JPH04283471A - Silent processor - Google Patents

Abstract

PURPOSE:To operate an optimal processing according to a situation by comparing the number of voice data obtained by adding an additional value to a reference value with reference voice data, and using the sum of the additional number and the minimum value as a threshold value. CONSTITUTION:A D/A converting processing is operated by an audio block 2, and the compressing and extending processing of voice information, and silence detection are operated by an ADPCM block 10. The compressed and extended voice information are stored in a buffer 11 and a host memory 12. The threshold value is set at the time of receiving a voice signal, and used as the reference of the judgement of the sound and silence of sampling data. Then, the voice data are stored through the memory 12 in an RAM 13. Then, a CPU 14 reads out the voice data from the RAM 13, and retrieves the minimum data from among the voice data having sound. Moreover, the CPU 14 calculates the number of the added voice data obtained by adding a prescribed number to the minimum value, and compared it with the minimum value. A control part uses the sum of the minimum value and the additional value as the threshold value from this compared result. Thus, the optimal silent processing can be attained according to the situation.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for encoding and recording digitized audio information, and more particularly to a silence processing apparatus.

0002

2. Description of the Related Art When audio information is encoded and recorded on a recording medium such as a CD (compact disc)-i or CD-ROM, conventionally, information in a silent state that does not include the necessary audio information is also encoded. It was recorded on the above recording medium. Therefore, there is a problem in that the recording capacity of the recording medium is consumed even by information in a silent state, and the storage capacity is not used efficiently.

[0003] In order to solve these problems, the applicant of the present application has proposed a silence processing device as shown below. As shown in FIG. 7, audio information converted into an electrical signal by a microphone or the like is supplied to an audio block 2 via an input terminal 1, where it is digitally converted and sent to an encoder 3. . Noise is reduced by reducing the amount of information in the digitally converted audio signal and performing noise shaping.The quantizer 6 rounds the audio signal, and the output signal from the encoder 3 is sent to the recording medium. It is sent to the management circuit 7. Further, the output signal of the audio block 2 is sent to the utterance/non-sound determination section 5 that constitutes the encoder 3, and the utterance/non-sound determination section 5 monitors the output signal of the audio block 2 in units of one sound unit. If there is at least one output signal having an amplitude value exceeding the threshold value in a sound unit, that sound unit is determined to be a sound unit, and a signal to that effect is sent to the system control section 4. Note that if there is no output signal having an amplitude value exceeding the threshold within one sound unit, that sound unit is determined to be a silent sound unit, and a signal to that effect is sent to the system control section 4. In addition, the above sound unit is
It is a unit representing a set of input audio information sampled a predetermined number of times in order to encode an audio signal, and for example, information corresponding to 28 samplings becomes information of one sound unit. Further, the collective unit of the sound units is called a sector, and one sector is made up of N sound units, and the number of silent sound units in one sector is expressed by the silence rate. For example, if M silent sound units are counted in one sector, the silent rate is M/N. The system control unit 4 counts silent sound units based on the supplied sound unit signals, calculates a silence rate from the counted value, and compares the calculation result with a predetermined value α. If the silence rate value is equal to or greater than the predetermined value α, the system control unit 4 determines that the sector is a silent sector and sends a signal to that effect to the recording medium management circuit 7. Note that if the silent rate is less than or equal to the predetermined value α, the system control unit 4 determines that the sector is a sound sector, and sends a signal to that effect to the recording medium management circuit 7.
Send to. The recording medium management circuit 7 is connected to the system control unit 4
If a signal indicating silence is supplied, the encoder 3
The encoded data sent by the system is not recorded on the disk as a recording medium, but only a flag indicating silence is recorded in the header information of the disk or in a sector information storage section installed separately from the disk.On the other hand, the system control When a signal indicating that there is a sound is supplied from the section 4, the encoded data sent from the encoder 3 and the sound presence flag are recorded on the disk or the sector information storage section. Also, as described above, C
When reproducing audio information recorded on a D-ROM or the like, the recording medium management circuit 7 stores information for each sector in advance in a sector information storage section in which a signal indicating whether the sector is sound or silent is stored. The audio information corresponding to the accessed and audio sector information is read from the disk. As described above, according to the silence processing device of this embodiment, when the signal sent from the microphone is in a silent state, that information is not recorded on the recording medium, so that the recording capacity of the recording medium can be used efficiently. can. In the above embodiments, a CD-ROM or a CD-i is used as the recording medium, but the present invention is not limited to this, and a medium for recording signals processed as digital signals, such as a semiconductor memory such as a RAM, may also be used. It's okay to have one.

0004

[Problems to be Solved by the Invention] However, the above-mentioned threshold value, which is the criterion for determining whether the sound unit is uttered or not, is fixed in the conventional silence processing device. was the value given. However, if the above threshold value is not set to an appropriate value in a system to which this silence processing device is applied, such as recording a telephone line or a normal conversation, the sound information may actually be recorded. There is a sleet that is considered silent. Furthermore, even in the same system, it is necessary to change the threshold value depending on the situation at that time. For example, in a telephone line, appropriate silence processing cannot be performed unless the threshold value is appropriately changed depending on whether the noise around the caller is loud or low.

[0005] Furthermore, when reproducing audio information recorded on a CD-ROM or the like as described above, it is necessary to access the sector information storage section, but at this time, only sector information representing silence is continuous. If the sector information is recorded as such, sector information indicating silence will have to be read out continuously, resulting in a problem that the playback time will become longer. The present invention takes these points into consideration, and an object of the present invention is to provide a silence processing device that can perform appropriate silence processing depending on the situation.

[0006]

[Means for Solving the Problems and Their Effects] The present invention provides a method for determining the presence or absence of sound in audio data consisting of digital signals based on a threshold value, and recording only the audio data that is determined to be sound. In the processing device, a storage unit that stores audio data supplied for a certain period of time from when it is determined that there is a sound;
The number of reference audio data, which is the number of audio data having a reference value, is counted from among the audio data stored in the storage unit, and the number of audio data, which is the number of audio data having an augend value obtained by adding an additive value to the reference value, is counted. Count the number of audio data to be added, and when the comparison result of the counted value between the reference number of audio data and the number of audio data to be added is greater than or equal to a predetermined value, add the reference value and the additional value. and a control unit that sets the value as the threshold value.

[0007] With this configuration, the control section counts, for example, the number of audio data having the minimum value from the audio data group stored in the storage section as the reference audio data number. Further, the control unit counts the number of audio data having a value obtained by adding 1 to the minimum value as a first number from the stored audio data group. For example, if 1/2 of the reference number of audio data is larger than the first number, the calculation is terminated, and if it is smaller, the number of audio data having the value obtained by adding 2 to the minimum value is calculated as the number of audio data in the stored audio. It is counted as the second number from the data group. Then, the control unit determines whether 1/2 of the first number is larger than the second number, and the calculation proceeds in the same manner thereafter. Then, when the calculation is completed, the control section sets the addition result value of the addition number added to the minimum value and the minimum value as a threshold value. In this way, the control section operates so that an appropriate threshold value can be set.

Furthermore, the present invention provides a silence processing device that determines the presence/absence of audio data in the form of a digital signal based on a preset threshold and records only the audio data determined to be audio. A storage section that stores audio data supplied for a certain period of time from the time when it is determined to be a sound, and a minimum value among the values of the audio data stored in the storage section, and a minimum value that is set in advance as the minimum value. The present invention is characterized by comprising: a control unit that changes the input level of input audio information that calculates the difference between the input audio information and the threshold value and performs silence processing according to the difference.

With this configuration, the control unit adjusts the signal of the input audio information so that the relative relationship between the signal level of the input audio information and the preset threshold signal level is appropriate. It acts to change the level appropriately.

Furthermore, the present invention determines the utterance or silence of digitally converted audio information, counts the silence information, and calculates the number of silences in the minimum unit based on the result of counting the number of silences in the minimum unit of reading and writing audio information to a recording medium. At the same time, a signal indicating whether there is a sound or no sound is sent to the sound/silence signal storage section, and while a signal indicating no sound is being sent, no audio signal is recorded on the recording medium. In the silence processing device that reproduces the audio signal recorded on the recording medium based on the sound/silence signal stored in the sound/silence signal storage section, when a plurality of signals indicating silence are continuously supplied, one The present invention is characterized in that it includes a voice/silence signal storage section that stores only signals indicating silence.

[0011] With this configuration, the number of signals indicating silence stored in the utterance/silence signal storage section is reduced, and the time required to read out the signals is shortened.

0012

[Embodiment] FIG. 1 shows an embodiment of the silence processing device of the present invention.
, anti-aliasing filter processing, smoothing filter processing, A/D conversion processing of audio information detected by a microphone etc., and D/A conversion of audio information made of digital signals read from CD-i etc. The audio block 2 that performs processing is an ADPC that performs compression and decompression processing of audio information in the form of digital signals, and silence detection.
Connected to M block 10. ADPCM block 10
is a buffer 11 that stores compressed and expanded audio information.
, and the host memory 12. host memory 1
2 is a memory for temporarily storing sector information, etc., which will be described later, and is connected to a RAM (random access memory) 13, and the host memory 12 is connected to a buffer 11 and a RAM 1.
It also performs DMAC (direct memory access control) operations between 3 memory cells.

The ADPCM block 10, host memory 12 and RAM 13 described above are connected to a CPU (central processing unit) 14 via an address bus and a data bus, respectively.
Further, the address bus and data bus are connected to a write circuit (not shown) for writing audio information to, for example, a CD-i.

The operation of the silence processing apparatus constructed as described above will be explained below. In the silence processing device of this embodiment, a threshold value, which is a criterion for determining the presence or absence of each sampling data constituting a sound unit, is first set after receiving audio information. The setting of this threshold value will be explained below. The audio block 2 performs the above-mentioned filter processing on audio information input via a microphone, etc., and then converts the audio data of the A/D-converted digital signal into an ADP.
It is sent to the CM block 10. Note that all bit data of audio data sent from audio block 2 is 0.
All audio data sent by the audio block 2 for about 1 to 2 seconds from the time when so-called sound other than so-called silence is detected is stored in the buffer 11, regardless of whether it is silent data or sound data.

Next, the audio data stored in the buffer 11 is transferred to the RA via the DMAC in the host memory 12.
After being stored in the M13, the CPU 14 reads out the audio data stored in the RAM 13 via the data bus, searches for the audio data with the smallest value from among the audio data that has sound, and searches The value, ie the minimum value, is stored in host memory 12. Incidentally, for the sake of explanation, the above minimum value is assumed to be "a". In addition, in step 1 shown in FIG. 2 (indicated by S in the figure), the above “α” corresponds to “α”.
a”.

Next, the CPU 14 reads the audio data from the RAM 13, searches for audio data having the same value as the above-mentioned minimum value "a" stored in the host memory 12 from among the read audio data, and searches the retrieved audio data. The number of data is counted and the counted value is stored in the host memory 12. Note that this count value for purposes of explanation will be referred to as "A". Also, step 1 above
In this case, the above "A" corresponds to "X".

Next, the CPU 14 reads the minimum value "a" from the host memory 12 and adds 1 to the minimum value "a", and then reads the audio data from the RAM 13.
The voice data having the value "a+1" which is the value of the minimum value +1 is searched, the number of searched voice data is counted, and the counted value is stored in the host memory 12. For the sake of explanation, this count value will be referred to as "B". Further, in step 2 shown in FIG. 2, the above-mentioned "B" corresponds to "Y".

Next, as shown in step 3, the CPU 14
is the count value “A” and the count value “B” from the host memory 12.
'' is read out, and it is determined whether the count value "B" is less than or equal to 1/2 of the count value "A", that is, the count value "A/2" ≧ the count value "B".

As a result of the judgment, the count value "B" is equal to the count value "A".
”, the threshold setting operation is finished,
The sum of the above minimum value “a” and 1 added to the minimum value is “
a+1" is set as the threshold value.

As a result of the judgment, the count value “B” is the count value “A”.
'', the CPU 14 executes the above (
a+1) is newly set to a. That is, generally, as shown in step 4, the value (α+1) is set as the new value of α. Then, return to step 1 again, and repeat steps 1 to 4 in the same manner. To explain in detail, as shown in step 2, add 1 to the value of "a+1" again,
Audio data having a value of "(a+1)+1", ie, "a+2" is retrieved from the RAM 13, and the number of retrieved audio data is counted. For the sake of explanation, this count value will be referred to as "C". Then, the CPU 14 determines that the count value "C" is the above count value "B".
If this condition is satisfied, a value obtained by adding an additional value to the minimum value is set as the threshold value. In this case, the minimum value is "a" and the added value is 2 since 1 is added twice, so the threshold value is "a+2".

[0021] If the count value "C" is still not less than B/2, the CPU 14 further adds 1 to the minimum value "a", counts the number of audio data having the value "a+3", and calculates this value. The same operation as described above for the count value is performed. Thereafter, the calculation is repeated in the same manner to set the threshold value.

[0022] By setting the threshold in this way, it is possible to eliminate
Noise at a certain level is determined to be silent, and only audio information exceeding the certain level can be recorded as sound. For example, when recording a telephone conversation, by using the above-mentioned silencing method, it is possible to silence the noise that originally exists in the telephone line, making it possible to perform more effective silencing processing.

In the above description, the threshold level is changed, but the level of the input audio signal may be changed instead of changing the threshold level. That is, the threshold level is set in advance to a certain value, which is the e-level value for purposes of explanation. Similar to the method described above, the CPU 14 detects the minimum value "a" from among the voice data stored in the RAM 13. And the CPU 14 is
For example, find the difference between these values by subtracting the minimum value a from the threshold level e, and if the difference is positive, that is, if the threshold value is greater than the minimum value, input The signal level of all the audio information is increased by the voltage corresponding to the above difference. On the other hand, if the difference is negative, that is, when the above threshold is smaller than the above minimum value, all the signal levels of the input audio information are increased by the voltage corresponding to the above difference. The signal level is lowered by a voltage corresponding to the above difference.

[0024] By adopting such a method, the relative relationship between the input audio signal level and the threshold level can be appropriately set, so that more effective silencing processing can be performed.

[0025] By the above method, the threshold value is set to ADPC.
The ADPCM block 10 then determines the presence or absence of sound for each sound unit in the same manner as in the conventional example, and divides the audio data supplied from the audio block 2 into a set of 144 sound units. The number of sound units determined to be silent in one sector is counted, and the counting result is sent to the CPU 14. Incidentally, the timing at which the counting result is sent to the CPU 14 is after the encoding in the ADPCM block 10 for one sector is completed.

[0026] The CPU 14 compares the count value set in advance in the CPU 14 with the above-mentioned number of silent sound units supplied from the ADPCM block 10, and if the number of silent sound units is larger than the set value, is determined to be a silent sector and sends data of “1” to ADPCM.
If the number of silent sound units is smaller than the set value, it is determined that the sector is a sound sector, and data of "0" is sent to the ADPCM block 10 and host memory 12.

[0027] The ADPCM block 10 includes a silent sector,
The audio data encoded by the ADPCM block 10 is sent to the buffer 11 regardless of the sound sector, and the audio data stored in the buffer 11 is transferred to the RAM 1.
The CPU 14 sends out whether or not to store the sound in 3.
Judgment is made based on silent sector information. In addition to sending out the sound/non-sound sector information, the CPU 14 also sends page information indicating in which page of the buffer 11 the audio data encoded in the ADPCM block 10 is stored to the host memory 12.

That is, when sector information of "1" indicating a silent sector is sent from the CPU 14 as shown in FIG. 3, it is stored in the page of the buffer 11 designated by the CPU 14 as shown in step 2 of FIG. The current audio data is not written into the RAM 13 and the process moves to step 4 in FIG. In step 4, sector information, which is "1" in the above example, is stored in the host memory 12. on the other hand,
If the CPU 14 sends sector information of "0" indicating that it is a sound sector in step 2, then in step 3 the audio data stored in the page of the buffer 11 specified by the CPU 14 as shown in FIG. RAM1 of
3, and in step 4, the sector information, which is "0" in the above example, is also stored in the host memory 12.

Since the audio data encoded in the ADPCM block 10 corresponding to the sector determined to be a silent sector in this way is not stored in the RAM 13, the data is compressed accordingly.

[0030] Furthermore, as mentioned above, the host memory 12
When sector information is supplied from the PU 14 and the sector information is stored, if data of "1" indicating silence is stored continuously as shown in FIG. 1'' data may be compressed into one piece of data and stored in the new sector information storage area in the host memory 12. By operating in this way, the number of sector information can be reduced.

As mentioned above, the data is compressed and stored in the RAM.
A case in which the audio data stored in the host memory 13 is reproduced based on the sector information stored in the host memory 12 and the encoded audio data in the ADPCM block will be described below with reference to FIG. The CPU 14 sends page information indicating to which page of the buffer 11 the audio data is sent to the RAM 13. The host memory 12 first reads sector information in step 1, and if the content is "1" indicating a silent sector as shown in steps 2 and 3, the data of all "0" in one sector is sent to the CPU 14.
Write to the page of the buffer 11 specified by . On the other hand, if the content of the sector information is "0" indicating a sound sector, the RAM 13 in which encoded audio data in the ADPCM block of that sector number is stored is accessed, and the page in the buffer 11 specified by the CPU 14 is accessed. Store the read audio data to. Note that even when new sector information is stored in the new sector information storage area as shown in FIG. 5, the host memory 12 sends audio data to the buffer 11 based on the new sector information.

In this way, the ADPCM block 10 accesses the buffer 11 every sampling period, and as shown in step 4 of FIG.
All you have to do is decode it. There is no need for the ADPCM block 10 to be involved in the reproduction of silent sectors.

[0033] The 8-bit audio data ADPCM-decoded in the ADPCM block 10 is sent to the audio block 2 in step 5, and is subjected to D/A conversion and filter processing in the audio block 2 in step 6. This becomes audio information.

In this way, in reproducing a silent sector, “
The ADPCM block 10 does not give any consideration to silent sector reproduction because the information is processed in the host memory 12 based on "sector information" and "voice sector data" obtained by "silent sector detection" and "silent sector compression". In addition, by compressing silent sector information using new sector information as shown in Figure 5, it is possible to cut out silent parts, such as in conversations, so information can be decoded faster. Obtainable.

[0035]

As described in detail above, according to the present invention, by changing the threshold value or the signal level of the input audio information, which is the standard for determining whether there is a sound or no sound in the input audio information, Appropriate silence processing can be performed even when there is a lot of ambient noise. Further, according to the present invention, when silence information is continuously supplied, the voice/silence information storage section stores only one piece of silence information, so that when reproducing a recording medium, the time required to reproduce the silence portion is reduced. It can save time.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an embodiment of a silence processing device of the present invention.

FIG. 2 is a flowchart for explaining a threshold setting operation in the silence processing device shown in FIG. 1;

3 is a flowchart for explaining a silence compression operation in the silence processing device shown in FIG. 1. FIG.

4 is a diagram for explaining a silence compression operation in the silence processing device shown in FIG. 1. FIG.

5 is a diagram for explaining the operation of compressing sector information in the silence processing device shown in FIG. 1. FIG.

6 is a flowchart for explaining the reproduction operation of audio data processed by the silence processing device shown in FIG. 1. FIG.

FIG. 7 is a block diagram showing the configuration of a conventional silence processing device.

[Explanation of symbols]

2...Audio block, 10...ADPCM block,
11...Buffer, 12...Host memory, 13...RAM,
14...CPU.

Claims (3)

[Claims] Claim 1: When a sound is determined to be present in a silence processing device that determines the presence or absence of sound in digital signal audio data based on a threshold value and records only the audio data that is determined to be sound. A storage section for storing audio data supplied from the storage section for a certain period of time, and a reference audio data number, which is the number of audio data having a reference value, from among the audio data stored in the storage section, and is added to the reference value. The number of audio data to be added, which is the number of audio data having the augend value obtained by adding the values, is counted, and the comparison result of the counted value of the reference audio data number and the number of audio data to be added is equal to or greater than a predetermined value or a predetermined value. A silence processing device comprising: a control unit that sets the threshold value to a value obtained by adding the reference value and the additional value when the threshold value is equal to or less than the threshold value. Claim 2: A silence processing device that determines the presence or absence of voice data in digital signals based on a preset threshold and records only the voice data that is determined to be voiced. A storage unit that stores audio data supplied for a certain period of time from the time when the storage unit detects the minimum value from among the values of the audio data stored in the storage unit, and detects the minimum value from among the values that the audio data stored in the storage unit has, 1. A silence processing device comprising: a control unit that changes the input level of input audio information that calculates a difference between the input audio information and the input audio information and performs silence processing according to the difference. Claim 3: Determine the presence or absence of utterances in the digitally converted audio information, count the silence information, and calculate the number of silences for each minimum unit based on the result of counting the number of silences in the minimum unit of reading and writing audio information to a recording medium. A signal indicating sound or silence is sent to the sound/silence signal storage section, and while the signal indicating silence is being sent, no audio signal is recorded on the recording medium; In the silence processing device that reproduces the audio signal recorded on the recording medium based on the sound/silence signal stored in the storage unit, when a plurality of signals indicating silence are continuously supplied, one silence is indicated. A silence processing device comprising a voice/silence signal storage section that stores only signals.