GB2196463A

GB2196463A - Portable recorder with voice recognition function

Info

Publication number: GB2196463A
Application number: GB08720582A
Authority: GB
Inventors: Kazuo Hashimoto
Original assignee: Individual
Current assignee: Individual
Priority date: 1986-09-02
Filing date: 1987-09-02
Publication date: 1988-04-27
Anticipated expiration: 2007-09-02
Also published as: JPS6363191A; GB2196463B; JP2639387B2; GB8720582D0

Abstract

A recorder is provided with a voice recognition function for operation as a memorandum storage for messages with search function for message identifiers. Each message stored on a recording medium includes a specific identifying word recorded at a predetermined portion thereof. The identifier may be a name, a number, or other symbol. A RAM stores codes for the identifiers and start end stop location codes for the messages associated therewith. The voice recognition section of the recorder recognizes an operator's utterance of a message identifier and, upon determining that a code corresponding to the identifier is stored in RAM, transports the recording medium to the proper location for playback of the specific message, as represented by the stored code. The specific message is played until reaching the stored stop location. When several messages are stored with a common identifier, a secondary identifier, such as a number or a first name may be stored.

Description

SPECIFICATION Portable recorder with voice recognition function TECHNICAL FIELD This invention relates generally to a portable recorder having voice recognition function.

BACKGROUND ART A tape recorder records conversations of many people, or records messages from many people as ordered memoranda on the tape. A major problem encountered in such a device is that much time is required until a specific message can be accessed because the tape must be rewound or advanced. Therefore, a convenient portable recorder has no utility as a storage of memoranda.

DISCLOSURE OF THE INVENTION It is an object of the present invention to provide a portable recorder having a voice recognition fnction for playing back only a specified message. Namely, for each of a plurality of messages from (or to) several people and sequentially recorded on a recording medium (such as tape, for example), there is recorded after each message on the tape the name of a specific person or a simple word for memorization in a voice recognition circuit. A specific message is played back only upon utterance by the operator to the tape recorder of the recorded name or simple word associated with the message, upon rewinding or forwarding of the tape.

Namely, utterance of a recorded name or simple word causes the tape recorder to play back only the specific message corresponding to the specific utterance.

It is another object of the present invention to provide a portable recorder having a voice recognition function, which recorder plays back only a specific message when a plurality of the same names exists in association with messages recorded on the tape.

Namely, an ordering number is recorded on the tape after the recorded name, and the ordering number is memorized in the voice recognition circuit. Then, when the operator plays back the specific message, the operator utters the name and the number corresponding to the name. For example, in association with the name SATO, the operator utters "SATO-1", "SATO-2" or "SATO-3". Thus, the tape recorder plays back the message in accordance with the operator's utterance.

Namely, a distinct word is used to discriminate among the plural same names (such as SATO), the word (for example, Sato TAROU, Sato JIROU) being memorized in the voice recognition circuit. Thus, the tape recorder is provided with the capability to discriminate and play back only the specific message from among the plural recorded messages.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 is a block diagram of the preferred embodiment of the invention; Figure 2 is an explanatory view of an example shown on a liquid crystal display in the invention; Figure 3 is an explanatory view of a symbol shown on the display corresponding to a word which is recognized and of an output code output upon voice recognition corresponding to the symbol; Figure 4 is an explanatory view of a RAM showing a stopping position and a starting position of the tape, corresponding to the code of the name which is registered while recording; Figure 5 is a flow chart for a recording operation; and Figure 6 is a flow chart for a play-back operation.

BEST MODE FOR PRACTICING THE INVEN TION In Fig. 1, 1 is a micro processor CPU, Ii and 12 are input ports, Oi, 02, 03, 04 and 05 are output ports, I/O is an input-output (double way) port, INT is an interrupt terminal, TEST 1 is a count input terminal which counts transit of the tape (namely, CPU 1 has a counter), and 2 is a keyboard section having keys 0-9 which are used for recognizing the number of voice recognitions, for operating voice recognition of the invention, for registering voice recognition, for rewinding or forwarding the tape, for playback of the message, for stopping the tape, and for recording the message.

The key which is used in the typical operations of rewinding, forwarding, playback or recording the tape is set on the surface of the recorder to be operated without difficulty. The keys 0-9 which are less frequently used for registering voice recognition are set on the reverse side of the tape recorder.

In Fig. 1, 3 is a tape sensor section, namely a shutter which, corresponding to transit of the tape, blocks light between a light emitting diode LED and a phototransistor PT. Thus, blocking the light signifies transit of the tape and the sensor section provides the tape count. Reference numeral 4 is a liquid crystal display which displays the operation mode and transit of the tape, etc., as shown in Fig. 2.

Reference numeral 5 is a microphone, while 6 is a voice recognition section which is made of VLSI (very large scale integration).

Shown in Fig. 1 are RAM 1, a voice recognition memory which registers voice recogni tion patterns, and RAM 2, a memory for memorizing transit of the tape and data which is voice-recognized in a specific order. RAM 1 and RAM 2 are backed backed up by a battery so that, even if power supply is cut off for the tape recorder section, the tape data does not disappear.

Referring once more to Fig. 1, 7 is a driver and 8 is a solenoid, for rewinding and forwarding the tape, etc. Reference numeral 9 is a motor control section which controls a tape operating motor to be ON and OFF, as well as a turning direction thereof.

The recording and playback section for the tape messages is well known. Accordingly, description of this section is omitted from the specification.

Concrete operation of the system is explained upon reference to Figs. 2-6. First, however, the voice recognition patterns must be registered in RAM 1 before using the tape recorder. In order to register the voice recognition patterns, the operator activates the voice recognition register key of keyboard section 2. At the same time, a symbol "REGST", which means registration of the voice recognition pattern, is displayed on display section 4 by the program shown in the drawings, as illustrated at Fig. 2A. As the operator activates the recording key "REC", display section 4 displays a symbol "REC".

The tape recorder does not operate at this time, even if the operator activates the operation key for the tape recorder while registering. In this state, the voice recognition section 6 has already entered the registering mode through output port 02 in accordance with the program executed by the CPU 1.

The operator now utters "record" to the microphone 5. If the uttering voice is correctly registered in voice recognition section 6, "OK" is displayed on the display, as shown at Fig. 2A. An operator can register the voice five times while the symbol "REC" is displayed, as shown in Fig. 3. When the tape recorder is used by only one person, the above described indication "OK" disappears after passage of one second. After the disappearance of the "OK" display, the operator utters "OK" five times and the pattern of these utterances is registered in the voice recognition memory RAM 1.

If a drop of recognition percentage is permitted at the voice recognition time, it is possible to register the voices of five operators, the maximum permitted in the presently preferred embodiment. After the registration (up to five times) is finished, the operator activates the playback key "PLAY" and display section 4 displays the indication "PLAY".

In playback, the operator also registers a voice five times. As relates to the indications "REW" and "FF" (forwarding) and "STOP" (stopping), which are not shown in Fig. 3, the operator also registers the voice five times and "1" is displayed upon finishing registration of the voice pattern in the tape recorder, as shown at Fig. 2D. This corresponds to "1" in the "displaying symbol" of Fig. 3, and this represents the registered number of the name of the operator using the tape recorder.

Thus, the operator may utter "HASHI MOTO" as shown in Fig. 3 and the pattern of the utterance is registered in RAM 1. "OK" is displayed on the display section and after passage of the one second interval, the indication "OK" disappears and the operator registers HASHIMOTO five times.

When an operator uses a single tape recorder, each operator respectively registers his (her) name once, upon a reduction in the recognition percentage. In that regard, the memory size may be increased and each operator's name may be registered a plurality of times, together with a number after the name or, if necessary, the full name may be registered.

Still further, if voice recognition technology improves, each user's name may be registered by a representative of each use.

Thus, the operator may operate the recorder by voice, to obtain voice actuated operation, and may designate the contents of a recording corresponding to each operator by finishing registration of the operator's name and the order for recorder operation.

To perform the above, the operator pushes the key in the keyboard section 2 to operate in a voice recognition mode. The operator may further rewind the tape to the start point thereof by uttering "rewind" as necessary.

Specifically, the operator utters "rewind" and this utterance is recognized by the voice recognition section 6. "Output code when recognized" is then provided at 11-15 of Fig.

3. The corresponding code is output from voice recognition section 6, and is detected by the program in CPU 1 through input port 12. As a result, the tape is rewound.

When the tape is rewound to its starting point, the tape is stopped by a well-known function. At this time, the operator utters "RECORD" into the device. Upon recognition of the uttered input command by the device, operation shifts to step 501 in the flow chart shown in Fig. 5.

If the operator's name is "HASHIMOTO".

the operator utters "HASHIMOTO". Once this utterance is recognized, operation shifts to step 502 and the output code of the name "HASHIMOTO" is provided, particularly corresponding to one of codes "101"-"105" of Fig. 3, such as "102" for example. This code is memorized in the address 0 in RAM 2, as shown in Fig. 4.

In step 503, the recording state is started for the tape recorder and, at step 504, the starting position of the tape, for example "0", is memorized at address 1 in RAM 2 as shown in Fig. 4. When recording is finished, the operator utters "stop". This utterance is recognized at step 505 of Fig. 5 and, as a result, the recording mode is stopped at step 506 and, at step 507, the tape stopping position, for example "55", is memorized at address 2 in the RAM 2 as seen in Fig. 4.

Addresses 3-5 shown in Fig. 4 illustrate that output code 106, corresponding to another uttered name, KIMURA, has been stored at address 3 in RAM 2, that the starting position of the tape for this recording of KIMURA is 55, as stored at address 4 of RAM 2 shown in Fig. 4, and that the tape stopping position at the end of this recording is 150, as stored at address 5 in RAM 2, shown in Fig. 4.

In the above examples, the specific contents recorded on the tape and the position of the recorded contents on the tape are determined by three addresses, specifically three bytes which are set as a single group.

When the contents recorded in accordance with the foregoing process are played back, upon selection of a name, utterance of "play" by the operator in the voice recognition mode, and recognition of the utterance by voice recognition section 6, the program shifts to step 601 in Fig. 6. The operator utters his or her name, such as "Hashimoto" for example.

When this utterance is recognized operation shifts to step 602.

In step 602 CPU 1 refers to RAM 2, and specifically to address 0 therein as shown in Fig. 4. The code corresponding to the above name is referenced and the tape start position corresponding to the code is read.

In this example, the starting position corresponding to the name, as read from address 1, is 0. At step 603 it is determined if the tape start position is the same as the current tape position. If the present tape position is different from the accessed tape start position, operation shifts from step 603 to step 604 in Fig. 6. Therein, the tape position is adjusted by rewinding (REW) or forwarding (FF) the tape in accordance with the difference between the start and current tape positions.

For the illustrative example, the tape must be played back from the tape beginning point (tape start position is 0). If the tape is not at the tape beginning point, the tape is rewound to the beginning point and, at step 606, the tape is played back. In step 605, before beginning of a playback, the operator's name or, in the practical example under consideration, the output code "102" associated with the recorded contents, is displayed when recognized, as shown at Fig. 2F. Consequently, the structures of keyboard 2 and of display section 4, which is made of liquid crystal, become complex and expensive.

However, it is possible to display and input alphabetic characters when the above noted code "102" is being input. Thus, it is pos sible to display "HASHIMOTO" on the display. In that event, the "0" shown in Fig. 2F represents the tape position, i.e., indicates the value of the tape count. Accordingly, for the example under consideration, "0" is displayed because the tape is played back from the tape beginning point of zero.

In the next stop 607 it is determined whether the specific recorded contents are or are not finished. Namely, the tape stopping position "55", stored in address 2 of RAM 2 in Fig. 4, is compared with the present position (i.e., the value of the tape count). As the tape rewinds to the tape stopping position and operation shifts to step 608, the contents of RAM 2 are referenced. Specifically, it is tested whether other occurrences of the same name, or an abbreviated name, are memorized in RAM 2.

If the same name is not discovered in RAM 2, operation shifts to step 609 wherein transit of the tape is stopped. If the name is discovered in RAM 2, operation returns to step 602. Therein the present tape position is compared with the tape starting position of the recorded contents which should be played back, it is decided whether or not the tape is to be forwarded, or whether or not the tape should continue to play back.

Herein, the above operation is repeated and, when the playback of all specific recorded contents is finished, the tape stops completely. When tape transit stops completely it is possible to output the specific voice from a speaker (not shown).

Thus, an operator pushes the key for operation of the voice recognition feature, as a result operation of the voice recognition feature is released, and the operator can handoperate the tape recorder by pushing the key.

The portable arrangement hereinabove described as having a recorder with a voice recognition function thus advantageously permits such a recorder to be used for recording as memoranda various messages from a plurality of people. When a message is to be played back, the operator only utters specific words, or a specific name which is memorized when a message is recorded. Position of the message on the tape is referenced by forwarding or rewinding of the tape, and it is possible to play back only a specific message. The invention thus improves the usefulness of a portable recorder as a memorandum device.

Claims

1. A portable recorder with a voice recognition function for a person's name, time, and date, and a brief message, said recorder operable as a memo pad, comprising: storage means for memorizing locations represented by numerical counts provided by a counter which displays a running count of locations along a recording medium, said memorized locations corresponding to a first, start, point and a second, stop, point on said recording medium; registration means for registering a specified word at said start point, whereby a word such as a person 5 name or a signal is specifically registered at said start point as a message identifier; recording means for recording a user's message until a stop point after said registration of said specified word; holding means for recording of said registration when a user turns off a power source;; standby means for returning the recording medium to a beginning position when the user turns on a playback switch; playback means for playing back a message recorded on the recorded medium and for detecting said specified word as spoken by the user, said playback means further operable for searching for the start point of the message corresponding to said specified word on the recording medium to play back the message; and searching means for reproducing a message from the start point of the message, -corre- sponding to next specified word spoken by the user, after stopping at the stop point of a previous message on said recording medium.

2. A portable recorder with voice recognition function according to claim 1, wherein said specified words for activating the voice recognition function consist of brief words previously registered into a voice recognition circuit.

3. A portable recorder with voice recognition function according to claim 1, comprising further means operable after the recorder is turned on and the recording medium is returned to the beginning position for the next usage, said further means operable for searching for said specified word in response to utterance of said word by a user.

4. A portable recorder with voice recognition function according to claim 1, comprising means for distinguishing among a plurality of occurrences of the same specified word by registering a number following said same word whereby said plural occurrences are distinguished by speaking, for example, Mr. B. 1 and Mr. B. 2, to search said words success#ively.

5. A portable recorder with voice recognition function according to claim 1, wherein the recording medium includes a magnetic tape or a disk.

6. A portable recorder with voice recognition function according to claim 2, wherein the previously registered words are registered a plurality of times in said voice recognition circuit by uttering said words into said voice recognition circuit a plurality of times.

7. A portable recorder with voice recognition function according to claim 6, wherein the registered words represent a digit.

8. A portable recorder with voice recognition function according to claim 6, wherein the registered words represent a name.

9. A portable recorder with voice recognition function according to claim 6, wherein the registered words are selected from a group consisting of a digit, a name, and a time indication.

10. A tape recorder for recording a plurality of pieces of information on a tape, and including means to store for each piece of information data relating to the position of the peice of information on the tape and associated data relating to the identity of the information, and means responsive to identity data to move the tape to a position corresponding to the associated position stored in the storage means.

11. A tape recorder or memoranda recording system substantially as described in the description with reference to the drawings.