Data processing: speech signal processing – linguistics – language – Speech signal processing – Recognition
Reexamination Certificate
1999-08-23
2002-01-08
Dorvil, Richemond (Department: 2641)
Data processing: speech signal processing, linguistics, language
Speech signal processing
Recognition
C704S247000, C704S211000, C704S205000, C704S206000, C704S273000, C704S270000
Reexamination Certificate
active
06338036
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of Invention
This invention relates to a method of notifying a speaker of whether a sound of the speaker is input in an appropriate state when the sound spoken by the speaker is recognized. The invention also relates to a sound recognition device that uses this method, and to a recording medium on which is recorded a processing program that identifies an input state of sound to be recognized.
2. Description of Related Art
Recently, sound recognition technology has been widely used in various fields. In particular, it has been recently used for children's toys and household electrical appliances which have become daily necessities.
If sound recognition technology is used for a device used by a variety of non-specific users, rather than a specific user, in order to recognize sounds spoken by the users at high reliability, it is important to guide the users in the use of the device, such as how to properly input sound, in an easy-to-understand manner, and thus provide an easy-to-use device.
For example, as one device which uses sound recognition which is targeted for a variety of users, a so-called sound clock has been recently developed. That is, when a button or the like disposed on the clock is pressed, a sound informs the user of a current time.
This sound clock is convenient because it is possible to find out the current time in the dark. For example, when a user wakes up in the middle of the night, he/she can find out the current time while in the dark. Furthermore, those who are blind can take advantage of this device. In addition, it is also possible to apply this type of sound clock to children's toys.
In this type of sound clock, setting a current time and alarm can be performed by sound, in addition to outputting the time by sound. For example, if the current time is 6:30 a.m., the user speaks the necessary words in a determined order, such as “a.m.”, “6”, and “30” by using a sound clock in a current time setting mode. In addition, on the sound clock side, the sound spoken by the user is recognized, and based upon the recognition result, the time setting process is performed. Setting an alarm time can be performed in the same manner, and the user speaks a desired alarm time in an alarm time setting mode.
While time setting can be performed by this type of operation, the user may have a concern as to whether the sound spoken by himself/herself has been input in an appropriate state (an appropriate state for a recognition process).
In order to solve this problem, there are methods such that a sound can be input while a recognition result for the word is responded per word spoken by a user. For example, in the example of the content spoken by the user described earlier, the user speaks “a.m.” and a response such as “a.m.” is returned from a device as the recognition result. Next, when the user speaks “6”, a response such as “6” is returned from the device. Furthermore, when the user speaks “30”, an operation is performed such that a response such as “30” is returned from the device. In addition, in this case, when the sound spoken by the user is inappropriate and the sound is not recognized, an operation can be performed such that a response from the device side is not created, and/or a response such as “please speak again” is performed.
Thus, when the recognition result cannot be responded per word spoken by the user and the sound is not recognized, if an operation is performed such that some response is returned, the user can find out whether the content spoken by himself/herself is not appropriate, and how the sound is recognized, so that the user feels relieved and can easily use the device.
However, as described earlier, if the sound is recognized per word and responded to the user, if one setting operation such as time setting is performed, this time-consuming operation can create problems. Furthermore, if this type of sound recognition technology is applied to a device which requires low cost, such as daily necessities and toys, it is necessary to reduce the cost as much as possible, so there are significant restrictions on processing ability of a CPU and on the memory capacity. Therefore, the CPU needs to bear a large burden on the device side, and operations which use a large amount of memory must be reduced as much as possible.
In order to solve this problem, for example, in the case of the time setting described earlier, instead of recognizing the sound and responding with a recognition result when a user speaks one word, it is conceivable to have the user speak words that form one group, such as “a.m.”, “6”, and “30”, intermittently while leaving a small interval after every word, as the necessary content to set the time, and to perform sound recognition with respect to this spoken content. In this case, because there is no word-for-word response of the recognition result described earlier from the device for each of a plurality of words forming one group, it is possible to shorten the time setting period.
However, in a method in which a relatively long series of sounds forming a plurality of words is input to the device from beginning to end, as described earlier, the user may have a concern as to whether a sound per word spoken by himself/herself has been input in an appropriate state. Therefore, it is becoming necessary to inform the user of whether the sound spoken by the user was input in an appropriate state, without having troublesome processing.
SUMMARY OF THE INVENTION
Therefore, one aspect of this invention is to improve the convenience of a device during a sound inputting operation, and to inform a user whether the sound is appropriately input by performing a simple process when a sound is recognized with respect to a sound spoken by the user.
In order to accomplish this aspect, the method of notifying of an input state of sound to be recognized includes detecting an effective sound division for a sound to be recognized based upon the sound power which is obtained from a sound wave form of a sound to be recognized that is spoken by a speaker, determining whether the sound to be recognized has been input in an appropriate state, depending upon a time length of the effective sound division and magnitude of sound power within the effective sound division, and generating information showing that the sound is appropriate immediately after the completion of inputting of the sound to be recognized when it is determined that the sound is appropriate.
Furthermore, the sound to be recognized, for which it has been determined whether the sound has been input in an appropriate state, may be sound spoken with a plurality of words as one group, and may be spoken having a space, between each sound for each word forming this one group, as divisions for each word.
The information which is generated when the sound to be recognized, for which it is determined whether the sound has been input in an appropriate state, is determined to be appropriate, may be at least one of a sound signal, light, a sound message, and a display on a display screen which is instantly output in the spaces that form divisions for each of the words that form the one group.
The plurality of words may include one group belonging to a first through an nth (n is a positive integer) word group, the order of being spoken being determined from a word which belongs to the first word group to a word which belongs to the nth word group, and a reference which determines the time length of the effective sound division being set for each word group.
Additionally, the sound recognition device may also include a sound inputting device which inputs sound to be recognized spoken by a speaker and outputs the sound as digitized sound data; a sound analysis device which analyzes the sound data which has been output from the sound inputting device per predetermined time interval and calculates sound power and characteristic data per predetermined time interval; a sound division detection/determination device which detects effective sound division for the sound to be
Dorvil Richemond
Nolan Daniel A.
LandOfFree
Confirmation notification by apparatus using audio... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Confirmation notification by apparatus using audio..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Confirmation notification by apparatus using audio... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2839778