JPH0833738B2 - Noise level learning device - Google Patents

Description

The present invention detects a voice section from an input sound in which a voice section and a section including only noise in the vicinity where voice does not exist are continuous in voice recognition and the like. The present invention relates to a noise level learning device for ambient noise.

(Prior Art) FIG. 3 is a functional block diagram of a conventional noise level learning device. In FIG. 3, 31 is a handset of an automobile telephone, 32 is a preprocessing unit, 33 is a lock release lever for releasing the lock of the handset 31, 34 is release of the lock release lever 33 (corresponding to off-hook of a normal telephone). Detection unit, 36 is a noise level learning unit, 37 is a voice section detection, voice recognition,
The following parts are shown. This conventional example is an example of a noise level learning device for detecting a voice section in a voice recognition device for a car telephone.

Next, the operation of the above conventional example will be described with reference to FIG. When the speaker unlocks to pick up the handset, it detects it () and learns the noise level (). The noise level learning is 0 after the lock release is detected.
The power level of noise is observed for 4 seconds, the noise level is learned as an average level during that period, and then the voice section detection and voice recognition are performed based on the result ().

Thus, even with the above conventional noise level learning device,
Noise level learning is possible for the time being.

(Problems to be Solved by the Invention) However, the above conventional noise level learning device has the following problems. That is, the handset of a car phone is often installed between the driver's seat and the passenger's seat, but the noise level such as gear noise is high in that place, until the handset is actually lifted after the handset is unlocked. During that time, it learns a large level of noise. Furthermore, if noise is generated when the handset is picked up, it is also learned as a part of ambient noise.
In order to detect a voice section, if a method of inputting a level higher than the level of ambient noise continues, it is used as a voice section. It will be difficult to detect.

The present invention solves such a conventional problem, and an object of the present invention is to provide a noise level learning device capable of learning a correct noise level.

(Means for Solving the Problems) In order to achieve the above object, the present invention provides a voice input means (11, 12) for inputting voice and noise for learning a noise level based on the output of the voice input means. Level learning means (16), means (14) for receiving a speaker's utterance notice,
And a means (15) for causing the noise level learning means to start learning of the noise level after a lapse of a predetermined time after receiving the utterance announcement.

(Operation) Therefore, according to the present invention, noise level learning is performed after a lapse of a certain time after the speaker's utterance announcement, so that a large level of noise is input to the means for inputting voice immediately after the utterance announcement. Also, it has an effect that it can be excluded from the target of the noise level learning, and the noise level learning can be performed with high accuracy.

(Embodiment) FIG. 1 shows an embodiment of the present invention. This embodiment is also an example of a voice recognition device for a car telephone, like the conventional example. In FIG. 1, 11 is a handset that is a means for inputting voice, 12 is a preprocessing unit, 13 and 14 are means for receiving a speech announcement of a speaker, and 13 is a lever that unlocks the handset. Reference numeral 14 is a detection unit for releasing the lever 13, 15 is a portion for delaying the detection signal of the detection unit 14, 16 is a noise level learning unit, and 17 is a subsequent portion such as voice section detection and voice recognition.

Next, the operation of the above embodiment will be described with reference to FIG. When the speaker releases the lock in order to pick up the handset 11, the unlock detection unit 14 detects it (). The detection signal is delayed by the delay unit 15 for 0.3 seconds (). After the delay of the fixed time (0.3 seconds), the noise level learning unit 16 starts learning the noise level (). In learning the noise level, the power level of the noise is observed over 0.4 copies, and the noise level is learned as the average level during that period. After that, the post-processing unit 17 performs post-processing such as voice section detection and voice recognition based on the result.

As described above, according to the present embodiment, the learning of the noise level is not performed for a certain period of time after the handset is unlocked. There is an advantage that it is possible to prevent learning of noise of a large level and to perform accurate noise level learning.

(Effects of the Invention) As is apparent from the above-described embodiments, the present invention has the advantage that accurate noise level learning can be performed by performing noise level learning after a lapse of a certain period of time after receiving a voice announcement of a speaker. As a result, there is an effect that the accuracy of voice section detection and voice recognition performed later is also improved.

[Brief description of drawings]

FIG. 1 is a functional block diagram of an embodiment of the present invention, FIG. 2 is an operation explanatory diagram of the same embodiment, FIG. 3 is a functional block diagram of a conventional noise level learning device, and FIG. Is. 11,31 ...... Handset, 12,32 ...... Preprocessor, 13,33 ...... Lever for unlocking handset, 14,34 ...... Handset unlocking detector, 15 ...... Detection signal Delay part of
16,36 …… Noise level learning unit, 17,37 …… Post-processing unit for voice section detection, voice recognition, etc.

Claims (3)

[Claims] 1. A voice input means for inputting a voice, a noise level learning means for learning a noise level based on the output of the voice input means, a means for receiving a speaker's utterance notice, and a utterance notice after receiving the utterance notice. A noise level learning apparatus comprising: a noise level learning means for starting a noise level learning after a lapse of a predetermined time. 2. A noise level learning device according to claim 1, wherein a handset of a telephone is used as the voice input means. 3. The noise level learning device according to claim 2, wherein the telephone is an automobile telephone and a lever for unlocking the handset is used as a means for receiving a speech announcement of a speaker.