As shown in Figure 1, disposal system 100 of the present invention mainly comprises: a speech input device 110, a sound encoding device 120, a voice identification apparatus 130, a memory device 140, a speech simulation control device 150, a finger-impu system 160, a display device 170, a switchgear 180 and a system host board 190 etc.
System host board 190 mainly comprises: bios memory body device 194, a display device controller 195 and a power connector 196 of a system processor 191, a System on chip controller 192, a keyboard controller 193, one basic input and output system (Basic Input andOutput System), when external one extraneous power supply (not shown), in order to the operating voltage of each correlation unit in the system to be provided.
Set forth for convenience, the system processor 191 in the system host board 190 that the present invention exemplified is directly via an address bus (Address Bus; AB), a data bus (Data Bus; DB) and a control bus (Control Bus; CB), be connected with System on chip controller 192; As for keyboard controller 193, bios memory body device 194 and display device controller 195 etc., can be designed to directly or partly control signal be arranged according to the difference that designs is the control that is subjected to System on chip controller 192.Certainly, more existing on the market motherboard devices, system processor wherein via an impact damper and other controller after, be connected to the controller product sheet of each stratum again; And, the system processor in the motherboard, the also non-microprocessor that uses Ying Daier (Intel) X86 that company provides that only is defined as.
In system host board 190, keyboard controller 193 wherein is provided with the clock signal of a keyboard data input end (KB-DATA) and a keyboard clock pulse input end (KB-CLK), to receive the data of being keyed in from the finger-impu system 160 of ibm personal computer or its compatibility computer.During enforcement, keyboard controller 193 for example can adopt industry to be familiar with being numbered 8742 keyboard controller.193 technology that transmit data of finger-impu system 160 and keyboard controller for example can be with reference to United States Patent (USP) the 4th, 460, and No. 957 and the 5th, 146, No. 567.
Below further specify keyboard data and deliver to the situation of host computer system.As shown in Figure 1, speech simulation control device 150 is the control end (G) that has a control signal output ends (CONTROL) and be connected to switchgear 180, uses the guiding path of gauge tap device 180.In other words, open beginning setting state (hereinafter further specifying) after the start, the keyboard data of finger-impu system 160 (Keyboarddata) is connected to the keyboard data input end (KB-DATA) of keyboard controller 193 via the first keyboard data line 161A, switchgear 180, the second keyboard data line 161B; And the keyboard clock pulse (Keyboardclock) of finger-impu system 160 is via the first keyboard clock pulse line 163A, switchgear 180, the second keyboard clock pulse line 163B and be connected to the keyboard clock pulse input end (KB-CLK) of keyboard controller 193.By this, make the data that finger-impu system 160 is keyed in, to deliver to keyboard controller 193, optionally so the system processor of delivering in the system host board 190 191 handle.During enforcement, above-mentioned switchgear 180 can adopt the general integrated switching circuit of PMOS of numbering QS3244.
Deliver to the practice of host computer system as for the speech simulation keyboard data, as shown in Figure 1, function of the present invention mainly is to utilize above-mentioned speech input device 110, and sound encoding device 120, voice identification apparatus 130, phonetic memory body device 140 and speech simulation control device 150 are finished.
Speech input device 110, for example can directly adopt a microphone input media, in order to voice signal (voicesignal) with the user sent, be converted into analog voice data (analog voice data), and by a speech data incoming line 111, and it is delivered to sound encoding device 120.
The main function of sound encoding device 120 is in order to will being encoded from the voice signal of voice input media 110, and via a data line 113, with export one through the digital voice data (digital voice data) of coding to voice identification apparatus 130.During enforcement, coding and decoding device (CODEC) that sound encoding device 120 can adopt company of Texas Instrument (TI) to provide to be encoded to TCM320C36.
Phonetic memory body device 140 only uses a read-only memory (ROM) 140A usually, but in order to promote the ability of system handles voice, also can use a SRAM (SRAM) 140B in addition.The speech samples data (voice template data) of read-only memory (ROM) 140A in order to store common people's specific format in advance, it can adopt INTEL, SGS or MXIC company that the read-only memory device that is numbered 27C010 is provided; As for the speech samples data (voice template data) of SRAM (SRAM) 104B in order to the own specific format of storage user, it can adopt the memory device of CYPRESS CY7C109 that company provides, or the memory device of employing MI CRON MT5C128KA1 that company provides etc.
Voice identification apparatus 130 is connected in memory device 140 and sound encoding device 120, and as shown in Figure 1, it is provided with: one first group of signal wire 131, one second group of signal wire 133, one the 3rd group of signal wire 135 and one the 4th group of signal wire 137.
First group of signal wire 131 is connected with read-only memory (ROM) 140A, voice identification apparatus 130 via first group of signal wire 131 to read the data of read-only memory (ROM) 140A.Wherein: the ROMCS of voice identification apparatus 130
++, expression is selected signal output part in order to the wafer of selecting read-only memory (ROM) 140A; The MEWR of voice identification apparatus 130
++, expression is read signal output part in order to the memory body that reads read-only memory (ROM) 140A or SRAM (SRAM) 140B; As for the A (0:16) of voice identification apparatus 130, then represent memory body access address signal output part in order to access read-only memory (ROM) 140A or SRAM (SRAM) 140B.
Second group of signal wire 133 is connected with SRAM (SRAM) 140B, voice identification apparatus 130 via second group of signal wire 133 with access SRAM (SRAMA) 104B.Wherein: the SRAMCS of voice identification apparatus 130
++, expression is selected signal output part in order to the wafer of selecting SRAM (SRAM) 140B; The MEWR of voice identification apparatus 130
++, expression is in order to write the memory body write signal output terminal of SRAM (SRAM) 140B; As for the MD (0:15) of voice identification apparatus 130, then export/go into end in order to the data-signal of expression connection read-only memory (ROM) 140A or SRAM (SRAM) 140B.
The 3rd group of signal wire 135 is connected with sound encoding device 120, and voice identification apparatus 130 utilizes the 3rd group of signal wire 135, to receive the digital voice data from sound encoding device 120.Wherein: the DX0 of voice identification apparatus 130, expression voice identification apparatus 130 is delivered to the data output end of sound encoding device 120; The DR0 of voice identification apparatus 130, expression voice identification apparatus 130 receives the data input pin of sound encoding device 120 forward data; The SCLK of voice identification apparatus 130, the tandem clock pulse output terminal of expression voice identification apparatus 130 is with the use that cooperates 130 data of sound encoding device 120 and voice identification apparatus to transmit; As for the FSYNC of voice identification apparatus 130, the synchronous signal output end of expression voice identification apparatus 130 is recognized the synchronous use that 30 data transmit so that sound encoding device 120 and speech recognition to be provided.
The 4th group of signal wire 137 is connected with speech simulation control device 150, and voice identification apparatus 130 and speech simulation control device 150 utilize the 4th group of signal wire 137 to use with the transmission as data to each other.Wherein: the RST of speech simulation control device 150, for reset signal (Rest signal) output terminal, in order to provide reset signal to voice identification apparatus 130; The WR of speech simulation control device 150
++, for writing out the output terminal of signal, expression speech simulation control device 150 is planned data and is written to voice identification apparatus 130; The RD of speech simulation control device 150
++, for reading the output terminal of signal, 150 simulations of expression speech simulation control device are from voice identification apparatus 130 reading of data; The ACK of speech simulation control device 150
++, being the output terminal of confirmation signal, expression speech simulation control device 150 has been ready for the data that can accept voice identification apparatus 130; The LAST of speech simulation control device 150 for notice is finished the output terminal of receiving signal, has received data in order to cooperate expression speech simulation control device 150; As for the D (0:7) of speech simulation control device 150, for end is gone in the output of data-signal, for the use of voice identification apparatus 130 and 150 data transmission of speech simulation control device.
In addition, another vital task of voice identification apparatus 130 is: in order to the digital voice data (digital voice data) and the degree of approximation that be stored in the speech samples data (voice template data) of memory device 140 of comparison from sound encoding device 120; And as a digital voice data and a certain speech samples data is when reaching the degree of approximation of a preassigned, must be via the certain line 139 in one the 4th group of data line 137, to export a speech samples index data with respect to the speech samples data (voice template index data) to speech simulation control device 150.During enforcement, voice identification apparatus 130 for example can adopt DSP Communication, the phonetic order processor that inc provided (Voice Command Processor) (DVC 306).
Speech simulation control device 150 refers as preamble, and it is connected in voice identification apparatus 130 and keyboard controller 193.Speech simulation control device 150 will be mainly in order to will being changed from the speech samples designation data of voice device for identifying 130, and make it be modeled into the data of keying in for as keyboard.Now, speech simulation control device 150 makes keyboard clock pulse end (CLK) output one keyboard clock signal, via a gate circuit (open collector) (154) and be sent to the keyboard clock pulse input end (KB-CLK) of keyboard controller 193 via the second keyboard clock pulse line 163B; By this, cooperate the keyboard clock signal, make the data that 150 simulations of speech simulation control device are keyed in as keyboard, via another gate circuit (open collector) 152 and be sent to the keyboard input end (KB-DATA) of keyboard controller 193 via the second keyboard data line 161B.
The situation of keyboard controller 193 takes place data to be delivered to simultaneously for fear of speech simulation control device 150 and finger-impu system 160.The treatment principle that solves this situation is that when speech simulation control device 150 was realized finger-impu system 160 in the transmission data, speech simulation control device 150 waited keyboard input media 160 transmission data automatically and finishes the transmission of data of beginning to carry out.
In other words, speech simulation control device 150 is that the signal condition that utilize to detect the second keyboard data line 161B and the second keyboard clock pulse line 163B is or changes in special time.If change, expression finger-impu system 160 is transmitting data, and 150 transmission that suspend data earlier of speech simulation control device are to allow finger-impu system 160 proceed the transmission of data; Otherwise, if no change, expression finger-impu system 160 is not transmitting data, speech simulation control device 150 is before carrying out the data transmission, the control signal of its control signal output ends of Schilling (CONTROL) output one logic high potential, serves as the state that keeps logic low potential through a reverser 156 with the first keyboard clock pulse line 163A that forces finger-impu system 160, can not disturb speech simulation control device 150 to carry out the transmission of data to guarantee finger-impu system 160.Certainly, after in case speech simulation control device 150 is whenever finished once or the data of pre-determined number transmit, promptly automatically terminate the logic high potential output of its control signal output ends (CONTROL), transmit the right of priority of data to keyboard controller 193 to allow system recovery have by finger-impu system 160.
Certainly, in order to allow the user when carrying out phonetic entry, have an opportunity to hear or to see the own voice that sent, system of the present invention can connect a loudspeaker 112 and the display device controller in system host board 190 195 and connect a terminal display 170 at sound encoding device 120 places.
As for resistor 158A, 158B, 158C, the 158D among Fig. 1, be to design for the operation that cooperates gate circuit 152,154; And relevant resistor 142A, 142B, 142C, 142D, 142E, 142F then are to design in order to ensure effective promotion of each respective signal.
As shown in Figure 2, the present invention is the cooperation that utilizes software program, so that system initialization, and judge and make system to be that a kind of operator scheme that enters in the speech recognition operation pattern automatically, for example for entering " voice keyboard " operator scheme, " form instruction " operator scheme or " voice training " operator scheme etc.
" voice keyboard (voice keyboard) " operator scheme: default of the present invention is become, can be when receiving the user speech input device 110 sent the alphabetical voice of for example button " A ", " B " or " C " etc., it comes to the same thing on finger-impu system 160, directly keys in the operator scheme of letters such as " A ", " B " or " C ".
" form instruction (windows commands) " operator scheme: system of the present invention is set to, can be when receiving the user speech input device 110 sent the form instruction of windowing software for example, it comes to the same thing in finger-impu system 160, directly keys in the operator scheme of the instruction of windowing software.
" voice training (voice training) operator scheme: default of the present invention is become; the user deposits the voice of oneself in SRAM (SRAM) earlier (140B) intentionally; as a kind of speech samples data (voicetemplate data) of making form by oneself; in the future speech input device 110 is sent voice;, and then carry out the specified operator scheme of these speech samples data for the reference of system as speech recognition as own.
Below further set forth characteristics of the present invention with the operation of program flow diagram coupled system again.As shown in Figure 2:
Step 50 is an initial step, when the external extraneous power supply of the power connector in the system host board 190 196, in order to for example to provide speech simulation control device 150 used+5 couchers to make voltage, so that speech simulation control device 150 is for preparing to enter the state of initial operation.
Step 51 is the step of beginning initialization system, for example sets the initial value that opens of speech simulation control device 150 external pins, and for example determines relevant peripheral hardware, to set voice identification apparatus 130 Inside Parameter Values.
Step 52 is instructed to voice identification apparatus 130 for speech simulation control device 150 send, so that voice identification apparatus 130 can be in order to answer the step of speech recognition operation pattern.
Step 53 has or not the speech data of speech recognition operation pattern to enter the step of voice identification apparatus 130 for repeating in inspection.If do not have, then continue to carry out to check; If have, then execution in step 54.
Step 54 is for judging whether the speech data that is heard is " voice keyboard " operator scheme.If then execution in step 57, so that voice identifying device 130 and speech simulation control device 150 enter the processing of " voice keyboard " operator scheme; If not, then enter and execution in step 55.
Step 55 is for judging whether the speech data that is heard is " form instruction " operator scheme.If then execution in step 58, so that voice identification apparatus 130 and voice mould control device 150 enter the processing of " form instruction " operator scheme; If not, then enter and execution in step 56.
Step 56 is for judging whether the speech data that is heard is " voice training " operator scheme.If then execution in step 59, so that voice identification apparatus 130 and speech simulation control device 150 enter the processing of " voice training " operator scheme; If not, then return back to step 52.
The present invention has following effect:
Because speech simulation keyboard of the present invention is a computer peripheral system independently, thereby its voice signal input do not have direct connection with application program, owing to adopt the QWERTY keyboard controller to be input interface, and simple in structure, and can improve the entire system work efficiency.
Above embodiment is in order to set forth the concrete practice of speech simulation keyboard data of the present invention.Certainly, under the situation that does not break away from category of the present invention, can do some the modification of peer to peer technology, some circuit component among the figure for example, but the circuit that functions of use is suitable is got and is treated, and some circuit can be given integrated circuit and form the suitable integrated circuit of function, for example whole sound encoding device 120 is a new circuit etc. with making voice identification apparatus 130; On the contrary, the function of a certain particular electrical circuit, for example utilize speech simulation control device 150, because its function with speech simulation and control, for this reason, nature can also select one to have roughly the same speech simulation function and and have roughly the same that the device of control function replaces.In a word, all modifications scheme for the present invention's design all should belong to protection domain of the present invention.