JPH0427559B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method of perceiving music and sound effects by vibrating the body with audio signals of music and sound effects, among the methods of perceiving music and sound effects through the ears and through vibrations of the body. The present invention relates to an extraction circuit for a sensory acoustic vibration signal that distinguishes between music, sound effects, and human voices among audio signals.

Music and sound effects are transmitted to the body through the air or the ground. The sound effects referred to here include, for example, the sound of a cannon being fired, the sound of a train or steam locomotive running, the sound of a large tree being cut down, the sound of an explosion, the sound of a car or tractor running, or the sound of an engine, etc., which are accompanied by a tremor or impact. It's something like this. There are sensory acoustic devices that actively convert these signals using electronic equipment (microphones) and attempt to vibrate the body using audio signals.

This tactile sound device uses a certain low frequency range of the amplified cut output frequency applied to the speaker (or headphone), for example, a low frequency range of 150 Hz or less, in order to create a sense of presence in sound reproduction. This is a device that amplifies bass frequencies to generate bodily-sensible acoustic vibrations, giving the viewer sound and bass vibrations at the same time to enhance the sense of deep bass, and is primarily effective for music listening. It is.

When this tactile sound device is activated and combined with the sound coming in through the ears, a sense of realism that cannot be obtained with conventional sound reproduction devices can be obtained, and when combined with movies etc., the dramatic effect is enhanced. Confirmed by experiment.

In addition, feature films and TV dramas are composed of a combination of human voices (conversation), music, and sound effects, and these audio signals, especially human voices (conversation), are amplified so that viewers can experience them. It has been experimentally discovered that when applied as acoustic vibrations, it feels very unnatural.

Therefore, in order to obtain good effects such as a sense of realism in feature films and TV dramas, there is a need for a method that allows the sensory sound system to operate with low-frequency sounds such as music and sound effects, but not with human voices. be.

Here, the frequency components of human voices, sound effects, and music will be compared.

The frequency components of the human voice range below 80 Hz (mainly for male voices) in the low range, and above 10 kHz in the high range for both male and female voices.

There are many types of sound effects, and there are sounds with various frequency spectra, but sounds that cause an earth-shaking or shocking sensation are characterized by a wide frequency distribution and rich deep bass components. .

Music often has a relatively uniform frequency distribution compared to other sounds.

For this reason, if you simply pass the signal through a low-pass filter that cuts around 80 Hz, most of the frequency components that are effective for sensory acoustic vibrations will be cut out, and the effect will be significantly impaired. Become.

This is because, among the frequency components below 150 Hz that are effective as sensory acoustic vibrations, the spectrum around 100 Hz has the highest frequency distribution density (the distribution density below 50 to 60 Hz is low).

Therefore, if a low-pass filter is used that cuts out human voices by setting the cutoff frequency to 60 hertz or lower, there will be no frequency components that are effective for sensory acoustic vibrations, and the effect will deteriorate.

Sounds that cause earth tremors (for example, sounds that include components between 20 and 200 Hz) have many components below 60 Hz, so you can use the above low-pass filter, but the components around 100 Hz will be cut out. This results in a less realistic feel and a loss of harmony with the sound heard by the ear.

FIG. 1 is an example of a circuit that does not cause such problems. To explain this, 1 is an input terminal which receives an audio signal from an amplifier.
Between the input terminal 1 and the output terminal 2, there is a first low-pass filter 3 with a cutoff frequency of about 150 heretz, and a voltage-controlled variable gain amplifier (UCA).
4 is connected.

The voltage-controlled variable gain amplifier 4 continuously controls the degree of amplification based on the input from the control end 4a to give a natural feel.

Input terminal 1 also has a cutoff frequency of 60
The input side of a second low-pass filter 5 of about Hertz is connected, and the output side of this second low-pass filter 5 is connected to the control circuit of the voltage-controlled variable gain amplifier 4 described above via a detection circuit 6 and an integration circuit 7. end 4
connected to a.

In this circuit, the input terminal 1 is connected to the input audio signal, which is a sound effect that is accompanied by a rumbling sound that is rich in ultra-low frequencies below 60 Hz (for example, from 20 Hz to
When a sound containing a component of 200 Hz comes in, one low-pass filter 3 outputs an output signal with a component frequency of 150 Hz or less, and the other low-pass filter 5 outputs a component frequency of 60 Hz or less. Ru. The output signal of the second low-pass filter 5 is converted to a DC level through a detection circuit 6 and an integration circuit 7, and is applied to the control terminal 4a of the voltage-controlled variable gain amplifier 4 to increase its amplification gain. The signal that has passed through the filter 3 is greatly amplified and is then applied to the output terminal 2.

Therefore, if the above-mentioned sensory acoustic device is connected to the output terminal 2, the viewer can perceive the sensory acoustic vibrations caused by the sound effects. This output signal is 100
Since it includes frequency components around hertz, it can produce realistic vibrations.

When a human voice signal (for example, a frequency that includes a component of 80 Hz or higher) is input as an input signal, there is no ultra-low frequency component that passes through the second low-pass filter 5, so the voltage-controlled variable gain amplifier 4 changes the gain. becomes low, and the output signal becomes 0 at output terminal 2,
Or, it becomes so small that human voices do not generate any bodily-sensible acoustic vibrations.

Although the initial purpose is achieved with this circuit, the music or sound effects for which you want to generate sensory acoustic vibrations do not contain components below 60 Hz, or even if they do, they do not contain components below 60 Hz. There are also quite a few products with very few ingredients. For example, if the component frequency is between 80 hertz and 200 hertz, the first low-pass filter 3 passes the component frequency of 150 hertz or less, but the output signal of the second low-pass filter 5 does not come out, so the output terminal 2 will have no output.

Considering the generation and characteristics of the human voice, the sound source generated from the vocal cords is a pulse wave similar to a sawtooth wave whose fundamental wave component varies over a fairly wide range of around 120 Hz in an adult male voice during normal conversation. This sawtooth wave from the vocal cord sound source (hereinafter referred to as a sawtooth wave) is known to have complex formant frequencies such as resonance and anti-resonance of the vocal tract consisting of the pharynx, oral cavity, nasal cavity, etc. (This resonant frequency is generally distributed much higher than the fundamental frequency of the sawtooth wave of the vocal cord sound source.)

This can be schematically illustrated as shown in FIG. 2. Figure a shows the sawtooth wave of the vocal cord sound source. Since the sawtooth wave emitted from the vocal cords contains abundant harmonics, when a formant frequency is applied to the resonant circuit section of the vocal tract, it becomes as shown in Figure b. By the way, in a general transmitter that handles frequencies, when a sawtooth wave is amplitude-modulated on a carrier wave, a waveform as shown in c in the figure is obtained. Figure 2 b and c have very similar waveforms, so if we consider the speech waveform in Figure 2 b to be a type of amplitude modulated wave, by detecting it, we can easily determine the fundamental wave component of the vocal cord sound source. It becomes possible to take it out.

The present invention provides a sensory acoustic vibration signal extraction circuit which solves the above-mentioned problems by distinguishing between human voices and other voices, and which can be applied to various applications with good results.

Hereinafter, a circuit for extracting a sensory acoustic vibration signal according to the present invention will be explained based on FIG.

In the circuit shown in FIG. 3, input terminal 1 has
Low pass filter 3 (cutoff circumference fraction is approximately 150
Hertz), an input level extraction circuit 8 and a human voice fundamental wave emphasis extraction circuit 9 are connected. The output sides of the input level extraction circuit 8 and the human voice fundamental wave emphasis extraction circuit 9 are connected to two input terminals 10a and 10b of an arithmetic circuit 10. Further, the output signal of the low-pass filter 3 is input to a voltage-controlled variable gain amplifier 4. The output side of the arithmetic circuit 10 is connected to the control terminal 4a of the voltage controlled variable gain amplifier 4.

The input level extraction circuit 8 is a circuit that integrates the detected input signal and outputs a DC level proportional to the input level.

Theoretically speaking, the human voice fundamental wave emphasis extraction circuit 9 consists of a detection circuit 11 and 80 to 80, as shown in FIG.
A band filter 12 selects 300 Hz.

When a human voice waveform as shown in FIG. 2b is input to this circuit and detected, the formant frequency is removed and a positive envelope or a negative envelope is obtained. When extracting either the positive or negative envelope, the one with the larger amplitude can be extracted to increase sensitivity.

On the other hand, if an audio signal other than a human voice is passed through the circuit shown in FIG. 4, not much output will be produced at the output terminal. This is because the waveform of the signal has a wide frequency range, so the percentage of components that pass through the bandpass filter 12 is small, and the waveform has a rule that makes it look like an amplitude modulated wave as shown in Figure 2b. This is due to the fact that it does not have a typical shape.

That is, in FIG. 4, the output level that is higher than the input level is human voice, and the output level that is lower than the input level is other voices, and it is possible to extract the output of only the fundamental wave component of the vocal cord sound source. Then, it is sufficient to convert this output level into a direct current and compare it with the output of the input level extraction circuit 8.

The circuit shown in Figure 5 is an implementation circuit that adds functions to the circuit shown in Figure 4 to process the envelopes on the positive side and negative side, and to convert the output level to DC, which is not necessary for human voice recognition at the front stage. High pass filter (low cut filter) 13 that cuts out extremely low frequencies
A fully rectified detection circuit 14 is provided to extract the positive and negative envelopes, and an AC-DC conversion circuit 1 is provided at the subsequent stage.
5 and a low-pass filter 16. The function of the high-pass filter 13 is such that even if the signal is frequency-multiplied by the full-wave rectification type detection circuit 14, the number of components passing through the band-pass filter 12 is reduced. According to the experimental results, it was appropriate to set the cutoff frequency to about 130 hertz.
In this way, only the human voice signal is extracted, and this is converted to a DC level and output.

Then, the arithmetic circuit 10 subtracts the output of the human voice fundamental wave level emphasis extraction circuit 9 from the output of the input level extraction circuit 8. The output signal of the input level extraction circuit 8 and the output signal of the human voice fundamental wave emphasis extraction circuit 9 are each weighted to an appropriate level, and also in the time axis domain, the time constant of the integration circuit, Alternatively, appropriate weighting is performed, such as by selecting the cutoff frequency of the smoothing low-pass filter 3.

In actual operation, if the input signal contains a human voice, the output of the input level extraction circuit 8 is high, and the output of the human voice fundamental wave emphasis extraction circuit 9 is high, so the output of the arithmetic circuit 10 is is a small signal (a solution in the negative direction), and when the input signal does not contain human voices and when music or other sounds are input to the input level extraction circuit 8, the output of the input level extraction circuit 8 is large and the output of the human voice fundamental wave emphasis extraction circuit 9 is small, so the output of the arithmetic circuit 10 becomes a large signal (a solution in the positive direction). It should be noted that if there is no input to the input level extraction circuit 8 and the human voice fundamental wave emphasis extraction circuit 9, this is not the purpose of the description and will therefore be omitted.

In this way, if the input signal has the same frequency as the human voice,
If the signal is other than human voice, the amplification degree of the voltage-controlled variable gain amplifier 4 can be increased based on the calculation result, and a large output can be obtained. Note that if a human voice is included, it goes without saying that the output of the voltage-controlled variable gain amplifier 4 will be zero or small.

The circuit shown in FIG. 6 is a modification of the circuit shown in FIG. This circuit does not use full-wave rectification types with frequency multiplication for the detection circuits 17 and 18, but uses the single-wave rectification type detection circuit 17 to detect the positive envelope, and also uses the single-wave rectification type detection circuit 17 to detect the positive envelope. A circuit 18 extracts the negative envelope. The polarity of the negative side detection output is inverted by the inverter circuit 19.

The respective detection outputs enter respective hand-pass filters 20 and 21, and are then sent to the AC-DC conversion circuit 2.
2 and 23 and are converted to a DC level, the signals are combined and input to the low-pass filter 16. In addition,
Although the high-pass filter 13 in the input section does not need to be provided in principle, it is better to provide it in order to improve the discrimination ability. The detection circuits shown in Figures 4 to 6 and
The AC-DC conversion circuit is preferably an absolute value circuit or a linear detection circuit in order to reduce calculation errors due to non-linearity.

Next, details of the embodiment will be explained with reference to the circuit shown in FIG. In this circuit, input terminal 1
Two channels are provided to receive signals through two channels, and the mixer circuit 24 mixes the signals. On the output side of this mixer circuit 24, there is a subsonic filter 2 that removes unnecessary subsonics of 20 or less frequencies.
5 is connected.

The output side of the subsonic filter 25 is connected to a circuit that goes to a voltage-controlled variable gain amplifier 4 through a low-pass filter (cutoff frequency 150 Hz) 3 for extracting sensory acoustic vibration signals, and an equalizer circuit 26.
It is branched into a circuit that goes to . This equalizer circuit 26 provides an appropriate equalizing curve in advance to the low and high frequencies in order to improve the human voice discrimination ability.

A level compression circuit 27 is connected to the next stage of the equalizer circuit 26, and its output side receives the inputs of an ultra-low frequency extraction circuit 28, an input level extraction circuit 8, and a human voice fundamental wave emphasis extraction circuit 9. It is connected. The output side of these is the input terminal 1 of the arithmetic circuit 10.
It is connected to 0c, 10a, and 10b.

The level compression circuit 27 compresses the dynamic range of the input signal to prevent errors in identification and calculation (if the level is too low, it will be difficult to distinguish, and if the level is too high, the operational amplifier will saturate). The ultra-low frequency extraction circuit 28 is a circuit that detects sounds rich in ultra-low frequencies that are accompanied by earth rumblings, and has the same function as the part surrounded by the dotted line in FIG. 1. Further, the input level extraction circuit 8 and the human voice fundamental wave emphasis extraction circuit 9 correspond to the circuit for the same signal in FIG. 3, respectively.

In this configuration, the arithmetic circuit 10 functions to subtract the output of the human voice fundamental wave emphasis extraction circuit 9 from the sum of the output of the very low frequency extraction circuit 28 and the output of the input level extraction circuit 8. This solution controls the voltage-controlled variable gain amplifier 4 in the same way as in FIG. 3. Note that good results can be obtained by setting the weighting of the very low frequency extraction circuit 28 relatively high and by providing a dead zone so that it does not operate at a level below the threshold value.

As explained in the embodiments above, the present invention uses a human voice fundamental wave emphasis extraction circuit to remove formant frequencies given by the vocal tract etc. from the voiced part of the human voice, and extract the vocal cord sound source from the input signal. The fundamental wave component can be emphasized and extracted. The input level extraction circuit and the arithmetic circuit reduce the amplification degree of the voltage-controlled variable gain amplifier when the level of one of the extracted human voice signals is high, so that the low frequency range extracted from the input signal is reduced. By passing a signal through this amplifier, if the input signal contains a human voice, it becomes a zero or small output signal;
Signals that are not valid for sensory acoustic vibrations are removed,
Voices other than human voices can be used as low frequency signals.

[Brief explanation of drawings]

Fig. 1 is a block diagram showing an example of the basic circuit of the present invention, Fig. 2 is a waveform diagram schematically showing audio waveforms such as human voice and music, and Fig. 3 is a block diagram showing an example of the circuit that is the basis of the present invention. FIG. 4 is a block diagram showing the basic part of the human voice fundamental wave emphasis extraction circuit in FIG. 3, and FIGS. 5 and 6 are human voice A block diagram of the fundamental wave emphasizing extraction circuit. FIG. 7 is a block diagram showing an extraction circuit of a sensory acoustic vibration signal according to another embodiment. 1...Input terminal, 2...Output terminal, 3, 5, 1
6...Low pass filter, 4...Voltage controlled variable gain amplifier, 4a...Control end, 6, 11...Detection circuit, 12...Band pass filter.

Claims (1)

[Claims] 1. An input level extraction circuit that connects a voltage-controlled variable gain amplifier to an input terminal via a low-pass filter that extracts a low-frequency signal for giving bodily-sensible acoustic vibrations, and outputs a DC level proportional to the input signal to the input terminal. and a human voice fundamental wave emphasizing circuit that detects the input signal to emphasize and extract the fundamental wave component of the vocal cord sound source from the voiced part of the human voice by excluding formant frequencies given by the vocal tract, etc. the output terminals of the input level extraction circuit and the human voice fundamental wave emphasis extraction circuit are connected to the input terminal of an arithmetic circuit, and the output terminal of the arithmetic circuit is connected to a control terminal of the voltage-controlled variable gain amplifier. , wherein when the output level of the human voice fundamental wave emphasis extraction circuit is extracted, the amplification degree of the voltage-controlled variable gain amplifier is reduced in accordance with the output level of the arithmetic circuit. extraction circuit.