close

Вход

Забыли?

вход по аккаунту

?

DESCRIPTION JP2001237920

код для вставкиСкачать
Patent Translate
Powered by EPO and Google
Notice
This translation is machine-generated. It cannot be guaranteed that it is intelligible, accurate,
complete, reliable or fit for specific purposes. Critical decisions, such as commercially relevant or
financial decisions, should not be based on machine-translation output.
DESCRIPTION JP2001237920
[0001]
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an
input level adjusting circuit used in wired communication and wireless communication.
[0002]
2. Description of the Related Art In recent years, the penetration rate of PHS and mobile phones
has been rapidly increasing, and it has become possible to talk with anyone anytime, anywhere,
but along with it there is a problem of manners used. When used in transportation such as a train
bus or in a place where people gather, the ringing tone may make the third party uncomfortable.
In addition, they tend to forget to care about the size of their voice because they talk to a person
who is not on the spot, and the situation in which only one side of the conversation can be heard
makes the third party uncomfortable.
[0003]
The ringing tone has been avoided by the function that the terminal itself vibrates. In addition, in
the case of conversation, an external switch generally called a manner mode is provided, and the
voice output level of the other party is secured by raising the input level of the terminal even
when speaking in a low voice by turning on the switch. .
10-04-2019
1
[0004]
FIG. 9 is a block diagram showing the configuration of a conventional input level adjusting circuit
of this kind. In FIG. 9, 1 is a microphone for inputting an audio signal, 2 is an amplifier for
amplifying an audio signal inputted through the microphone 1, and 3 is an analog-digital
conversion for converting an analog audio signal amplified by the amplifier 2 into a digital signal.
A voice coder 4 encodes the digital voice signal output from the analog-to-digital converter 3.
Reference numeral 5 is a switch, and 6 is a controller that controls the operation of the switch 5
to increase the gain of the amplifier 2.
[0005]
Next, the operation will be described. The voice is taken from the microphone 1, amplified by the
amplifier 2, and converted into a digital signal by the analog-to-digital converter 3. The voice
signal converted into the digital signal is encoded by the algorithm defined by the voice encoder
4 and output to the voice communication terminal system. Depending on the usage situation,
when the voice is lowered for use, pressing the switch 5 causes the controller 6 to increase the
gain of the amplifier 2 by a predetermined amount to obtain the level of the output voice heard
by the other party in the conversation. increase.
[0006]
However, in the above-described conventional level adjusting circuit, by pressing the switch 5,
the gain of the amplifier 2 is increased by the controller 6 to increase the level of the output
sound heard by the other party in the conversation. Because the input level at the other party's
terminal is fixed, in some cases the other party may be made to manipulate the reception volume,
and if the gain is easily increased, background noise will be amplified. There was a problem of
transmission.
[0007]
The object of the present invention is made in view of the above-mentioned point, and in order to
cope with various usage situations such as PHS and mobile phones, the output voice to be output
to the other party is made constant for the user's voice level. An object of the present invention is
to provide an input level adjustment circuit capable of automatically adjusting an input level at a
communication terminal so as to maintain the same.
10-04-2019
2
[0008]
According to the present invention, an input level adjusting circuit comprises: a microphone for
inputting an audio signal; an amplifier for amplifying the level of the audio signal input through
the microphone; and an amplifier for amplifying the level of the audio signal. An analog-to-digital
converter for converting an audio signal into a digital signal, a speech encoder for encoding an
output signal from the analog-to-digital converter, and whether the output from the analog-todigital converter is an audio signal A voice detector that outputs voice presence / absence
information, a level averaging circuit that determines input level information of the microphone
by determining the input level of the microphone, a switch for setting the gain adjustment of the
If the adjustment is set and the voice presence / absence information from the voice detector
indicates a voice signal, the input level from the level averaging circuit And a controller for
adjusting the gain of the amplifier according to bell information.
[0009]
The above input level adjustment circuit according to the present invention can be mounted on a
voice communication terminal device.
Also, the speech detector performs an autocorrelation process on the speech signal digitized by
the analog-to-digital converter in a frame unit of a predetermined number of samples and
outputs an autocorrelation coefficient; A pitch detector for obtaining a maximum value from the
zeroth-order autocorrelation coefficient value and the autocorrelation value excluding the zeroth
order within the number of samples based on the autocorrelation coefficient, and outputting the
maximum coefficient value, and the pitch detector A value obtained by multiplying the zerothorder autocorrelation coefficient value by a constant is used as a threshold for sound / silence
determination, and the threshold value is compared with the maximum count value from the
pitch detector to calculate the maximum count value. It can be configured to be provided with a
sound / non-speech determining unit that determines sounding when the threshold value is
greater than the threshold value and silence when the maximum count value is less than the
threshold value and outputs sound presence / absence information.
[0010]
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a block diagram showing
the configuration of an input level adjusting circuit according to an embodiment of the present
10-04-2019
3
invention.
In FIG. 1, the same parts as those of the conventional example shown in FIG.
As a new code, 7 is a voice detector that determines whether the output from the analog-todigital converter 3 is a voice signal and outputs voice presence / absence information; 8 averages
the output level from the analog-to-digital converter 3 Level adjustment circuit for outputting the
input level information of the microphone 1, gain adjustment is set by the switch 5 at 6A, and the
audio presence / absence information from the audio detector 7 indicates an audio signal, the
level average circuit 8 outputs It is a controller that adjusts the gain of the amplifier 2 according
to the input level information.
[0011]
Next, an operation according to the above configuration will be described. The audio signal input
from the microphone 1 is amplified by the amplifier 2, converted to a digital signal by the
analog-to-digital converter 3, and encoded by the audio encoder 4. The encoded voice signal is
output to a voice communication terminal system not shown.
[0012]
The voice detector 7 inputs the digitized voice signal from the analog-to-digital converter 3,
determines whether the signal input from the microphone 1 is voice or not, and outputs voice
presence / absence information. The level averaging circuit 8 averages the output level from the
analog-to-digital converter 3 and outputs the input level information of the microphone 1. The
switch 5 is a switch for setting the gain of the amplifier 2 to be automatically adjusted by the
input level from the microphone 1.
[0013]
When the controller 6A is set to the automatic adjustment state by the switch 5, when the voice
presence / absence information determined to be a voice signal is inputted from the voice
detector 7, the input from the level averaging circuit 8 The gain of the amplifier 2 is adjusted
10-04-2019
4
according to the level information to control the input signal to an appropriate level so that the
magnitude of the input signal becomes a predetermined level. As described above, by using the
voice presence / absence determination result for the gain adjustment of the amplifier 2, only the
background noise is not amplified, and only the voice signal can be automatically appropriately
adjusted.
[0014]
Here, an internal configuration of the voice detector 7 is shown in FIG. As shown in FIG. 2, as the
voice detector 7, an autocorrelator 71 which performs autocorrelation processing on the voice
signal digitized by the analog-to-digital converter 3 in frame units of a predetermined number of
samples and outputs an autocorrelation coefficient. And a pitch detector that obtains a maximum
value from the zeroth-order autocorrelation coefficient value and the zeroth-order
autocorrelation value within the number of samples based on the autocorrelation coefficient from
the autocorrelator 71 and outputs the maximum coefficient value 72. A value obtained by
multiplying the zeroth-order autocorrelation coefficient value from the pitch detector 72 by a
constant is used as a threshold for the presence / absence determination, and the threshold and
the maximum count value from the pitch detector 72 are calculated. In comparison, the system is
provided with a voice / non-voice determination unit 73 that determines as voiced when the
maximum count value is equal to or greater than the threshold value and as silence when the
maximum count value is less than the threshold value and outputs voice presence / absence
information. There is.
[0015]
Next, the operation of the speech detector 7 shown in FIG. 2 will be described. The autocorrelator
71 autocorrelates the voice signal digitized by the analog-to-digital converter 3 in a frame unit of
a predetermined number of samples and outputs an autocorrelation coefficient. The
autocorrelation coefficient is input to the pitch detector 72, and the pitch detector 72 detects the
maximum coefficient value between certain specific orders excluding the zeroth order. Here, the
next numerical value having the maximum coefficient value is the pitch period, which is a feature
of human speech (vowel). The zeroth-order autocorrelation coefficient value after detection and
the detected maximum coefficient value are output to the speech / non-speech determiner 73.
[0016]
10-04-2019
5
The zero-order autocorrelation coefficient is multiplied by a fixed constant which is the sound /
non-speech determiner 73, and the multiplied value is used as the threshold of the sound / nonspeech determination. The speech / non-speech determiner 73 compares the threshold with the
maximum coefficient value from the pitch detector 72, and determines that the speech above the
threshold is speech and the speech below the threshold is silence. Based on the determination
result of a certain number of times, output as sound / silence information.
[0017]
Next, each component of the autocorrelator 71 shown in FIG. 3 will be described. FIG. 3 is a block
diagram showing an internal configuration of the autocorrelator 71. As shown in FIG. In FIG. 3,
71a is a memory for storing a plurality of samples of digitized audio signals, 71b is a memory
similar to the memory 71a, but a delay circuit for shifting and outputting sample data from the
clock of the clock generation circuit 71c. Is a clock generation circuit that generates a shift
operation timing clock of the delay circuit 71b, 71d is a multiplier that multiplies the output of
the memory 71a and the output of the delay circuit 71b and outputs the result to a combining
circuit, 71e combines the output of the multiplier 71d Synthesis circuit to output the
[0018]
Next, the operation of the autocorrelator 71 shown in FIG. 3 will be described. The pitch
detection of human voice is generally detected between 20 msec and a cycle (50 Hz) to 2.5 msec
and a cycle (400 Hz). Here, as an example, in order to detect this range, it is assumed that 320
samples (40 msec) in which the sampling frequency of the analog-to-digital converter 3 in FIG. 1
is 8 kHz are processed as one frame. If an autocorrelation from the 20th order (1/8000 Hz × 20
= 2.5 msec) to the 160th order (1/8000 Hz × 160 = 20 msec) is taken in this frame, the
detection range of the above period is satisfied.
[0019]
The 320 samples of the digitized (data sampled at 8 kHz, hereinafter referred to as sample data)
audio signal from the analog-to-digital converter 3 are stored in the memory 71a and in the delay
circuit 71b. An output in which 320 sample data (BD1 to BD320) output from the memory 71a
and an output sample signal (WD1 to WD320) of the delay circuit 71b are shifted one sample at
10-04-2019
6
a time with the clock generated by the clock generation circuit 71c. Are multiplied by one sample
shift in the multiplier 71d, and the multiplication outputs (GA1 to GA320) are output to the
combining circuit 71e. Here, the clock frequency is made sufficiently faster than the sampling
frequency, and the shift for 160 sample data, which is a range obtained in one sample time, is
completed. The multiplication outputs (GA1 to GA320) of the multiplier 71d are synthesized
every one sample shift, and are output as a synthesized output of the synthesis circuit 71e.
[0020]
Here, the combined output to be output is referred to as an autocorrelation coefficient, and the
combined output before shift is a zeroth order autocorrelation coefficient value, and every one
sample shift (delay), the combined output is second order, third order,. *, N-1st and nth
autocorrelation coefficient values are called. Here, as an example, it is assumed that
autocorrelation coefficient values from the 0th order to the 160th order are handled.
[0021]
FIG. 4 shows the input state of the multiplier 71d of FIG. As shown, starting with shift 0 (delay 0),
each sample is shifted (delayed) and multiplied. Further, FIG. 5 shows the timing for taking out
the multiplication result (autocorrelation coefficient). As shown in FIG. 5, the autocorrelation
coefficient is extracted in accordance with the shift timing of the delay circuit 71b.
[0022]
Next, FIG. 6 is a block diagram showing an internal configuration of the pitch detector 72. As
shown in FIG. In FIG. 6, a switch 72a switches the input of the memory 72b, a memory 72b
stores the autocorrelation coefficient value, a switch 72c switches the output from the memory
72b, a comparator 72d and a memory 72e.
[0023]
Next, the operation of the pitch detector 72 shown in FIG. 6 will be described. The voiced voice of
human voice has a feature of repeating the same waveform as a pitch cycle. This cycle can be
10-04-2019
7
confirmed by the cycle (interval) of the peak of the autocorrelation coefficient value (a high value
appears at certain intervals). For this reason, in the case of unvoiced sound or white noise in
which the interval between the zeroth-order autocorrelation coefficient value and the peak of the
next largest value in the detection range is the pitch cycle, no feature is seen that the same
waveform repeats, so There is no tendency to represent the period in
[0024]
The autocorrelation coefficient value input from the autocorrelator 71 is sequentially stored from
the 0th order to the 160th while the input to the memory 72b is switched by the changeover
switch 72a. First, the zero-order autocorrelation coefficient value is output as it is. The 20th to
160th autocorrelation coefficient values are input to the comparator 72d by sequentially
switching the output of the memory 72b by the changeover switch 72c.
[0025]
The comparator 72d stores the twentieth-order autocorrelation coefficient value in the memory
72e, compares the data input one after another with the data stored in the memory 72e, and
compares the large coefficient value with the memory. Store in 72e. By repeating this operation
from the 20th order to the 160th order, the largest value in this range is output as the maximum
coefficient value.
[0026]
Next, FIG. 7 is a block diagram showing an internal configuration of the voiced / silence judging
unit 73. As shown in FIG. In FIG. 7, 73a is a constant multiplier that multiplies the zeroth-order
autocorrelation coefficient value input from the pitch detector 72 by a predetermined constant
and outputs the result as a threshold for the presence determination, and 73b is the pitch
detector 72 is a comparator for comparing the maximum count value of the autocorrelation
coefficient input from 72 with the threshold value from the constant multiplier 73a, and 73c is a
changeover switch for sequentially switching the input of the comparison result of the
comparator 73b to the memory 73d 73d is a memory that sequentially stores comparison
results, and 73e is a determination unit that makes a majority decision on the comparison results
and outputs voice presence / absence information.
10-04-2019
8
[0027]
Next, an operation according to the noise / non-speech determiner 73 shown in FIG. 7 will be
described. In order to confirm sound, discrimination is performed by comparing the maximum
peak value (maximum count value) of the obtained autocorrelation coefficient with a value
obtained by multiplying the zeroth-order autocorrelation count value by a predetermined
constant. If voiced, the maximum peak value also appears as a somewhat high value. As an
example, assuming that the constant is 0.7, the 0th-order autocorrelation coefficient is 0f, and
the autocorrelation value of the order in which the maximum value is detected is Nf, noise and
silence are determined by the following equation (2) Be done.
[0028]
Sound: Nf> 0f × 0.7 Silence: Nf ≦ 0f × 0.7 (2)
[0029]
The zeroth-order autocorrelation count value from the pitch detector 72 is multiplied by a
predetermined constant by the constant multiplier 73a, and is output to the comparator 73b as a
threshold for the presence / absence determination.
The comparator 73 b compares the threshold value from the constant multiplier 73 a with the
maximum count value from the pitch detector 72. The comparison result is stored in the memory
73d by sequentially switching the input to the memory 73d whenever the comparison result is
output by the changeover switch 73c. When the comparison result is stored for the number of
memories of the memory 73d, the comparison result is output to the determination unit 73e, the
majority determination of the comparison result is performed by the determination unit 73e, and
is output as the voice presence / absence information. Here, as an example, if it is determined
that the sound is present 7/10 times, the voice presence / absence information is output as voice
information.
[0030]
Next, FIG. 8 is a block diagram showing an internal configuration of the level averaging circuit 8.
10-04-2019
9
As shown in FIG. In FIG. 8, 8a is a changeover switch for sequentially switching the input to the
memory 8b of the digital audio signal input from the analog-to-digital converter 3 via the switch
8a every one sample, and 8b sequentially for every sample of the digital audio signal A memory
8c is an averaging circuit unit for averaging stored numbers.
[0031]
Next, the operation of the level averaging circuit 8 shown in FIG. 8 will be described. The digital
audio signal from the analog-to-digital converter 3 is sequentially stored in the memory 8b via
the changeover switch 8a each time one sample data is input. The stored data is averaged by the
averaging circuit unit 8c and the number of stored data is output as input level information.
[0032]
Here, if the averaging number is 160 and one data is S (t) as an example, the average value Kav is
as shown in equation (1).
[0034]
The controller 6A shown in FIG. 1 uses the result of comparison by giving a threshold to the
average value Kav, and uses it as input level information to grasp the level of the input signal.
In addition, taking the average can suppress the effects of small fluctuations.
[0035]
As described above, according to the present invention, when the voice presence / absence
information from the voice detector indicates a voice signal, the gain of the amplifier is adjusted,
so that the input level at the voice communication terminal is obtained. The effect is very large
because it can be adjusted to the user's voice level and automatically corresponds to the speech
of the conversation according to the usage situation.
10-04-2019
10
Документ
Категория
Без категории
Просмотров
0
Размер файла
20 Кб
Теги
jp2001237920, description
1/--страниц
Пожаловаться на содержимое документа