next section or next sentence when receiving segmentation markers or subordinate sentence once more by the time again, method is the same.Handle up to whole dish tape and can run into full end-of-tape marker, at this moment will be used as last subordinate sentence mark to this mark handles, deposit corresponding address information in the sentence table, with the whole zero setting of remainder of segment table and sentence table, represent the end of segment table and sentence table simultaneously with this.As previously mentioned, segment table and sentence table all can not use up, and always have complete zero search terms at last, and these zero search terms are the sign of segment table and sentence end of list (EOL) just.Sound literary composition machine will be put in order the entire process flow process of dish language and characters composite tape content typing and see Fig. 5.

The key of understanding this algorithm is to keep the implication of the currency of three pointers all the time firmly in mind, though all be to point to a certain address in literal and the character library flash memory, meaning is different: the currency of segment pointer is the address of the next section search terms that will write in the segment table; The currency of sentence pointer is the address of the current field of the next sentence search terms that will write in the sentence table; The currency of literal pointer is the address of that character in literal and flash memory that will write now in the literal of current sentence.

Two FLASH data structure in memory have been arranged.It is the various characteristics broadcasts of unit that sound literary composition machine just can carry out with sentence, section according to the order that the user imports from keyboard, how to need to prove the position that finds the voice and the lteral data of their correspondences according to fullstop and segment number here.Numbering from 0 to 8191 with reference to Fig. 3, Fig. 4, sentence has had fullstop, it be multiply by 8 add the plot that sentence is shown, and just finds this search terms in the sentence table.The start address of the speech data that can obtain this from preceding 3 bytes of this search terms voice and prompt tone flash memory (because the end address of the sentence start address of next just, complete just end of tape address, last end address is so needn't charge to the end address of sentence in the sentence table); From being right after 3 bytes and can obtaining the start address of this lteral data literal and character library flash memory of this search terms, 1 number of characters that byte is this foreign language literal down again, last 1 number of words that byte is this Chinese written language.Be easy to take out successively this all literal codes by these data, be transformed to character pattern data via character library then and deliver to LCD.The structure of character library adopts 16 * 16 dot matrix word libraries of standard in the PC, and usage is very simple, no longer narration.Be that segment number is arranged earlier during by the section redirect, at this moment segment number multiply by 2 and add the segment table plot, just find the search terms of this section in segment table.The section search terms has only one 2 byte integer, expression the time this section in first number, can in the sentence table, find the start address of the various data of this section thus by the fullstop lookup method.Use this search method, foregoing various be that the broadcast mode of benchmark obviously is not difficult to realize with sentence, section.

Claims (4)

1, a kind of by earphone (1), microphone (2), voice magnetic head Signal plug (3), literal magnetic head signal socket (4), telephone line interface (5), detuner and front end circuit (6), simulation Port Multiplier and analog signal processing circuit (7), audio coder ﹠ decoder (codec) (8), digital signal processor (9), master controller (10), voice and prompt tone flash memory (11), literal and character library flash memory (12), LCD (13), dataphone interface (14), push-button array (15), the solid pronouncing and literal learning machine that battery and charging circuit (16) etc. constitute, it is characterized in that: said earphone (1), microphone (2) and voice magnetic head signal socket (3), telephone line interface (5) is connected with simulation Port Multiplier and analog signal processing circuit (7) respectively, these signals are switched and nurse one's health by the latter and deliver to the back level behind the suitable level and continue to handle, telephone line interface (5) also is the output channel of voice signal simultaneously, it is connected to simulation Port Multiplier and analog signal processing circuit (7) with earphone (1), and the voice signal that the back level is sent here is by simulation Port Multiplier and analog signal processing circuit switching and be amplified to suitable level and power; Simulation Port Multiplier and analog signal processing circuit are connected by bidirectional lines with audio coder ﹠ decoder (codec) (CODEC) (8), the voice signal that back level can be sent here carry out becoming after the AD conversion digit time sequence and be sent to the back level, and also sequence digit time that the back level can be sent here be carried out becoming analog voice signal after the DA conversion and delivered to prime; Digital signal processor (DSP) (9) is connected with master controller (MCU) (10), audio coder ﹠ decoder (codec) (8) and voice and prompt tone flash memory storer (11) with bidirectional lines, make the digital signal processor (9) deposit voice and prompt tone flash memory (11) after digitized voice signal compresses in fetching from prime under the control of master controller (10), also can under the control of master controller, from voice and prompt tone flash memory, take out and be sent to prime after the digital voice data that has compressed decompresses; Literal and character library flash memory (12) and two-way connection of master controller (10), so that by master controller the lteral data that receives is deposited in literal and the character library flash memory, or the literal that will show from literal and character library flash memory takes out; Literal magnetic head signal socket (4) is connected with detuner and front end circuit (6), simultaneously detuner and front end circuit (6) are connected with master controller (10), are sent to master controller (10) after the literal code signal demodulation of literal magnetic head signal socket being read by detuner and front end circuit; LCD (13), dataphone interface (14) and the push-button array (15) in addition that are connected with master controller simultaneously.

2, solid pronouncing and literal learning machine according to claim 1 is characterized in that: by dump on the external memory medium in inner voice and prompt tone flash memory (11) speech data and the literal code data in literal and the character library flash memory (12) keeping with sentence, the section be the strict corresponding relation of unit.

3, solid pronouncing and literal learning machine according to claim 1, it is characterized in that: said telephone line interface (5) is connected with simulation Port Multiplier and analog signal processing circuit (7), thereby when not doing the study use, this learning machine can be used as the circuit answer phone and uses.

4, solid pronouncing and literal learning machine according to claim 1, it is characterized in that: computer is gone in the complex data unloading that records in the internal memory by built-in parallel dataphone interface, also the normal data in the computer can be imported internal memory, and then the exterior storage medium of its use can be all computer software storage mediums such as disk, CD.

Images