Audio response terminal for use with data processing systems

[Detailed Description of the Invention]

[Field of the Invention] The present invention relates to the voice (audio) response terminal for using it with a data processing device. Although such an audio response terminal may also contain a visible display and other data output devices, they are not the indispensable requirements for the present invention. [Description of the Prior Art] Now, it is becoming an everyday occurrence to communicate an input terminal to a Use data processing device and a dialogue. General communication is performed at present via a visible display like the IBM3278 data display terminal which has a keyboard for an operator to input information and a cathode-ray tube (CRT) which displays both of a response of the information and data processing device which were inputted (IBM: registered trademark). However, when having inputted the flow of a continuous text, in an example of application like the manufacturing process as which a message is seldom displayed on a CRT screen, an operator does not watch a CRT screen continuously. For the operator of the blindness which uses a data processing device, or weak eyesight, it is not useful to borrow the help of the CRT screen for an output message display. The device which answers with a sound to an operator's special demand is known. By recording a word or the big dictionary of a syllable on a voice recording medium like magnetic tape, or recording in digital one on a disk, and choosing a suitable word or syllable by a data processor, and connecting them, an old thing is a thing of the form which constitutes an audio response, and is Oh. The peculiar restriction referred to as that can use only the word already recorded for a response and it cannot use a new word and combination in this method is Oh. This method uses the bulk store where cost to the extent that it cannot admit starts again. [Present invention] The present invention is turned to providing the terminal unit for data processing devices which generates the voice response of the information and receipt Ivy message which were inputted from the data processing device. Then of a terminal is also good at a mere audio response unit, or a visible display or other arbitrary data output devices may be combined with an audio response unit. The present invention uses a well-known phoneme (phoneme) type voice synthesis machine for a person skilled in the art. it is a thing suitable for especially realizing by the integrated circuit A tip with one single example -- Then, a fixed resonance filter, and fixed control parameter value -- therefore, the generating pipe model which comprises the resonance filter which has the resonance frequency controlled, and in which two or more tuning is possible is included. The digital changes circuit is included, and by adding repetitively the schedule percentage of the difference between a present value and a desired value to a present value, the value of an utterance control parameter is gradually changed as it changes from a phoneme (phoneme) to a phoneme. An utterance pipe model is carried out using the capacitive switching method that the sound of the human being of comparatively low frequency can be realized without needing many components. The glottal sound source generator is suitable for glottis vibration generating a glottis oscillating signal in digital one in the mode orthopedically operated easily by the form of an arbitrary request in spectrum. A selecting means for the present invention to choose data from Then and a data buffer with the audio response unit used with the data processing device containing the data buffer which memorizes output data, The conversion method for changing selected data into a series of digitized phoneme signals, An audio response unit including the voice synthesis machine for changing the digitized phoneme signal into a voice (audio) output signal and the means for connecting a voice response signal to a speech output unit is provided. [Example] As for the example of the present invention, a visible display is connected to a data processing device via a data buffer. Then of a data processing device is also good at the microprocessor stored in the cabinet good Then and same as the display connected directly or remotely or the processing unit of other arbitrary form with the remote calculation processing system connected to the display via the communication line. Drawing 1 shows the block diagram of such a common system. Visible display 1 is connected to data buffer 2. For a visible display, this buffer is a refreshment buffer and memorizes the character which should be displayed on the screen of visible display 1. Processing unit 3 is connected to data buffer 2. Auxiliary processing unit (control device) 5 gives a control facility to an audio response unit. It is connected to data buffer 2 and control device 5 receives the input from input keyboard 6 and 1 set of mode keys 7. Voice synthesis machine 9 is connected so that an input may be received from control device 5. The audio signal output from voice synthesis machine 9 is connected to speech output unit 10 (a speaker or headphone may be used). Intermediary To have [ as ] which will separate a speaker if the socket of a speaker and headphone is provided in the example and the jack plug of headphone is inserted in a socket. Drawing 2 shows control device 5 of Drawing 1 in more detail. As for it, this control device 5 is connected to data bus 21 and address bus 22 including microprocessor 20. The microcode for controlling operation of control device 5 is memorized by read-only memory (ROS) 23, and random access memory (RAM) 24 for work is also connected to data bus 21 and address bus 22. In the example, it is equipped with these components (microprocessor 20, read-only memory 23, and random access memory 24) on one composition card. Display adapter 25 is connected to data bus 21, address bus 22, and data buffer 2 again. Keypad adapter 26 is connected to keypad 7, and audio adapter 27 is connected to voice synthesis machine 9. Both keypad adapter 26 and voice adapter 27 are connected to data bus 21 and address bus 22. Keypad 7 is sketched by Drawing 3. There, it is two utterance (speaking) format keys 31, 32 or 10 number keys 33 thru/or 42, and a total of 16 keys containing four special function keys 43, 44, 45, and 46. Operation is explained. An utterance mode is made when selection of four There and an utterance mode pushes key 31 or 32. If key 31 is pressed, it will change from "pronunciation" mode to a "spelling reading (1-character utterance of 1 character)" mode, or its contrary. If key 32 is pressed, it will switch between a "punctuation" mode and a "phonographic (Huonetech) spelling" mode. If key 31 is pressed, it will return to "pronunciation" from either a "punctuation" or "phonographic spelling reading." Whenever it pushes any of these two keys they are, a new mode name is announced every. "Pronunciation" mode: The contents displayed on the screen are pronounced in the same way as reading a book (a space and a punctuation are not announced). The space of each word is removed, and if it does not contain the character which each word repeats 3 times or more continuously including a vowel, it will try to pronounce it. If that is not right, spelling reading of the word will be carried out (every character is uttered). Distinction of a capital letter (upper row) and a small letter (lower berth) is not carried out. The number which comprises the number of 2 or more characters is separately recognized for every number, for example, 132 is not pronounced with a one alder dreadlocks and Saaty two (to Facial, or To and San yuu -), but is pronounced with a one three two (to Ichi and San -). If only one character is repeated 3 times or more, it will be first uttered only once to the sound, and the number of times of an appearance will be announced. It is shown instead of this repeating utterance of the character repeatedly that the number of times of an appearance was computed from the text data. "Punctuation" mode: The contents of the screen are although it is uttered like [ in the case of pronunciation modes ], In addition, a space is identified with sound, respectively, a new line is identified by different Poet, if a space is repeated 3 times or more, the number will be given, a top capital letter is identified, and a punctuation mark is pronounced. Distinction of a space and space is not carried out. A top capital letter is identified and spelling reading of the capital letters other than a head is carried out. "Spelling reading" (1-character utterance of 1 character) mode: All the word of 1 character of 1 character is pronounced at this time, a space and a capital letter are identified, and a space or the number of times of a repetition of a character is given, and a punctuation mark is pronounced. "Phonographic spelling reading" (1-character utterance of 1 character) mode: Since the confusion in the case of the character (for example, v and b) of similar sound is canceled, spelling can be chosen using the international phonetic alphabet. It sets it a standard any of the cursor of a display, or an internal pointer to be, and utterance is performed. In the case of an internal pointer, a cursor stops at a state of rest and a user reads from a screen by moving a pointer. At the time of power supply ON, a device is placed on cursor mode and pronunciation mode. Conversion in pointer mode is attained by pressing key 43, pressing two number keys 33 thru/or 42 following it, and specifying the column index (01 thru/or 24) on a display. A pointer is placed for a sequence to begin, a column index is pronounced (the it is xx sequence like), and utterance of the sequence starts. If key 41 is pressed after pressing key 43, it will return to cursor mode. Cursor mode: When it is in cursor mode, screen data is chosen for utterance by choosing the word which positions a cursor and should be uttered by the regular cursor control key on the keyboard of a display by a keypad. The character, word, or sequence specified by the cursor is then uttered. A cursor directs the word, when it is under the character of one of the words. When it is in cursor mode and "pronunciation" mode, the function relevant to each key on a keypad is explained below. When it is in "spelling reading" or a "phonographic spelling reading" mode, all of 1 character of 1 character are pronounced. At the time of the mode of a "punctuation", every character is pronounced like the case of "pronunciation", and the above additional information is added. Utterance can be stopped by pressing "stop" key 41 at the arbitrary times. An output will be started, if a suitable key is pressed so that it may state below. The present function is also closed when other functions are needed. Key 33: The word in front of the word on which the cursor is placed is uttered. If it is in the place whose cursor is a word of the beginning of a sequence, the voice "the first word" can be heard. Key 34: The word on which the cursor is placed is uttered. Key 35: The word after the word on which the cursor is placed is uttered. If it is in the place whose cursor is after the last of a sequence, the voice called "last word" can be heard. Key 36: The sequence in front of the sequence on which the cursor is placed is uttered. If the cursor is placed on the 1st row, the voice called "1st row" will be heard. Key 37: The sequence on which the cursor is placed is uttered. Key 38: The sequence after the sequence on which the cursor is placed is uttered. If the cursor is placed on the last sequence, the voice called "last sequence" will be heard. Key 39: The character in which the cursor was placed is uttered. If it is an alphabetical letter, an equivalent sound will be uttered phonographically (phonetic). Key 40: The position (for example, the [ xx sequence, ] the yy column) of a cursor is announced. Key 41: It is an utterance stop key. Key 42: Regardless of the position of a cursor, utterance starts on top of a screen and continue to the termination of a screen. In cursor mode, it is cautious of there being "no recurrence voice." Keys 44, 45, and 46 can be used at the arbitrary times. Those functions are mentioned below. Pointer mode: If key 43 is pressed, it will change from the cursor mode to pointer mode. Two digits (01 thru/or 24) which specify a column index following this must be pushed. a pointer is placed by the character of the beginning of the selected sequence -- a column index -- "-- the -- xx sequence " is announced and utterance of the sequence starts. A pointer moves in the bottom of a word as each word is uttered. the word currently uttered if utterance is stopped and resumed in this way -- or spelling reading is carried out (every character is uttered), and it is repeated about A word. The key function relevant to each key is as follows in pointer mode. Key 33: A pointer returns to along the sequence of Powered by to the character of the beginning of a front word, and the word is uttered. This key continues and may be pressed repeatedly, and whenever it is pushed, only one pointer returns to along a sequence. If it is in the place whose pointer is a word of the beginning of a sequence, when a key is pressed next, a pointer does not move but is uttered with "the first word." Key 34: The word at which the pointer was positioned is uttered. A pointer does not move. Key 35: A pointer carries out Along movement, and progresses to a sequence to the character of the beginning of the following word, and the word is generated. A key may be pressed continuously repeatedly and uttered with the time "last word" of reaching the last word. key 36: -- a pointer is positioned and made into the character of the beginning of one-row Go back on a screen -- a column index -- "-- the -- xx sequence " is announced and utterance of the sequence starts. A key continues and may be pressed repeatedly, and whenever it is pushed, one row of pointers return on a screen. If a pointer is in the 1st row, when this key is pressed, it will not move, but "the 1st row" will be announced. Key 37: After announcing a column index, the sequence containing a pointer is uttered. A pointer is moved to along a sequence and a pointer is placed under the next word of the uttered word as utterance of a sequence progresses. key 38: -- one row of pointers are moved downward on a screen -- a column index -- "-- the -- xx sequence " is announced and utterance of the sequence starts. A key continues and may be pressed repeatedly, and whenever it is pushed, a pointer falls a screen top by one row. If a pointer is in the last sequence, when this key is pressed, a pointer will not move but it will be announced that it is "the last sequence." Key 39: A pointer position is announced (the [ pointer ] the [ xx sequence, ] the yy column). If this key is pressed during utterance, before a position is announced, utterance will stop. A pointer is uttered or is located in the place of one-character the word currently uttered one character. Key 40: A pointer moves to the cursor and the position is announced (the [ xx sequence, ] the yy column). Key 41: Read-out stops. A pointer is uttered or is placed on the position of 1-character the word currently uttered 1 character. Key 42: It is resumed in the position of a pointer and read-out continues to the termination of a screen. Keys 44, 45, and 46 can be used for arbitrary time. Those functions are shown below. Keys 44, 45, and 46 give the special means which can be called to arbitrary time. When one of those keys is pushed, other arbitrary operations are advancing and Buddy shelves and the operation of those are stopped. Key 44: A user enables it to choose the specific column by which this key should be uttered. The remaining data of each sequence is disregarded. If the number of one column is displayed on the screen in this way, the user can choose so that Okay. utterance of the column uttered without hearing the data of each sequence may be turned down. A user chooses the zone which should be uttered by pushing two digits (01 thru/or 80) which specify the first column, and also pushing two digits (01 thru/or 80) which specify the column of the last which should be uttered, after pressing key 44. The time "zone xxyy" of this selection operation being completed is announced, the check of zone selection is made, a pointer is positioned at the first selected column on sequence 1, and a user progresses to pointer mode. After pressing key 43, it can return to cursor mode by pressing key 41. A sequence is started and completed in the column as which now Then was also chosen in the form of which mode. If a cursor is out of a zone and key 34 is used, the word called "invalidity" will be announced. If it refers to the word outside a zone when key 33 or 35 is used, "invalidity" will be announced also in this case. If are expected of escaping from a band limit and uttering the whole sequence, and nine (key 41) is pushed after pressing key 44, it can return to the mode of all the screen zones of 01 thru/or 80. At the time of power supply ON, zones are 80 columns (namely, full length of a sequence). Key 45: The contents of the sequence will come to be uttered if it desires to determine a state in case a user is displayed on a display directions sequence (sequence 25), and key 45 is pressed. Before the displayed symbol is uttered, it is changed into meaningful words and phrases. Key 46: The position of the emphasis (highlight) field is determined by pressing key 46 and key 33. If key 33 is pressed after pressing key 46, when it is in cursor mode, the at The search which a screen begins will be started. If it is in pointer mode, a search will start in the present pointer position. When the emphasis field is discovered, it is stopped by search, and a pointer is placed on the place of the character of the beginning in the field, and the position (sequence xx to highlight, column yy) is announced. Now, a user is in pointer mode and utterance can begin him by pressing key 42. When key 33 is pressed after pressing key 46, the search to the next emphasis field will be performed. A search is performed about the whole screen. An original position will be repeated if the emphasis field is not discovered further. If there is no emphasis field on a screen, it will be uttered, saying "he has no emphasis field." When key 33 is pressed after pressing key 46 following it, a search is made to start again in the starting point of a screen, if it comes to make a return in cursor mode for the operation in the emphasis field. Key 33 is pressed after pressing key 46, when it is necessary to skip the already uttered field. Data is transmitted between the memory storage on microprocessor 20 and data bus 21, and an adapter device. A control device uses the input-and-output (I/O) device mapped by the memory. By [ whose they are memory places / that ] reading and writing like, this is the method of communicating with peripheral equipment. the address assigned to them receives them without relation to physical memory storage -- Read(ing) -- writing -- it is monitored by the hardware of an adapter. The function of an address reading-and-writing device is expressed as follows using a hexadecimal sign method.

[Table]

Star reading An adapter device interrupts and it drives it. That is, when an adapter device has the necessity of communicating with a processor, an adapter device generates an interruption signal, and a processor always suspends the present work in order to poll an adapter. Interruption of the highest priority is a thing from an audio adapter, and continues with the thing from a keypad adapter, and the thing from a data buffer. If operation is explained, data buffer 2 (Drawing 2) will memorize the character of all the screens of 25 rows. Each sequence may also be included up to eight scans of the whole buffer synchronizing with the electron beam scan of the cathode-ray tube of a display. Display adapter 25 (Drawing 2 and details are Drawings 4) receives five different Ivy inputs from a data buffer. When the scan of a visible display has generated the cursor on a screen, a cursor signal arises on line 50. This signal is detected by cursor detection logic 51, and, as a result, produces a cursor register load signal on Repeat 52. A cursor register has two portions, i.e., cursor column register 53, and cursor sequence register 54. These registers 53 and 54 are loaded from column counter 55 and sequence counter 56. A column and sequence counters 55 and 56 are driven by the input on line 57 from a display buffer. This input is drawn from a display counter and the addressing step to one is carried out for every character position scan by a visible display. Whenever a column counter scans 80 characters, it is reset, and it carries out the addressing step of the sequence counter to one about the scanning line of each sequence then. A sequence counter is reset every 25th sequence. When cursor detection logic 51 shows existence of a cursor, the present value of column counter 55 and sequence counter 56 is loaded in cursor column register 53 and cursor sequence register 54. ing in which both the cursor register always has the present address of the cursor position in this way. Both the cursor register is connected to address bus 22 (Drawing 2) and data bus 21 (Drawing 2). As shown above, when a microprocessor issues a reading command to address 400C, cursor sequence register 54 is read and the data is given to data bus 21. When reading to 400D is taken out, cursor column register 53 is read. Sequence register 58 connected to address bus 22 and data bus 21 always receives the input from a microprocessor, when address 4008 is written in. Comparing device 59 sends a load signal to random access memory 60, when coincidence is obtained in the contents of sequence register 58 as compared with the value in sequence counter 56. Random access memory 60 has two portions, i.e., RAM-A, and RAM-B. The input to RAM60 is given from the display buffer on line 62 via gate 61. The RAM-A portion of memory 60 memorizes the data of a sequence now, and it is loaded when comparing device 59 takes out a load RAM signal on line 63. It is shown that one signal is sent to a microprocessor via line 63 simultaneously with it, and RAM60 is being loaded. In the example, a visible display has a directions sequence at the bottom of a screen, and the message which shows the state of a data processing device is displayed there to a user. The user of a device does not necessarily desire to be interrupted every, whenever one of the messages of these occurs. Then, a device is specified that it gives the audio display of these messages in a specific case. An indicator sequence signal is received from a display via line 64, and is sent to comparing device 65. The present indicator sequence character is memorized by the RAM-B portion of memory 60. Whenever a display scans an indicator sequence, every, the sequence memorized by RAM-B is compared with the present sequence received via gate 61. Change takes out a RAM load signal on line 66, and Buddy shelf Comparator 65 are exchanged with data with new data memorized by RAM-B. An interruption signal is generated on line 66 and it is sent to microprocessor 20 (Drawing 2). A certain thing of the data into which it was put to the visible display, for example, a password, and an identification number are not usually displayed. It is clear that this data's [ both of ] an audio output must not be produced as a result, when such data is scanned in a display buffer there, an undisplayed signal is received on line 67 and it always forbids gate 61. When the input from keypad 7 shows that it wishes for a user making it speak about a specific sequence, A microprocessor judges to which sequence the cursor is pointing now, and a processor loads the column index needed in sequence register 58 depending on whether it is a sequence which is the specific sequence which a user needs, or it precedes, or it is a sequence which follows. When comparing device 59 shows that a specific sequence is being scanned with a visible display, the sequence of a character is loaded to RAM-A via gate 61 from the data on line 62. When the sequence is loaded into a register, a microprocessor issues a series of 80 reading data row buffer commands (4008), and loads the sequence of a character into RAM24 (Drawing 2). Drawing 5 shows in diagram two main tasks performed by processor 20. These tasks are keypad processing task 70 and conversion task 71. Answering a specific combination of the input of a keypad, a keypad processing task obtains the sequence of a desired data character from data buffer 2 via display adapter 25, and loads it into text buffer 72 contained in RAM24. If the specific word in a sequence should be uttered, a keypad processing task will separate the word from start pointer 73 similarly memorized by RAM24 and termination pointer 74. A key processing task contains a series of subroutines respectively relevant to the key aggressiveness command listed before. When a user indicates by will that cursor mode is required in this way and key 34 (Drawing 3) is pressed, A keypad processing task calls the routine which judges the position of the cursor on a screen by reading cursor column register 53 and cursor sequence register 54 (Drawing 4), The sequence of the data from an appropriate post-display adapter is required, waiting and a character are loaded to text buffer 72 until use of it is attained, the word relevant to the cursor column begins an appropriate post-start pointer and a termination pointer, and it puts on the last point. If key 35 is pressed, a keypad processing task will call the routine for appointing the boundary of the next word in a sequence. When the word by which a keypad processing task should be uttered is determined, conversion task 71 starts the main operation. That is, each word is changed into a series of digitized phonemes, and it sends to appropriate back voice adapter 27 and voice synthesis machine 9. The flow chart of conversion task 71 is shown in Drawing 6. The task is processed by the table (table) type. That is, when a word or a character chain is identified, he follows a task to a suitable table (table) in order to find out the phoneme chain needed. The 1st table is a dictionary of the plain language in accordance with a general pronunciation rule. COUGH, BOUGH, PLOUGH, and ROUGH are memorized in the form of a suitable phoneme chain in this way. Then and it are spelled on the table from which the 2nd table changes the alphabet and a number into a phoneme directly (every character is uttered), and it is needed for mode operation. The 3rd table generates the phonographic equivalent sound of the alphabet. Namely, A= alpha, B= Bravo, etc. are generated in phonetic spelling reading mode. The 4th table is a group of the rule for changing into a phoneme chain in order to utter automatically the word which is not in a dictionary. With a general-purpose context converter, this table is Then and that rule is the 23rd volume of IBM Technical Disclosure Bulletin, The example shown in the paper "general-purpose context converter" of No. 1, P.W.Johnson published in June, 1980 at the 38 - 40th pages of issue, and C.J.Lovell both is followed. The flow of a conversion task separates first the word which should be changed into the next (Step 80), and it is judged whether next in Step 81, the present mode is "spelling reading or phonographic spelling reading." If it is yes, spelling reading or a phonographic spelling reading processor table will be used in Step 82, and the phoneme to the word will be generated. If the result of Step 81 is no, the task of Step 83 will judge about whether the word is in a dictionary table. If this is yes, the phoneme to the word will come to hand from a word dictionary table in Step 84. If the result of Step 83 is no, a task will judge in Step 85 whether the user pointed to punctuation mode. It guarantees that a flag will be Set(ed) in Step 86 for a context converter if it is yes, and an audio response is generated to a punctuation mark. Step 87 continues after Steps 85 and 86. This step is used for generating the phoneme to the word currently processed by controlling a general-purpose context converter. When a phoneme is generated depending on Steps 82 and 84 or any of 87 they are, it is put in in Step 88 to a phoneme buffer (75, Drawing 5). It judges whether Step 88 already has a word which should be processed, if it is, it will be returned to Step 80, and if there is nothing, a task will await that an input is further given from a keypad processing task. Whenever a phoneme is addressing[ to one ]-taken out from a phoneme buffer with an interruption processing machine under the demand from a voice adapter and a phoneme passes through a synthetic circuit, it interrupts by requiring a new phoneme every. The process of phoneme generating and phoneme utterance happens simultaneously in this way. A buffer permits a slight change of generating speed. A microprocessor maintains the pointer for a phoneme buffer "ON", and "appearance" pointer. If the extraction speed by a voice adapter exceeds the speed of supply by a conversion task in this way, a protection mechanism will detect an "appearance" pointer = "ON" pointer, and a special invalid phoneme will be sent to a voice adapter. A voice adapter stops the demand of a phoneme. If newly supplied by the conversion task, the voice adapter will operate and it will start extraction of the phoneme from "appearance" pointer again. [The advantage of an invention] Especially the example of the above-mentioned audio response unit was designed a blind person or for weak eyesight persons. Use of an audio response unit has an advantage in other fields not only to such an operator. For example, it is necessary to give a voice communication device to the worker who cannot watch a screen continuously in the production line for which the data processing device is used in order to supervise a manufacturing process. The main technical advantages of the present invention are being able to change the output signal of a data processing device into an audio signal. The 2nd technical advantage is having prepared the keypad in order to make an operator choose the data and the mode of utterance which should be uttered.

[Brief Description of the Drawings]

It is a block diagram of the control device with which Drawing 1 was shown in the block diagram of the example of the present invention, and Drawing 2 was shown in Drawing 1, The figure and Drawing 6 in which the schematic diagram of the keypad device with which Drawing 3 was shown in Drawing 1, the block diagram of the display adapter by which Drawing 4 was shown in Drawing 2, and Drawing 5 show main operations of the microprocessor of Drawing 2 are a flow chart of the conversion task of Drawing 5. 1 ...... A visible display, 2 ...... A data buffer, 3 ...... A processing unit, 5 ...... A control device, 6 ...... A keyboard, 7 ...... A keypad, 9 ...... A voice synthesis machine, 10 ...... A speech output unit (speaker), 20 ...... A microprocessor and 23 ...... read-only memory (ROS) and 24 ...... random access memory (RAM) and 25 ...... a display adapter and 26 ...... a keypad adapter and 27 ...... a voice adapter.