JPS602676B2 - audio output device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an audio output device used in places where noise is large and the noise level is not constant.

Naturally, in a noisy place, the audio output level must be high, but conventionally this adjustment has typically been done manually.

However, when there is no human intervention or in a small place, the sound remains silent, and therefore the information transmission by voice from the speaker has no choice but to take an unnatural form. recent years,
Semiconductor-based speech synthesizers are becoming more common, and problems arise when they are installed in, for example, vending machines. During the daytime, when the noise level is high, the speech synthesizer needs to generate a sound with a considerably high level of amplitude, while at night, etc., it is not necessary to produce a sound with an unnecessarily high level of amplitude. Furthermore, the same can be said of noise not only in its amplitude level but also in its frequency characteristics, and it is effective to synthesize speech having frequency characteristics different from the frequency characteristics of noise. The present invention proposes an audio output device that can satisfy such needs, and one embodiment thereof will be described below with reference to the drawings.

In Figure 1, Ichikawa is a speech synthesizer, which imitates the human vocal tract.

The pseudorandom signal output from the nozzle generator 11 and the periodic signal output from the periodic impulse generator 12 are switched by a voiced/unvoiced switch 13 and input to an amplitude determining section 14. The amplitude determination unit 14 multiplies the input signal by a predetermined amplitude and inputs the result to the vocal tract imitation filter 5. Vocal tract imitation filter evening 1
5 filters the input signal according to a predetermined filter constant and outputs it. The period of the periodic signal generated by the periodic impulse generator 12, the voiced/unvoiced switching signal of the voiced/unvoiced switch 13, the amplitude data of the amplitude determination section 14, and the vocal tract imitation filter 15
The filter constant of is given by the parameter supplier 16. The output of the vocal tract simulation filter 5 passes through the characteristic determiner 20 and is output from the speaker 40 as synthesized speech. 50' is a sound collector.

A sound collector 5 collects ambient sounds and converts them into electrical signals, which are input to an ambient sound level memory 301. The characteristic determiner 2 outputs the ambient sound level memory 30 in response to a command signal input from the command terminal 61 and passed through the gate 62 to start operation.
Its input/output characteristics are determined by referring to the signals from. Further, the command signal passing through the gate 62 is simultaneously input to the parameter supply device 16. Furthermore, ambient sound level memory 3
The output voltage of 0 is compared with the output voltage of the reference voltage generator 63 by a comparator 64, and the output signal of the comparator 64 is input to the gate 62 together with the command signal input from the command terminal 61. When the output voltage level of the reference voltage generator 63 is higher than the output voltage level of the ambient sound level memory 30, the parameter supply unit 16 supplies parameters to each block in the speech synthesizer 10 by a command signal from the command terminal 61. Start supplying the voice synthesizer 1
0 works. However, when the output voltage level of the reference voltage generator 63 is lower than the output level of the ambient sound level memory 30, the command signal from the command terminal 61 does not pass through the gate 62, and therefore the voice synthesizer 1 Well, it doesn't work.
The interior of the ambient sound level storage device 30 is configured as shown in FIG.

FIG. 3 is a diagram illustrating the operation of the ambient sound level storage device 30. In FIG. 2, a signal output from a sound collector 50 is input to a full-wave rectifier 31 and output as a rectified waveform. An integrator 32 integrates the output of the full-wave rectifier 31. In FIG. 3, a and b are input/output signal waveforms of the full-wave rectifier 31, and a tooth-shaped wave d is an integrated waveform in the integrator 32. The integrator 32 is The clock generator 33 is driven by a clock J, , ◇2 as shown in FIG. 3c. Clock low 2 resets the integrator, and clock low 2 holds the output of the integrator. Therefore, the integrator 32 integrates the input signal level from clock ◇2 to clock ◇, holds its peak value, and outputs it. FIG. 3d shows this situation. 34 is an analog shift register. S, to SN are cells of the analog shift register 34. Each cell receives a clock as shown in FIG. 3e from the clock generator 33? 3,? 4, the analog memory amount is shifted from S to SN. The outputs of cells S, to SN are configured to store signals of about one second in total. Signals older than one second are not stored in analog shift register 34. The outputs in each cell S, to SN of the analog shift register 34 are added by an adder 35. As a result, the output of the adder 35 is the total integrated value of the signal for one second. FIGS. 3f and 3g show how the analog shift register 34 shifts in synchronization with the clocks 3 and 04 from the clock generator 33. FIG. The output of the adder 35 is input to a logarithmic converter 36 and subjected to logarithmic conversion. The characteristic determiner 2 of FIG. 1 receives the output from the logarithmic converter 36.
The output of the logarithmic converter 36 is measured at the time when the command signal to start operation is received from the characteristic determiner gate 62, and the gain at a level higher than this value by a certain decibel is stored. The output from the speech synthesizer 10 is amplified by the set gain. As a result, speaker 40
The level of the voice uttered is always set to be higher by a certain decibel than the average voice level of the sound collector 50 over the past one second. It is a very effective means to vary the vocalization level in response to the surrounding noise level. This makes it possible to obtain a sound that is sufficiently audible to humans without producing more sound than necessary. In FIG. 1, when the comparator 64 detects that the output level of the ambient sound level memory 30 exceeds the output level of the reference voltage generator 63, even if an input command signal is generated from the command terminal 61, the gate 62 Does not pass input command signals.

As a result, if the noise level for the past one second exceeds the limit value, the speech synthesizer 10 is prohibited from producing a sound. In the presence of extreme noise, it makes sense to prevent further sound pressure and the generation of bell sounds. Further, when the input signal to the command terminal 61 is a pulse having a certain width, the gate 62 waits for input from the converter 64 as long as the pulse is in the ON state.

Therefore, it is possible to start vocalization after waiting for the noise level for the past one second to become below a certain value. This is useful because even in a noisy environment, the voice can be emitted when there is an opportunity to become quiet for a certain moment. As mentioned above, the ambient sound level storage device 30 shown in FIG. 2 stores only input signals for the past one second.

Therefore, data past one second or more is not stored at all. If the filter is simply a capacitor and a resistor, some past data will remain. However, if extreme noise has been input for more than 1 second in the past, the ambient sound level memory 3
Since it takes time for the zero output to decay, this is not an appropriate means for the purpose of the present invention. In this example, since the method does not refer to past data for more than 1 second, the noise becomes quieter and
You can start speaking without waiting more than a second. Further, by changing the values of the resistors R, , R2, R3 to RN in the adder 35 shown in FIG. 2, it is possible to weight the noise level within the past second and change the criterion. FIG. 4 is a diagram illustrating another embodiment of the present invention.

The signal input from the sound collector 50 is passed through the bandpass filter 7
5 to 77. The respective outputs of bandpass filters 75-77 are input to level storage devices 72-74, respectively. The level setting devices 72 to 74 have the same structure as the ambient sound level storage device 30 shown in FIG. The outputs of the level stores 72-74 are input to a lowest level detector 78. Sound collector 30
The lowest level detector 78 detects the frequency band having the lowest level among the frequency components of the noise collected from the first frequency component. According to this result, the bias frequency determining circuit 71 determines the bias frequency. The parameters given to the periodic impulse generator 12 from the parameter supply device 16 in FIG. 1 are added to the output from the bias frequency determining circuit 71 in the adder 70 in FIG. The period of the impulse signal will be determined. When producing voiced sounds, it is an effective means for humans to identify voices by enabling them to produce sounds with frequency characteristics as different as possible from the frequency characteristics of external noise. Further, the frequency characteristic of the speech synthesizer may be determined by setting the passing frequency band of the characteristic determiner 20 (FIG. 1) with reference to the level of the frequency band of the ambient sound level storage device. . According to the present invention, it is possible to always automatically generate sound that is a certain decibel higher than the ambient sound level, so it is possible to obtain the optimum sound level regardless of where the noise level is not constant or at night. Nothing goes inaudible or too loud.

Moreover, it is effective because it is possible to synthesize not only the amplitude level but also the sound having frequency characteristics different from the frequency characteristics of the noise.

[Brief explanation of the drawing]

FIG. 1 is a block diagram explaining one embodiment of the present invention, and FIG.
The figure is a block diagram explaining the inside of the ambient sound level storage device.
FIG. 3 is a signal waveform diagram illustrating the operation of FIG. 2, and FIG. 4 is a block diagram illustrating another embodiment. 10... Speech synthesizer, 12... Periodic pulse generator, 16... Parameter supplier, 20
...Characteristic determiner, 30...Ambient sound level storage device, 50...Sound collector, 75-77...Band pass filter, 78... Minimum level detector. Figure 1 Figure 2 Figure 3 Figure 4

Claims (1)

[Claims] 1. A speech synthesizer that starts operating in response to a command, a sound collector installed around the speech synthesizer to collect external sounds, and a sound collector that accumulates the output of the sound collector for a certain period of time. and an ambient sound level memory for storing the ambient sound level, and when an operation start command to the speech synthesizer is generated, the output of the ambient sound level memory is referenced to determine subsequent generation conditions of the speech synthesizer. An audio output device characterized by: 2. When an operation start command is issued to the speech synthesizer, the gain of the speech synthesizer is set so that the subsequent output level of the speech synthesizer is a certain amount higher than the output level of the ambient sound level memory at that time. 2. The audio output device according to claim 1, wherein the audio output device is configured to be set. 3. If the output level of the ambient sound level memory device exceeds a certain upper limit value and a command to start operation is issued to the speech synthesizer, the above-mentioned sound is output until the output level of the ambient sound level memory device falls below the upper limit value. 3. The audio output device according to claim 1, wherein the operation of the synthesizer is delayed. 4 The ambient sound level storage device has an ambient sound frequency characteristic measuring device that measures the level of each frequency band of the output from the sound collector, and accumulates the level of each frequency band from the ambient sound frequency characteristic measuring device for a certain period of time. When an operation start command is issued to the speech synthesizer, the level for each frequency band in the ambient sound level storage device is referenced to determine the frequency characteristics of the speech synthesizer thereafter. 2. The audio output device according to claim 1, wherein: 5 The speech synthesizer has a periodic impulse generator that can set the sound source pitch, and when the operation start command to the speech synthesizer is issued, the level of each frequency band in the ambient sound level memory is referred to and the above-mentioned speech after that is referenced. 5. The audio output device according to claim 4, wherein the impulse period of the synthesizer is determined.