JP2005253097A - Speech signal transmitting and receiving apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To perform received sound volume control corresponding to a noise level on the basis of speaker's intended volume even if a user moves up or down a manually adjustable sound volume channel in a speech signal transmitting and receiving apparatus having a transmitter and a receiver. <P>SOLUTION: A noise level detection circuit 5 detects a voice sound level entering a transmitting microphone 1 as a noise level when there is no speech input at a transmitter, and a microcomputer 6 stores a correspondence relationship between a noise level that is detected just before a user manually adjusts a speech volume, and the received sound level, and controls received sound volume responsive to the noise level detected by the noise level detection circuit 5 and the correspondence relationship between the stored noise level and the received sound level. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

The present invention relates to an audio signal transmission / reception device, and more particularly to an audio signal transmission / reception device that compresses an audio signal with high efficiency by digital signal processing.
About.

  In recent years, a code encoding linear prediction (CELP) such as Vector Sum Excited Linear Prediction (VSELP) has been used as a speech coding method at a low bit rate, that is, 4.8 to 9.6 Kbps. Has been advocated.

  The technical contents of this VSELP are described in Patent Document 1 “Digital audio coder with improved vector excitation source” and “VECTOR SUM EXCITED LINEAR PREDICITION (VSELP): SPEECH CODING AT 8 KBPS: Ira A. by Motorola, Inc. Gerson and Jasiuk: Paper presented at the Int. Conf. On Acoustics, Speech and Signal Processing -April 1990 ".

  There is a VSELP encoder as a speech coding apparatus by digital signal processing that compresses speech with high efficiency using the VSELP. This VSELP encoder analyzes parameters such as voice frame power, reflection coefficient and linear prediction coefficient, pitch frequency, code book, pitch, and code book gain from the inputted voice signal, and uses the analysis parameters to Is encoded. The VSELP encoder, which is a speech encoding device based on digital signal processing that compresses such speech with high efficiency, is applied to a portable telephone device.

  Since this portable telephone device is often used outdoors, it is often difficult to hear a call due to surrounding background noise. This is because the minimum audible value of the listener increases due to the masking effect due to noise, and the intelligibility and intelligibility of the received speech deteriorate. On the other hand, it is necessary to suppress noise or increase the speaker's voice volume on the transmitting side, increase the reproduction volume on the receiving side, and close acoustic coupling between the speaker and the telephone as a whole. For this reason, the portable telephone device has a switch for manually switching the reception volume according to the surrounding environment.

Japanese National Publication No. 2-502135 Japanese Patent Laid-Open No. 4-82331 JP-A-5-30177

  By the way, as described above, when using the portable telephone device, it is troublesome to manually switch the reception volume according to the surrounding environment. It would be convenient if the reception volume could be switched automatically.

  Here, in the case of a device equipped with a volume control that can be manually adjusted by the user, the reception volume changes by moving the volume control up and down. In this case, however, the reception volume automatically changes according to the surrounding environment. It is desirable to switch.

  The present invention has been made in view of the above circumstances, and even when the volume volume knob that can be manually adjusted by the user is raised or lowered, the received volume can be controlled according to the noise level based on the volume intended by the speaker. An object of the present invention is to provide such an audio signal transmitting / receiving apparatus.

  In order to solve the above-described problem, an audio signal transmitting / receiving apparatus according to the present invention is a sound signal transmitting / receiving apparatus including a transmitting unit and a receiving unit. Noise level detection means for detecting a voice level input to the microphone for transmission when there is no input as a noise level, a noise level detected immediately before the user manually adjusts the volume volume, and a level of the received volume Storage means for storing the correspondence relationship, and control means for controlling the received sound volume according to the noise level detected by the noise level detection means, wherein the control means includes the detected noise level and the storage means. The volume control is performed based on the correspondence relationship between the noise level and the received volume level stored in the above.

  Here, it is preferable that the noise level detection means detects a voice level input to the transmission microphone of the transmission unit immediately after the transmission call power is turned on.

  Further, it is preferable that the noise level detection means detects a sound level input to the transmission microphone every predetermined period in a standby state of the incoming signal of the transmission unit.

  Further, it is preferable that the noise level detection means detects a voice level input to the transmission microphone when a voice level of the receiving unit is a predetermined value or more.

  In the present invention as described above, when the noise level detection means does not input the transmission voice to the transmission unit, the voice level input to the transmission microphone is detected as the noise level, and the user manually adjusts the volume. The correspondence relationship between the noise level detected immediately before and the level of the received sound volume is stored, and the control means is based on the detected noise level and the correspondence relationship between the stored noise level and the received sound volume level, Control the listening volume.

  According to the present invention, the voice level input to the microphone for transmission is detected as the noise level, and the correspondence between the noise level detected immediately before the user manually adjusts the volume and the level of the received volume is stored. Then, by controlling the reception volume based on the detected noise level and the correspondence between the stored noise level and the reception volume level, the user's intended (changed) volume (manually with the volume control knob). The received sound volume can be controlled in accordance with the noise level based on the adjusted sound volume), and the received sound with high intelligibility that is not affected by the influence of background noise can be supplied.

  Hereinafter, preferred embodiments of an audio signal transmitting / receiving apparatus according to the present invention will be described with reference to the drawings.

  FIG. 1 is a block circuit diagram showing a schematic configuration of a mobile phone device according to an embodiment of the present invention.

  In this embodiment, as shown in FIG. 1, a vector sum excited linear prediction (VSELP) encoder 3 that compresses a speech signal with high efficiency by digital signal processing and an analysis obtained by the VSELP encoder 3 A noise interval detection circuit 4 that detects a background noise (hereinafter referred to as noise) interval using a parameter, a noise level detection circuit 5 that detects a noise level of the noise interval detected by the noise interval detection circuit 4, and The microcomputer 6 includes a microcomputer 6 that controls the reception volume according to the noise level detected by the noise level detection circuit 5.

  As a speech encoding method using the VSELP encoder 3, high quality speech transmission at a low bit rate is realized by code book search by analysis by synthesis. Further, in a speech coding apparatus (speech coder) to which a speech coding method using VSELP is applied, a pitch or the like forming the characteristics of an input speech signal is excited by selecting a code vector stored in a code book. The voice is encoded. Parameters such as a pitch frequency used in the encoding include frame power, reflection coefficient and linear prediction coefficient, code book, pitch and code book gain, and the like.

In this embodiment, among these analysis parameters, frame power R ₀ , pitch gain P ₀ indicating the strength of the pitch component, primary linear predictive coding coefficient α _1, and lag LAG relating to the pitch frequency are used for background noise detection. To do. For example, the frame power R ₀ is used because the voice level and the noise level are hardly the same, and the pitch gain P ₀ is used if the ambient noise is almost random. This is because the ambient noise is considered to have almost no pitch.

The primary linear predictive coding coefficient α ₁ is used because it is possible to determine whether α ₁ is large or small and whether the high frequency component or the low frequency component of the frequency is strong. Usually, the background noise is concentrated on the high frequency component of the frequency, and the background noise can be detected from the first-order linear predictive coding coefficient α ₁ . The first-order linear predictive coding coefficient α ₁ is the sum of the coefficients of the Z ⁻¹ term when the direct type higher-order FIR filter is decomposed into a cascade of second-order FIR filters. Therefore, when the zero point is in the range of 0 <θ <π / 2, the first-order linear predictive coding coefficient α ₁ becomes large. Therefore, when this α ₁ is larger than a predetermined threshold, it means a signal in which energy is concentrated in the low band, and when it is smaller than the predetermined threshold, it is a signal in which energy is concentrated in the high band. Become.

  Here, the relationship between θ and frequency will be described.

Assuming that the sampling frequency is f, a frequency of 0 to f / 2 corresponds to 0 to π in a digital system such as a digital filter. For example, if the sampling frequency f is 8 kHz, (0 to 4 kHz) corresponds to (0 to π), and thus π / 2 = 2 kHz. Therefore, the frequency component becomes lower as θ is smaller. Further, as θ decreases, α ₁ increases, and by examining the relationship between α ₁ and a predetermined threshold, it can be determined whether the low frequency component is strong or the high frequency component is strong.

Next, the noise section detection circuit 4 relates to the analysis parameter, that is, the frame power R ₀ , the pitch gain P ₀ indicating the strength of the pitch component, the first-order linear predictive coding coefficient α ₁ and the pitch frequency from the VSELP encoder 3. Receive lag LAG and detect noise interval. This is effective for avoiding increasing the amount of calculation because the size of a digital signal processing (DSP) device and a memory is limited at present when the cellular phone device is miniaturized.

The noise level detection circuit 5 detects the voice level of the noise section detected by the noise section detection circuit 4, that is, the voice level for transmission. Here, the detected transmission voice level may be the value of the frame power R ₀ of the frame that is finally set as the noise section by the determination using the analysis parameter of the noise section detection circuit 4. However, since there is a possibility of a detection error, the frame power _R0 is input to, for example, a 5-tap minimum value filter as described later.

  The microcomputer 6 controls the timing of the noise interval detection in the noise interval detection circuit 4 and the noise level detection in the noise level detection circuit 5, and also controls the volume of the reproduced sound according to the noise level.

  Such a present Example is comprised as a whole so that it may demonstrate below.

  That is, the input voice signal converted into an electric signal by the microphone 1 for transmission is converted into a digital signal by the analog / digital (A / D) converter 2 and supplied to the VSELP encoder 3. The VSELP encoder 3 analyzes an input signal that is a digital signal, compresses information, and performs encoding. At this time, analysis parameters such as the frame power of the input audio signal, the reflection coefficient and the linear prediction coefficient, the pitch frequency, the code book, the pitch, and the gain of the code book are used.

  The data compressed and encoded by the VSELP encoder 3 is supplied to the baseband signal processing circuit 7 and added with a synchronization signal, framing, error correction code and the like. The output data from the baseband signal processing circuit 7 is supplied to the RF transmission / reception circuit 8, modulated to a required frequency, and transmitted from the antenna 9.

Here, among the analysis parameters used by the VSELP encoder 3, as described above, the frame power R ₀ , the pitch gain P ₀ indicating the strength of the pitch component, the first-order linear predictive coding coefficient α _1, and the lag relating to the pitch frequency. LAG is supplied to the noise section detection circuit 4. The noise interval detection circuit 4 detects the noise interval using the frame power R ₀ , the pitch gain P ₀ indicating the strength of the pitch component, the primary linear prediction coding coefficient α ₁ and the lag LAG relating to the pitch frequency. Do. Information (flag information) on a frame finally determined as a noise section by the noise section detection circuit 4 is supplied to the noise level detection circuit 5.

The noise level detection circuit 5 is also supplied with a digital input signal from the A / D converter 2, and detects the signal level of the noise section according to the flag information. The signal level in this case may be the frame power R ₀ as described above.

  The noise level data detected by the noise level detection circuit 5 is supplied to a microcomputer 6 which is a control unit. The microcomputer 6 is also supplied with information from the receiving side level detection circuit 11 as will be described later. By changing the gain of the variable gain amplifier 13 as described later below, the received sound volume can be adjusted. Control.

  The received sound volume is a sound volume when a signal from a call partner transmitted to the mobile phone device of this embodiment is reproduced. The signal from the other party is received by the antenna 9 and supplied to the RF transmission / reception circuit 8.

  The input audio signal from the other party demodulated to baseband by the RF transmission / reception circuit 8 is supplied to the baseband signal processing circuit 7 and subjected to predetermined signal processing. The signal from the baseband signal processing circuit 7 is supplied to the VSELP decoder 10. The VSELP decoder 10 decodes the audio signal based on this information. The decoded audio signal is supplied to a digital / analog (D / A) converter 12 and converted into an analog audio signal.

  The decoded audio signal from the VSELP decoder 10 is also supplied to the reception side level detection circuit 11. The reception side level detection circuit 11 detects the level of the reception side voice and determines whether or not there is a currently received voice (input voice signal from the other party). Detection information from the reception side level detection circuit 11 is supplied to the microcomputer 6.

  The analog audio signal from the D / A converter 12 is supplied to the variable gain amplifier 13. Since the gain of the variable gain amplifier 13 is varied by the microcomputer 6 as described above, the reproduction volume (reception volume) emitted from the speaker 14 is increased by the microcomputer 6 in accordance with noise (background noise). Be controlled.

  Note that a display device 15, a power supply circuit 16 and a keyboard 17 are connected to the microcomputer 6. The display device 15 displays whether the cellular phone device according to this embodiment can be used, what the key switch pressed by the user on the keyboard 17 is, and the like.

  Next, detection of the noise level in the noise level detection circuit 5 constituting this embodiment will be described below.

  First, the section in which the noise level is detected is required to be a noise section detected by the noise section detection circuit 4. The timing for detecting this noise interval is controlled by the microcomputer 6 as described above. The detection of the noise section is to assist the detection of the noise level in the noise level detection circuit 5. That is, it is determined whether the corresponding frame is voiced sound or noise. If it is determined that the frame is noise, the noise level can be detected. Of course, it is clear that more accurate noise level detection should be performed when only noise is present. Therefore, in this embodiment, the noise level detection circuit 5 which is also a transmission voice level detection means detects the voice level input to the transmission microphone 1 when there is no transmission voice input.

  First, for example, −20 dB is set as the initial value of the noise level with respect to the volume level set by the user. When it is determined that the detected noise level is large with respect to the initial set value as will be described later, the reproduction volume level at the receiving unit is increased.

  The noise level is easy to detect as described above if the input speech for each frame is the background noise section. For this reason, in this embodiment, immediately after the transmission power supply of the transmission unit is turned on, it is input when the incoming signal of the transmission unit is in a standby state and during a call and the audio level of the reception unit is equal to or higher than a predetermined value. The voice level is detected as background noise, and the noise level of the frame during this period is detected.

  Here, the fact that the transmission call power supply of the transmission unit is turned on is an intention display for the user to start using the mobile phone device of this embodiment. At this time, this embodiment normally performs a self-diagnosis of each internal circuit, and then enters the standby state after confirming the connection with the base station when the user extends the antenna 9. Since the user's input (input voice) is received only after these series of operations, the user does not utter voice during this time. Therefore, if the voice level is detected using the transmission microphone 1 during this series of operations, the detected voice level is the ambient noise level, that is, the background noise level. Similarly, the background noise level can be detected during or immediately after the user performs the oscillation operation immediately before the start of the call.

  The waiting state for the incoming signal of the transmitting unit is a state in which the power of the receiving unit is turned on and the incoming call signal from the other party is awaited. In this state, of course, it is considered that there is no user's transmitted voice because it is not in a call. Therefore, in this standby state, the background noise level can be detected by measuring the surrounding volume level using the microphone 1 for transmission. This measurement may be performed at an appropriate interval and averaged.

  As described above, the background noise level can be estimated immediately after the transmission call power supply of the transmission unit is turned on and in the standby state of the incoming signal of the transmission unit, and the call can be started by the corresponding voice processing. It is preferable to dynamically follow changes during a call. Therefore, in this embodiment, the background noise level is detected also in accordance with the sound level of the receiving unit during a call.

  The detection of the noise level according to the voice level of the receiving unit during a call is preferably performed after detecting the noise section by the analysis parameter used in the VSELP encoder 3 on the receiving side as described above.

For example, noise can be detected more reliably by monitoring the frame power _R0 and detecting the noise level when the level is above a certain reference level or when the other party is speaking. The playback volume when the other party is speaking can be controlled in real time, and more comfortable call quality can be realized.

  As described above, in this embodiment, immediately after the transmission power of the transmission unit is turned on, the microcomputer 6 is in the standby state of the incoming signal of the transmission unit and when there is no voice of the transmission unit during a call. The detection timing of the noise interval detection circuit 4 and the noise level detection circuit 5 is controlled.

  Next, the noise interval detection operation in the noise interval detection circuit 4 will be described with reference to the flowcharts shown in FIGS.

First, when the flowchart of FIG. 2 is started, in step S1, the frame power R ₀ from the VSELP encoder 3, the pitch gain P ₀ indicating the strength of the pitch component, the primary linear predictive coding coefficient α ₁ and the pitch frequency are related. Receive a lag LAG.

  In the present embodiment, the determination in the following steps using the analysis parameters supplied in step S1 is basically performed in three frames. This is because errors are increased when background noise is discriminated in only one frame. When the noise interval is determined while looking at the range of each parameter over three frames, the noise flag is set to “1”, otherwise, it is set to “0”. The breakdown of the 3 frames is the current frame and the frames up to 1 and 2 frames before.

  Such discrimination based on the analysis parameters through three consecutive frames is performed in the following steps.

First, in step S2, it is determined whether or not the frame power _R0 of the input voice is smaller than a predetermined threshold value _R0th for three consecutive frames. If it is determined YES (R ₀ is smaller than R _0th for three consecutive frames), the process proceeds to step S3. If NO (R ₀ is _{equal to} or greater than R _0th for three consecutive frames), the process proceeds to step S9. This predetermined threshold value R _0th is a value that regards a level higher than that as sound, not as noise. That is, this step S2 is a signal level check.

In step S3, 1 order linear predictive coding of the input speech (LPC) coefficient alpha ₁ is 3 consecutive frames to determine whether a predetermined threshold value alpha _th smaller. If it is determined YES (α ₁ is smaller than α _th for three consecutive frames), the process proceeds to step S4, and if NO (α ₁ is _{equal to} or greater than α _th for three consecutive frames), the process proceeds to step S9. The predetermined threshold α _th is a value that hardly appears when noise is analyzed. That is, this step S3 is a check of the inclination of the voice spectrum.

In step S4, it is determined whether or not the value of the frame power _R0 of the current input voice frame is smaller than "5". If YES (R ₀ is smaller than 5), the process proceeds to step S5. If NO (R ₀ is 5 or more), the process proceeds to step S6. Here, the reason why “5” is used as the threshold value is that the frame when the frame power R ₀ is greater than “5” has a high probability of being a voiced sound.

In step S5, the input speech signal less than 0.9 value for three consecutive frames of the pitch gain P ₀ of, and the current pitch gain P ₀ is determined whether 0.7 or greater. Here, YES (the value of the pitch gain _{P 0} is smaller than 0.9 for three consecutive frames, and the current pitch gain _{P 0} is greater than 0.7) If it is determined that, the process proceeds to step 8, NO (pitch gain If it is determined that the value of P ₀ is 0.9 or more for three consecutive frames and the current pitch gain P ₀ is 0.7 or less, the process proceeds to step S9. Steps S3 to S5 are checks for the strength of the pitch component.

In step S6, in response to the determination result in step S4 (NO: _R0 is 5 or more), it is determined whether or not the frame power _R0 is 5 or more and less than 20. If it is determined YES (R ₀ is 5 or more and less than 20), the process proceeds to step S7. If NO (R ₀ is not 5 or more but less than 20), the process proceeds to step S9.

In step S7, the value of the pitch gain _{P 0} of the input audio signal is smaller than 3 consecutive frames 0.85, and the current pitch gain _{P 0} is determined whether 0.65 or greater. Here, YES (smaller than the pitch gain 0.85 value for three consecutive frames of _{P 0,} and the current pitch gain _{P 0} is 0.65 greater than) If it is determined that, the process proceeds to step 8, NO (pitch gain If it is determined that the value of P ₀ is 0.85 or more for three consecutive frames and the current pitch gain P ₀ is 0.65 or less, the process proceeds to step S9.

  In step S8, the noise flag is set to “1” in response to the determination result of YES in step S5 or step S7. Setting the noise flag to “1” means that the frame is noise.

  In step S9, if the determination in step S2, step S3, step S5, step S6, and step S7 is NO, the noise flag is set to “0” and the corresponding frame is audio. .

  Next, the flowchart of FIG.

  In step S10, it is determined whether or not the pitch lag LAG of the input audio signal is zero. If the determination is YES (LAG is 0), if the LAG representing the pitch frequency is 0, there is almost no probability of being a voice, so that frame is regarded as noise. That is, the process proceeds to step S11 and the noise flag is set to “1”. If it is determined NO (LAG is not 0), the process proceeds to step S12.

In step S12, it is determined whether or not the frame power _R0 is 2 or less. If YES (R ₀ is 2 or less), the process proceeds to step S13. If NO (R ₀ is greater than 2), the process proceeds to step S14. In this step S12, it is determined whether or not the frame power _R0 is quite small. If it is determined as YES, the noise flag is set to “1” in the next step S13, and the frame is regarded as noise.

  In step S13, the noise flag is set to “1” to make the frame noise as in step S11.

In step S14, the previous frame power _R0 is subtracted from the frame power _R0 of the current frame, and it is determined whether or not the absolute value exceeds 3. This is because when the change in the frame power R ₀ between the current frame and the previous frame suddenly increases, that frame is used as an audio frame. That is, if it is determined as YES in this step S14 (change in the frame power _R0 of the current frame and the previous frame has suddenly increased), the process proceeds to step S16, the noise flag is set to “0”, and the frame is set. Let it be an audio frame. If NO (the change in the frame power _R0 between the current frame and the previous frame does not increase rapidly) is determined here, the process proceeds to step S15.

In step S15, the previous frame power _R0 is subtracted from the frame power _R0 of the current frame to determine whether or not the absolute value exceeds 3. This is because when the change in the frame power R ₀ between the current frame and the previous frame suddenly increases, that frame is set as an audio frame. That is, if it is determined as YES in this step S15 (change in the frame power _R0 of the current frame and the frame two immediately before has suddenly increased), the process proceeds to step S16, the noise flag is set to “0”, and the frame is set. Let it be an audio frame. If it is determined NO (the change in the frame power _R0 between the current frame and the previous frame does not increase rapidly), the process proceeds to step S17.

  In step S17, the noise flag is finally determined to be “0” or “1”, and the flag information is supplied to the noise level detection circuit 5.

  As described above, the noise level detection circuit 5 detects the voice level in the noise section according to the flag information obtained by the operation in the noise section detection circuit 4 according to the flowcharts shown in FIGS.

  By the way, in the noise interval detection by the noise interval detection circuit 4, it is impossible to completely distinguish between the voice interval and the noise interval, and it is possible that the voice is erroneously detected as noise. This detection error mostly occurs in the consonant part of speech. If background noise is mixed in at approximately the same level as the consonant part, there is no problem because the reported noise level does not change even if it is detected erroneously. In such a case, since the level differs by 20 to 30 dB depending on the case, it becomes a serious problem. Therefore, in this embodiment, even if erroneous detection is performed, the noise level of the detected noise section is not used as it is, but the influence of erroneous detection is reduced by smoothing or the like.

  The detection of the noise level in which the influence of erroneous detection is reduced by such smoothing means will be described with reference to FIG.

In FIG. 4, a digital input signal from the A / D converter 2 is supplied to the input terminal 20. Further, the input terminal 21 is supplied so that the flag information from the noise section detection circuit 4 is input to the noise level determination unit 5a of the noise level detection circuit 5 constituted by a digital signal processor (DSP). The noise level determination unit 5a is also supplied with the frame power _R0 from the input terminal 22. That is, the noise level determination unit 5a determines the noise level of the input voice signal based on the flag information from the noise section detection circuit 4 or the frame power _R0 . Specifically, in step S17 of the flowchart shown in FIG. 3, the value of the frame power R ₀ when the noise flag is finally set to “1” is regarded as the background noise level.

At this time, since there is a possibility of a detection error, the value of _R0 is input to the 5-tap minimum value filter 5b, for example. This R ₀ is input only when it is recognized as background noise. The output of the minimum value filter 5b is input to a control CPU such as the microcomputer 6 at an appropriate cycle (for example, every 100 msec). Here, when the output of the minimum value filter 5b is not updated, the previous value is repeatedly used. The minimum value filter 5b does not output the middle value in the tap like the median filter described later, but outputs the minimum value. In the case of the same number of taps, it can cope with detection errors of up to four consecutive frames. Moreover, since the minimum value is set to the report level for errors beyond that, the influence can be reduced as much as possible.

In the microcomputer 6, the signal level R ₀ is further input to the 5-tap median filter 6 a in order to further improve the reliability of the input signal level R ₀ . The median filter 6a makes it difficult to report the level even if detection errors continue. In this filtering, the values in the taps of the filter are rearranged in ascending order and the intermediate values are output. The 5-tap median filter does not mistake the reporting level even if a detection error occurs up to two consecutive frames.

  The output signal of the median filter 6a is supplied to the volume position adjusting unit 6b. The volume position adjustment unit 6b varies the gain of the variable gain amplifier 13 based on the output signal of the median filter 6a. In this way, the microcomputer 6 controls the reception volume that is the reproduction volume. Specifically, the increase / decrease of the volume is controlled with the volume position set by the user as the center (base point). Alternatively, the noise level immediately before the user adjusts the volume may be stored, and the output volume may be increased or decreased based on the change in the level and the current background noise level.

  Note that the filter used here may be a smoothing filter such as a primary low-pass filter that smoothes the detected background noise level. Depending on the degree of the low-pass filter, the level difference can be reduced because the follow-up is delayed even if the level is suddenly changed by mistake in detection.

  In this way, even when the noise level is erroneously detected, the influence of the erroneous detection can be reduced.

  Here, a method of controlling the received sound volume based on the detected noise level will be described.

  When controlling the reception volume, normally, the initially set volume is changed according to the background noise as described above. If the user manually changes the volume, the received volume is controlled according to the background noise level based on the volume.

  Specifically, as shown in FIG. 5, for example, the received sound volume level (a, b, c, d, e) corresponding to the noise level in five stages (1 to 5: changing from small to large) is set as an initial value. Given and controlled based on this value.

  For example, when the volume volume knob that can be manually adjusted by the user is raised, the volume level increases. For example, when the detected noise level is 3, the received sound volume level is c before raising the volume volume knob, but the received sound volume level after raising the volume volume knob is d.

  Further, for example, when the volume volume knob that can be manually adjusted by the user is lowered, the volume level is lowered. For example, if the detected noise level is 3, the received volume level is d before the volume volume knob is lowered, but the received volume level after the volume volume knob is lowered is c.

  In other words, if the volume volume knob that can be manually adjusted by the user is raised or lowered, remember the correspondence between the noise level immediately before changing the volume volume knob and the received volume, and when the user changes the volume volume knob, The reference value of the received sound volume is dynamically changed by changing the correspondence (mapping) of the sound volume level. In this way, the received sound volume can be controlled according to the noise level based on the volume (the volume manually adjusted by the volume volume knob) intended (changed) by the speaker.

  Here, an algorithm for controlling the received sound volume when the volume on the receiving side can be internally changed in 2 dB steps will be described.

  For the volume on the receiver side, the variable range of automatic volume adjustment according to the noise level is set to 5 levels, and the volume value corresponding to these levels is set to a 6 dB step. Variables in which volume values set corresponding to each stage are stored are lvl [0] to lvl [4], and their value ranges are 0 to 12. That is, it is considered that 1 of the variable value corresponds to 2 dB.

  The initial values of the variables are stored in the nonvolatile RAM, for example, as lvl [0] = 0, lvl [1] = 3, lvl [2] = 6, lvl [3] = 9, and lvl [4] = 12. deep. These variable values correspond to +0 dB, +6 dB, +12 dB, +18 dB, and +24 dB, respectively, as actual volume levels. LVnow is the current volume value, and LVafter is the volume value to be changed after reading the noise level. The noise levels corresponding to the lvl [0], lvl [1], lvl [2], lvl [3], and lvl [4] are, for example, 0 to 5, 6 to 8, 9 to 15, and 16 to 45, respectively. 46-. This noise level corresponds to 1/16 of the noise level read by the noise level detection circuit 5 of FIG. 1, and varies depending on the gain of the microphone.

  Here, FIG. 6 is a flowchart showing an algorithm of the received sound volume control. The received sound volume control operation shown in FIG. 6 is executed in response to an interruption every 100 ms, for example.

  First, in the first step S21, it is determined whether or not there is a volume change by the user. If YES, that is, if there is a volume change, the process proceeds to step S22 to determine whether or not the operation is a volume up operation. If YES is determined, that is, if a volume up operation is performed, the process proceeds to step S23, and after lvl [i] = lvl [i] +3, i.e., 6 dB up with respect to i = 0-4, Return from return, that is, interrupt. If NO is determined in step S22, that is, if a volume down operation is performed, the process proceeds to step S24, and for i = 0 to 4, lvl [i] = lvl [i] -3, that is, 6 dB. After going down, return.

  If NO in step S21, that is, if it is determined that there is no volume change by the user, the process proceeds to step S25, where the microcomputer 6 reads the noise level detected by the noise level detection circuit 5 and is 1/16 times. After this is set to the noise level NL, the process proceeds to step S26.

  In step S26, when the noise level NL is 5 or less (NL ≦ 5), the volume value LVafter to be changed is set to lvl [0] (LVafter = lvl [0]), and otherwise NL ≦ 8. LVafter = lvl [1], otherwise LVafter = lvl [2] when NL ≦ 15, otherwise LVafter = lvl [3] when NL ≦ 45, otherwise LVafter = lvl [4] And Here, each comparison value with the noise level NL varies depending on the gain of the microphone for transmission.

In the next step S27, when the LVafter is larger than the upper limit value UP _lim , for example, UP _lim = 12 (LVafter> UP _lim ), the LVafter is limited to LVafter = UP _lim . Further, in the next step S28, when the LVafter is smaller than the lower limit value DWN _lim , for example, DWN _lim = 0 (LVafter <DWN _lim ), the LVafter is limited to LVafter = DWN _lim .

In the next step S29, the current volume value LVnow is smaller than the volume value LVafter should change the (LVnow <LVafter) time, is increased by a unit step _{V step} volume change _{LVnow (LVnow = LVnow + V step} ) when LVnow is greater than LVafter (LVnow> LVafter), it reduces the LVnow only _{V step} and (LVnow = LVnow -V _step). Here, the unit step V _step corresponds to 1 as described above, that is, 2 dB.

  In the next step S30, it is determined whether or not LVnow ≠ LVafter. If NO, that is, LVnow = LVafter, the process returns from the return, that is, the interrupt. If YES, ie, LVnow ≠ LVafter, the volume value is set to the value of LVnow and then the process returns.

  By such an incoming volume control operation, volume adjustment by the user and automatic volume control according to the noise level are effectively performed.

  Next, in order to confirm the effectiveness of the present embodiment described above, an example of actual background noise detection by simulation will be described.

  In general, what is represented by the Hot spectrum as a standard for indoor noise is generally used, but it is difficult to apply the Hot spectrum to a mobile phone device that is often used outdoors. Therefore, the noise actually recorded outdoors was used for the simulation. This noise was recorded at the premises of two stations (referred to as stations A and B). Then, when the voice and noise are added as a digital waveform on the computer, the noise is played in the listening room, and the voice when talking through the microphone using the mobile phone device is recorded in that state. Three types were examined. The noise level was assumed to be about 70 dBspl.

  As this simulation, a simulation with fixed decimal points was performed, and the frequency of detection, errors, and detected noise level were investigated.

  Examples of detecting background noise are shown in FIGS. FIG. 7 to FIG. 10 show the results of the voice and the detected background noise when talking using the mobile phone device while flowing the background noise recorded in the station A or B station as a sample.

  Fig. 7 shows the result when talking in a male voice, "Human seeks rich nature" while playing background noise recorded in station A. Fig. 8 shows background noise recorded in station A. This is the result of a female voice saying “Please do not overdo it for health”. Fig. 9 shows the result of a man's voice saying "Human wants rich nature" while playing background noise recorded in the station B. Fig. 10 shows the background noise recorded in the station B. This is the result of a female voice saying “Please do not overdo it for your health”.

  In each detection result, a rectangular portion in the figure is a section in which a portion that seems to be background noise is detected. Although the voice part and the noise part cannot be completely separated, the detection can be performed in units of several tens of ms, and the voice part is hardly erroneously detected. The background noise detection error in the consonant part can be avoided by using the means such as the smoothing described above. In particular, the use of minimum filtering can avoid errors in level reporting due to detection errors.

  Such noise detection simulation is not limited to the above-mentioned simulation with fixed decimal points, but may be performed with floating decimal points on a workstation, for example, and the detection results obtained are almost the same.

  As described above, since the cellular phone device according to the present embodiment performs the noise interval detection using the analysis parameter used in the VSELP encoder, the background noise can be detected with high accuracy and high reliability with a small amount of calculation. Since the playback volume is controlled according to the background noise, it is possible to provide a reception sound with high intelligibility.

  In summary, the audio signal transmitting / receiving apparatus according to the embodiment of the present invention includes an audio signal transmitting / receiving apparatus having an audio signal transmitting encoding circuit that compresses an audio signal with high efficiency by digital signal processing. A noise interval detecting means for detecting a noise interval using an analysis parameter obtained by the transmission encoding circuit, a noise level detecting means for detecting a noise level of the noise interval detected by the noise interval detecting means, and the noise level Control means for controlling the received sound volume in accordance with the noise level detected by the detecting means.

  Therefore, since the noise interval detection means detects the noise interval using the analysis parameter obtained by the encoding circuit for transmitting the audio signal, the background noise can be detected with high accuracy and high reliability with a small amount of calculation, and the noise level. The detection means detects the noise level of the noise section, and the control unit controls the reception volume according to the noise level, so that the detection is highly reliable, and the reception volume of the reception volume is determined based on the detected background noise level information. Control and the like can be performed easily and reliably, and a reception sound with high intelligibility can be supplied.

  Further, the embodiment of the audio signal transmitting / receiving apparatus according to the present invention detects and controls the audio level input to the transmitting microphone as the noise level when the noise level detecting means does not input the transmitting audio to the transmitting unit. Since the means controls the reception volume according to the detected sound level, it is possible to supply a reception sound with high intelligibility that is not affected by the influence of background noise.

  Note that the audio signal transmitting / receiving apparatus according to the present invention is not limited to the above-described embodiment. For example, only one analysis parameter can be used to detect a noise interval. Furthermore, it is possible to detect only one frame instead of considering a plurality of consecutive frames. However, in these cases, it cannot be denied that the accuracy is lower than in the present embodiment. Furthermore, it goes without saying that the flow of noise interval detection is not limited to that shown in the flowchart.

It is a block circuit diagram for demonstrating the circuit structure of the Example of the audio | voice transmission / reception apparatus which concerns on this invention. 2 is a flowchart for explaining the operation of the background noise detection circuit of the embodiment shown in FIG. 2 is a flowchart for explaining the operation of the background noise detection circuit of the embodiment shown in FIG. It is a figure for demonstrating the means for preventing a background noise level from the influence of an error. It is a figure for demonstrating the specific example of the reception volume control by the detected noise level in a present Example. It is a flowchart for demonstrating reception volume control operation | movement. It is a figure which shows the background noise detection result obtained by performing the simulation by a fixed decimal point. (When speaking in a male voice with noise in station A) It is a figure which shows the background noise detection result obtained by performing the simulation by a fixed decimal point. (When talking in a female voice with noise in station A) It is a figure which shows the background noise detection result obtained by performing the simulation by a fixed decimal point. (When talking in a male voice with noise in the station B) It is a figure which shows the background noise detection result obtained by performing the simulation by a fixed decimal point. (When talking to a woman with noise in the station B)

Explanation of symbols

  1 microphone for transmission, 2 analog / digital (A / D) converter, 3 VSELP encoder, 4 noise interval detection circuit, 5 noise level detection circuit, 6 microcomputer, 7 baseband signal processing circuit, 8 RF transmission / reception circuit, 9 antenna, 10 VSELP decoder, 11 receiving side level detection circuit, 12 digital / analog (D / A) converter, 13 variable gain amplifier, 14 speaker

Claims (4)

In an audio signal transmitting / receiving apparatus including a transmission unit and a reception unit,
Volume adjustment volume for manually adjusting the reception volume,
A noise level detection means for detecting a voice level input to the microphone for transmission when there is no transmission voice input in the transmission unit;
Storage means for storing a correspondence relationship between a noise level detected immediately before the user manually adjusts the volume and the level of the received volume;
Control means for controlling the reception volume according to the noise level detected by the noise level detection means,
The audio signal transmitting / receiving apparatus characterized in that the control means controls the sound volume based on the detected noise level and the correspondence relationship between the noise level stored in the storage means and the level of the received sound volume.   2. The audio signal transmitting / receiving apparatus according to claim 1, wherein the noise level detecting unit detects an audio level input to the transmitting microphone of the transmitting unit immediately after the transmission call power is turned on.   2. The audio signal transmitting / receiving apparatus according to claim 1, wherein the noise level detecting means detects an audio level input to the transmission microphone every predetermined period in a standby state of an incoming signal of the transmission unit.   2. The audio signal transmitting / receiving apparatus according to claim 1, wherein the noise level detecting means detects an audio level input to the transmitting microphone when an audio level of the receiving unit is equal to or higher than a predetermined value.

Images