WO2018040430A1 - 一种回声时延的确定方法、装置及智能会议设备 - Google Patents

一种回声时延的确定方法、装置及智能会议设备 Download PDF

Info

Publication number
WO2018040430A1
WO2018040430A1 PCT/CN2016/113262 CN2016113262W WO2018040430A1 WO 2018040430 A1 WO2018040430 A1 WO 2018040430A1 CN 2016113262 W CN2016113262 W CN 2016113262W WO 2018040430 A1 WO2018040430 A1 WO 2018040430A1
Authority
WO
WIPO (PCT)
Prior art keywords
signal data
audio
determining
time information
input
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2016/113262
Other languages
English (en)
French (fr)
Inventor
刘荣
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangzhou Shiyuan Electronics Thecnology Co Ltd
Original Assignee
Guangzhou Shiyuan Electronics Thecnology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Guangzhou Shiyuan Electronics Thecnology Co Ltd filed Critical Guangzhou Shiyuan Electronics Thecnology Co Ltd
Publication of WO2018040430A1 publication Critical patent/WO2018040430A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04M—TELEPHONIC COMMUNICATION
    • H04M9/00—Arrangements for interconnection not involving centralised switching
    • H04M9/08—Two-way loud-speaking telephone systems with means for conditioning the signal, e.g. for suppressing echoes for one or both directions of traffic
    • H04M9/085—Two-way loud-speaking telephone systems with means for conditioning the signal, e.g. for suppressing echoes for one or both directions of traffic using digital techniques
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04M—TELEPHONIC COMMUNICATION
    • H04M9/00—Arrangements for interconnection not involving centralised switching
    • H04M9/08—Two-way loud-speaking telephone systems with means for conditioning the signal, e.g. for suppressing echoes for one or both directions of traffic
    • H04M9/082—Two-way loud-speaking telephone systems with means for conditioning the signal, e.g. for suppressing echoes for one or both directions of traffic using echo cancellers
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04M—TELEPHONIC COMMUNICATION
    • H04M1/00—Substation equipment, e.g. for use by subscribers
    • H04M1/02—Constructional features of telephone sets
    • H04M1/20—Arrangements for preventing acoustic feed-back
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04M—TELEPHONIC COMMUNICATION
    • H04M9/00—Arrangements for interconnection not involving centralised switching
    • H04M9/08—Two-way loud-speaking telephone systems with means for conditioning the signal, e.g. for suppressing echoes for one or both directions of traffic

Definitions

  • the embodiments of the present invention relate to the field of sound signal processing technologies, and in particular, to a method, an apparatus, and a smart conference device for determining an echo delay.
  • the echo means that the sound played by the speaker of the electronic device is again picked up by the microphone of the electronic device and sent back to other remote electronic devices having a call function, so that the far-end speaker can hear himself. the sound of.
  • echo cancellation techniques or Automatic Echo Cancellation (AEC) are needed to eliminate echo.
  • the basic principle of canceling echo based on AEC technology can be described as: obtaining a reference signal of a sound currently emitted by a speaker from a voice system, sampling a microphone input signal picked up by the microphone, and then using the reference signal as an adaptive
  • the input value of the filter is processed by the adaptive filter, and the output value of the filter simulates the echo signal in the sampled microphone input signal, and then the output from the adaptive filter is subtracted from the sampled microphone input signal.
  • the echo signal thus completing the echo cancellation, so that the speaker can not hear his own voice.
  • an electronic device with a call function when a call is made based on a voice call system, signal data on the two channels of play and acquisition needs to be buffered into respective buffers. Therefore, there is a delay in the signal data currently sent to the speaker to actually emit the sound, and there is also a delay in the microphone input signal collected by the microphone back to the acquisition path, thereby obtaining the reference signal to the microphone input signal. A time delay is formed.
  • the echo cancellation module in the electronic device is doing In echo cancellation, the signal data read from the buffer corresponding to the playback and acquisition threads is echoed at the same time and the read signal data is echo-cancelled, but because of the time delay, the echo cancellation module reads from the acquisition buffer. The time of the signal data is delayed by the time when the same signal data is read from the playback buffer, so the above time delay can also be referred to as an echo delay.
  • the order, the amount of calculation, and the amount of data required by the adaptive filter to process the signal data are greatly increased, and finally the filtering effect is affected. Therefore, when performing the echo cancellation operation, it is necessary to shorten the echo delay as much as possible. To shorten its echo delay, you first need to determine the length of its echo delay.
  • the voice communication system of the existing electronic device plays and collects the sound signal through different hardware channels and different threads. When collecting threads, it is difficult to directly determine the length of the echo delay because there is no signal data in the respective buffer areas.
  • the existing method for determining the echo delay generally analyzes the signal data similar to the reference signal in the microphone input signal based on the correlation, thereby estimating the echo delay, but the method must wait until the communication device establishes a connection with another communication device. When the other speaker speaks, the measurement can be started. It is a passive measurement, and the calculation and measurement operations are cumbersome and the reliability of the measurement result is low.
  • the invention provides a method and a device for determining echo delay, and an intelligent conference device, which simply and effectively determines the echo delay in an electronic device with a call function, and provides a basis for improving the stability of the echo cancellation performance.
  • an embodiment of the present invention provides a method for determining an echo delay, where the method includes:
  • the audio signal data is stored in a set reference buffer, wherein the audio test signal is at least one single frequency signal;
  • the echo delay is determined based on the first time information and the second time information.
  • an embodiment of the present invention further provides an apparatus for determining an echo delay, the apparatus comprising:
  • a reference data storage module configured to: when playing audio signal data including an audio test signal based on the audio output device, acquire the audio signal data and store the data in a set reference buffer, where the audio test signal is at least one way Single frequency signal.
  • An input data storage module configured to acquire input signal data picked up by the audio input device, and store the input signal data in an input buffer, wherein the input signal data includes an associated signal of the audio test signal data;
  • a first information determining module configured to determine current time information corresponding to when the audio test signal is searched in the reference buffer, and record the first time information
  • a second information determining module configured to determine current time information corresponding to the search for the associated signal data in the input buffer, and record the second time information
  • an echo delay determination module configured to determine the echo delay based on the first time information and the second time information.
  • the embodiment of the present invention further provides an intelligent conference device, which integrates an apparatus for determining echo delay according to an embodiment of the present invention.
  • the present invention provides a method and device for determining echo delay, and a smart conference device.
  • the audio signal data including the audio test signal is played based on the audio output device
  • the audio signal data is acquired and stored in the setting.
  • the reference buffer then acquiring the input signal data picked up by the audio input device and storing the input signal data into the input buffer; thereafter, searching for the associated signal data in the reference buffer audio test signal and the input buffer respectively And determining time information when the audio test signal and the associated signal data are searched, respectively recorded as the first time information and the second time information, thereby determining the echo delay based on the first time information and the second time information.
  • the echo delay in the electronic device can be simply and effectively determined by setting the test signal before the electronic device having the call function establishes a call connection with other electronic devices having the call function, so as to eliminate the echo delay And provide the basis for improving the stability of echo cancellation performance.
  • FIG. 1 is a flowchart of a method for determining an echo delay according to Embodiment 1 of the present invention
  • FIG. 2 is a flowchart of a method for determining an echo delay according to Embodiment 2 of the present invention
  • Embodiment 3 is a flowchart of a method for determining an echo delay according to Embodiment 3 of the present invention.
  • FIG. 4 is a structural block diagram of an apparatus for determining an echo delay according to Embodiment 4 of the present invention.
  • FIG. 1 is a flowchart of a method for determining an echo delay according to Embodiment 1 of the present invention.
  • the embodiment is applicable to determining an echo delay in an electronic device having a call function, and the method may be performed by an echo.
  • the determining device of the delay is executed.
  • the device can be implemented in hardware and/or software and can generally be integrated into an electronic device having a voice call function.
  • a voice call with other electronic devices having a voice call function is implemented mainly based on a voice call system in the electronic device.
  • the method for determining the echo delay provided by the embodiment of the present invention specifically determines the echo delay generated by the electronic device based on the set audio test signal before the current electronic device activates the voice call system and establishes a call connection with other electronic devices.
  • the electronic device may specifically refer to a device that can implement a voice call based on a voice call system, such as a mobile phone, a computer, or a smart conference device.
  • a method for determining an echo delay includes:
  • the audio signal data including the audio test signal is first sent, and then the audio signal data is stored in a play buffer based on the call principle of the voice call system.
  • the audio signal data is read from the play buffer and played by the audio output device of the electronic device, wherein the audio output device may be specifically referred to as an earpiece and a speaker in the electronic device.
  • An audio playback device for playing audio data It should be noted that, while the audio signal data is played based on the audio output device, the audio signal data needs to be acquired and stored in a set reference buffer.
  • the audio signal data stored in the reference buffer may be specifically read by the echo cancellation module and used as an input value of the adaptive filter in the echo cancellation module for subsequent echo cancellation operations.
  • an echo cancellation module is generally disposed in the electronic device for canceling the call.
  • the generated echo the working principle of the echo canceling module based on the embodiment of the present invention has been described in the above background art, and will not be described in detail herein.
  • the audio test signal is specifically a preset at least one single frequency signal
  • the single frequency signal may be an audio signal of any single frequency and is generally stored in a set signal buffer.
  • the single-frequency signal has the characteristics of short cycle and simple fluctuation, so it is suitable for use as an audio test signal.
  • the audio signal data is specifically understood as the signal data that can be played back from the audio output device, which is required for determining the echo delay. It should be noted that, in order to ensure effective measurement of the echo delay, the audio is performed in the embodiment of the present invention.
  • the signal data includes a piece of silent signal data in addition to the audio test signal.
  • the audio test signal may form corresponding associated signal data, and the associated signal data may specifically be associated signal data formed by the audio test signal.
  • the associated signal data is picked up by the audio input device and referred to as input signal data together with other signal data picked up by the audio input device.
  • the input signal data is first stored in a set input buffer area, and then the input signal data is read from the input buffer area in a set manner and Transfer to the echo cancellation module afterwards.
  • the echo delay is specifically understood as a time difference between when the audio signal data in the reference buffer is read by the echo cancellation module and when the associated signal data in the input buffer is read by the echo cancellation module.
  • the time difference can be obtained based on the method for determining the echo delay provided by the embodiment of the present invention. Specifically, as shown in steps S103 and S104, in order to determine the echo delay, it is first necessary to separately read signal data of a set length in the reference buffer area and the input buffer area; and then read the signal data.
  • the start time of searching for the audio test signal and the associated signal data from the reference buffer area and the input buffer area is not limited.
  • the step may be started after step S102 is completed.
  • a search operation of an audio test signal or the associated signal data Since the chronological order is not limited when searching the respective signal data in the two buffer areas, when the echo is determined to be delayed, the time information recorded when the audio test signal is searched for needs to be specifically a relative time period information. And the time period information is a time when the audio test signal or the associated signal data is searched for in the currently read signal data and the signal data is started to be read in the buffer area (reference buffer area or input buffer area). The time difference between the moments.
  • the first time information may be specifically understood as time period information formed from the first reading of the signal data in the reference buffer to the end of searching for the audio test signal in the read signal data;
  • the second time information can be specifically understood as time period information formed from the first reading of the signal data in the input buffer area to the end of searching for the associated signal data in the read signal data.
  • the time period information may be described based on a specific time, such as the time period information being the time when the audio test signal is searched and the time when the signal data is read from the buffer area (reference buffer area or input buffer area) for the first time.
  • the time period information may also be described based on a time frame (the duration of each frame is T, for example, 10 ms). For example, if the signal data read from the buffer area is recorded as the 0th frame for the first time, the time period information is The current frame number corresponding to the audio test signal.
  • the first time information is determined based on the foregoing steps S103 and S104, respectively.
  • the time difference between the second time information and the first time information may be determined. For example, if the first time information and the second time information are described based on a specific time, the a time period corresponding to the first time information, and a time period corresponding to the second time information, so that the difference between the two time periods can be determined as the echo delay; for example, the time frame is used to describe the When the first time information and the second time information are described, the number of time frames corresponding to the first time information and the number of time frames corresponding to the second time information may be determined, thereby the number of the two time frames may be The difference is determined as the echo delay.
  • a method for determining an echo delay first stores signal data including an audio test signal in a set reference buffer area and an input buffer area, and then passes the reference buffer area and the input buffer area.
  • the signal data in the search is used to determine the time information corresponding to the search for the audio test signal, and finally the echo delay of the call device is determined based on the determined time information.
  • the echo delay in the electronic device can be simply and effectively determined by setting the test signal before the electronic device having the call function establishes a call connection with other electronic devices having the call function, so as to eliminate the echo delay And provide the basis for improving the stability of echo cancellation performance.
  • Embodiment 2 is a flowchart of a method for determining an echo delay according to Embodiment 2 of the present invention.
  • the present embodiment is optimized based on the foregoing embodiment.
  • the determining is in the reference buffer.
  • the current time information corresponding to the audio test signal is recorded as the first time information, and further optimized to: read the audio signal data in the reference buffer in units of frames; and determine the obtained audio signal in the current frame. Whether the value of the data is 0, and if so, reading the audio signal data of the next frame and repeating the determination operation of the value; if not, determining that the audio is included in the signal data of the current frame
  • the test signal records that the frame number of the current frame is the first time information.
  • the determining the current time information corresponding to the audio test signal in the input buffer area is recorded as the second time information, and is embodied as: reading the input buffer area in units of frames Input signal data; determining a corresponding energy value of the associated signal data in the acquired current frame, as a test energy value; determining a total energy value corresponding to the input signal data in the current frame, and recording the current frame total An energy value; if the test energy value is greater than a set threshold and the ratio of the test energy value to the current frame total energy value is greater than a set parameter value, determining that the associated signal is included in input signal data of a current frame Data, the frame number of the current frame is recorded as the second time information; otherwise, the input signal data of the next frame is read and the determination and determination of the energy value and the energy ratio are repeatedly performed.
  • a method for determining an echo delay according to Embodiment 2 of the present invention specifically includes the following operations:
  • a silent signal is first sent (the silent signal is specific) It can be understood as signal data with a value of 0), and then the constructed audio test signal (for at least one single-frequency signal) is generated, thereby forming audio signal data including the silent signal and the audio test signal and playing based on the audio output device. And stored to the reference buffer while playing.
  • the audio input device after the audio output device plays the audio signal data, the audio input device first picks up a silent signal in the audio signal data and then picks up the associated signal data of the audio test signal. When the audio input device picks up the audio signal data, the audio input device also picks up some noise signals. In this embodiment, the picked up silent signals, associated signal data, and other noise signals are referred to as input signal data. .
  • the step S203 to the step S206 are the determining process of the first time information
  • the step S207 to the step S210 are the determining process of the second time information.
  • the determination of the first time information can be started, and the determination is as follows based on step S202.
  • the determination of the second time information can be started by storing the input signal data in the input buffer area.
  • the present embodiment starts the determining operation of the time information after completing the above steps S201 and S202.
  • the frame may be The unit reads the signal data (including the audio signal data and the cached data that may have existed before) in the reference buffer.
  • the reading of the audio signal data in the reference buffer area in units of frames may be specifically understood as reading unit frame data in the reference buffer area each time.
  • the number of sampling points included in the unit frame can be determined under the premise of the known sampling rate. If the adoption rate is 16 kHz and the duration of the unit frame is set to 10 ms, the number of sampling points included in the unit frame is 160, that is, Every time from the reference cache The number of read signal data in the area is 160.
  • Step S205 after acquiring the audio signal data of the current frame, it may be determined whether the value of the audio signal data in the obtained current frame is 0. If it is 0, the audio signal data in the current frame is a silent signal, and Step S205 is performed; otherwise, the audio test signal is included in the current frame, and step S206 needs to be performed.
  • the audio signal data of the next frame may be read in the reference buffer, and then return to step S204 for the next frame.
  • the audio signal data is judged.
  • S206 Determine the audio test signal included in the audio signal data of the current frame, and record that the frame number of the current frame is the first time information.
  • the obtained value of the audio signal data in the current frame is not 0, it may be determined that the audio test signal in the reference buffer is searched for in the current frame, thereby the frame number of the current frame may be recorded. For the first time information.
  • the input signal data is stored in the input buffer. Before the area, it is not clear whether data exists in the input buffer area. Therefore, when the second time information is determined, the signal data (including the input signal data and the input signal data) in the input buffer area can be read in units of frames. Cache data that may have existed before).
  • the input signal in the acquired input buffer cannot be obtained.
  • the data simply searches for the associated signal data by determining whether the value of the signal data is zero.
  • the energy value corresponding to the associated signal data may be determined.
  • the determination of the energy value of the signal data can be implemented by means of Fourier transform or fast Fourier transform, and for a single frequency signal, it can also be simply obtained based on a specific calculation formula.
  • the energy value of the associative signal data obtained is referred to as a test energy value.
  • the total energy value corresponding to the input signal data included in the current frame may also be determined.
  • the total energy value can be obtained by a corresponding convolution formula.
  • test energy value is greater than a set threshold and the ratio of the test energy value to the current frame total energy value is greater than a set parameter value, determining that the associated signal data is included in the input signal data of the current frame. , the frame number of the current frame is recorded as the second time information; otherwise, the input of the next frame is read.
  • the signal data is entered and the process returns to step S208.
  • determining whether the associated signal data is included in the current frame may be implemented based on a determination condition set in step S210, and if the determination condition is met, the current frame is considered to include the The signal data is associated. At this time, the frame number of the current frame may be recorded as the second time information; otherwise, the input signal data of the next frame needs to be acquired and returned to S208 to perform the determination operation again.
  • the first time information and the second time information respectively determined based on steps S203 to S206 and steps S207 to S210 are respectively equivalent to current frames corresponding to when the audio test signal is searched in the reference buffer area.
  • the echo delay is a length of time corresponding to the difference between the first time information and the second time information, wherein the difference is the number of frames of the time difference frame.
  • a method for determining an echo delay provided by Embodiment 2 of the present invention, which stipulates a process for determining first time information and second time information, by which an electronic device having a call function and other call functions can be used. Before the electronic device establishes the call connection, the echo delay in the electronic device is simply and effectively determined by the set test signal, thereby providing a basis for eliminating the echo delay and improving the stability of the echo cancellation performance.
  • FIG. 3 is a flowchart of a method for determining an echo delay according to Embodiment 3 of the present invention.
  • the embodiment of the present invention is optimized based on the foregoing embodiment.
  • the determining method further includes: setting Sinusoidal and cosine signals of the same frequency are stored in the set sine buffer and In the cosine buffer; wherein the sinusoidal signal and/or the cosine signal are used as the audio test signal.
  • determining the energy value corresponding to the associated signal data in the acquired current frame is recorded as a test energy value, and the specific optimization is: determining the association based on the set power calculation formula. a power value of the signal data in the current frame, wherein the power calculation formula is set to:
  • P F represents the power of the frequency F corresponding to the associated signal data
  • micBuf[i] represents the amplitude corresponding to the i+1th sampling point in the input buffer in the unit frame
  • sinBuf[i] represents the i+ in the unit frame.
  • cosBuf[i] represents the amplitude corresponding to the i+1th sampling point in the unit frame in the cosine buffer area
  • N represents the number of sampling points in the unit frame;
  • the energy value of the associated signal data in the current frame is determined based on the power value, and is recorded as a test energy value.
  • the total energy value corresponding to the input signal data in the current frame is also determined, and is recorded as the total energy value of the current frame, and the specific optimization is: determining the method according to the set total power calculation formula.
  • the total power value of the input signal data in the current frame wherein the total power calculation formula of the input signal data in the unit frame is set to: E T represents the total power value of the input signal data in the unit frame T, micBuf[i] represents the amplitude corresponding to the i+1th sampling point in the input buffer in the unit frame, and N represents the number of sampling points in the unit frame.
  • the embodiment of the present invention further optimizes: deleting the second time information in the input buffer area. Corresponding to the signal data of the set number of frames before the signal data, so that the input signal data in the input buffer area and the audio signal data in the reference buffer area enter the echo cancellation module simultaneously; wherein the set number of frames Equal to the number of time frames corresponding to the echo delay.
  • a method for determining an echo delay according to Embodiment 3 of the present invention specifically includes the following operations:
  • S301 Set a sine signal and a cosine signal of the same frequency, and store them in the set sine buffer area and the cosine buffer area respectively.
  • step S301 is equivalent to a pre-processing operation, that is, used to preset a single-frequency signal for testing.
  • the single-frequency signal is set to a sinusoidal signal or a cosine signal, and is pre- Stored in the set sine buffer and cosine buffer.
  • the voice test signal is at least one single frequency signal, that is, there is a possibility that the voice test signal is composed of multiple single frequency signals, and the multiple single frequency signals may be sinusoidal signals or cosines. Signal, so in order to guarantee the single frequency characteristic of the audio test signal, it is necessary to set the frequency of the sine signal and the cosine signal to be the same.
  • steps S302-S304 have been specifically described in the above embodiments, and will not be described in detail herein.
  • the first time information in this embodiment is the current frame number corresponding to the search for the audio test signal.
  • the following steps S305-S310 specifically describe the determining process of the second time information in this embodiment.
  • the associated signal data of the audio test signal also corresponds to a single frequency signal, and only the energy value corresponding to the associated signal data needs to be considered, so No need to calculate based on complex Fourier transform or fast Fourier transform, you can directly use the set energy calculation formula to simply find.
  • the power calculation formula is set to:
  • P F represents the power of the frequency F corresponding to the associated signal data
  • micBuf[i] represents the amplitude corresponding to the i+1th sampling point in the input buffer in the unit frame
  • sinBuf[i] represents the i+ in the unit frame.
  • cosBuf[i] represents the amplitude of the i+1th sampling point in the unit frame in the cosine buffer area
  • N represents the number of sampling points in the unit frame.
  • S307. Determine, according to the power value, an energy value of the associated signal data in the current frame, and record it as a test energy value.
  • the energy value is equal to the power value multiplied by the time, after calculating the power value, the energy value of the associated signal data in the current frame may be determined,
  • the energy value is a test energy value.
  • the total power value of the signal data in the current frame may also be determined based on the set total power calculation formula. Specifically, the formula based on the total power is set to: The total power value of the signal data in the current frame may be determined, where E T represents the total power value of the signal data in the unit frame T, and micBuf[i] represents the i+1th sampling point in the unit frame. Enter the amplitude corresponding to the buffer area, where N is the number of sample points in the unit frame.
  • the total energy value of the input signal data in the current frame may be determined and recorded as the current frame total energy value.
  • test energy value is greater than a set threshold and the ratio of the test energy value to the current frame total energy value is greater than a set parameter value, determining that the input signal data of the current frame includes the associated signal data
  • the frame number of the current frame is recorded as the second time information; otherwise, the input signal data of the next frame is read and the process returns to step S306.
  • determining whether the input signal data of the current frame includes the associated signal data is: determining whether W F is greater than W0 and whether W F /W T is greater than a set parameter value, wherein W0 is a set energy threshold, generally taking a historical experience value of the single frequency signal, and the set parameter value may be It is preferably 0.6.
  • the echo delay is a length of time corresponding to four frames, and the length of one frame is assumed to be 10ms, the length of 4 frames is 40ms.
  • the time when the associated signal data of the audio test signal enters the input buffer is delayed by the time of entering the reference buffer, that is, the echo cancellation module is in the reference buffer.
  • the associated signal data cannot be read in the input buffer area. Therefore, it is necessary to delete the signal data corresponding to the echo extension degree before the associated signal data in the input buffer area, so as to ensure that the echo cancellation module reads the audio test signal in the reference buffer area as well The associated signal data is read in the input buffer.
  • step S312 since the time difference between the play thread and the capture thread remains unchanged after the voice call system is started, after the operation of step S312 is performed according to this embodiment, the length of the signal data buffered in the reference buffer area will be longer than the input buffer area.
  • the length of the buffered signal data is longer than the data length corresponding to the echo delay.
  • the audio signal data in the reference buffer area and the input signal data related to the audio signal data in the input buffer area can simultaneously enter the echo cancellation module. This reduces the processing time of the adaptive filter for the audio signal data, thereby ensuring the efficiency of the echo cancellation.
  • the method for determining the echo delay provided by the third embodiment of the present invention determines the second time information.
  • the process of determining the test energy value and the total energy value of the current frame is embodied; at the same time, the deletion operation of the signal data is added, and after the echo delay is determined, the input signal is caused by deleting the signal data based on the echo delay.
  • the input signal data enters the echo cancellation module simultaneously with the audio signal data in the reference buffer, thereby greatly reducing the processing time of the adaptive filter processing the audio signal data.
  • the echo delay in the electronic device can be determined simply and effectively by setting the test signal before the electronic device with the call function establishes a call connection with other electronic devices having the call function, so as to eliminate the echo delay and Improve the stability of echo cancellation performance to provide a basis for greatly reducing the processing time for echo cancellation.
  • the determining apparatus includes: a reference data storage module 41, an input data storage module 42, a first information determining module 43, a second information determining module 44, and an echo delay determining module 45.
  • the reference data storage module 41 is configured to acquire the audio signal data and store the audio signal data in a set reference buffer area when the audio output data including the audio test signal is played based on the audio output device, where the audio test signal For at least one single frequency signal.
  • the input data storage module 42 is configured to acquire input signal data picked up based on the audio input device, and store the input signal data in an input buffer, where the input signal data includes an association of the audio test signal Signal data.
  • the first information determining module 43 is configured to determine current time information corresponding to the audio test signal searched in the reference buffer, and record the first time information.
  • the second information determining module 44 is configured to determine current time information corresponding to the search for the associated signal data in the input buffer, and record the second time information.
  • the echo delay determination module 45 is configured to determine the echo delay based on the first time information and the second time information.
  • the device first acquires the audio signal data by using the reference data storage module 41 when the audio signal data including the audio test signal is played based on the audio output device, and stores the audio signal data in the set reference buffer;
  • the input data storage module 42 acquires input signal data picked up based on the audio input device, and stores the input signal data in the input buffer; and then determines, by the first information determining module 43, the audio is searched in the reference buffer.
  • the current time information corresponding to the test signal is recorded as the first time information
  • the second information determining module 44 determines the current time information corresponding to the search for the associated signal data in the input buffer area, which is recorded as the second time.
  • Information ultimately, the echo delay determination module 45 determines the echo delay based on the first time information and the second time information.
  • the apparatus for determining the echo delay provided by the fourth embodiment of the present invention can determine the electronic device directly and effectively by setting the test signal before the electronic device having the call function establishes a call connection with other electronic devices having the call function. Echo delay to provide a basis for eliminating echo delay and improving the stability of echo cancellation performance.
  • the first information determining module 43 is specifically configured to:
  • the second information determining module 44 specifically includes:
  • a signal reading unit configured to read input signal data in the input buffer area in units of frames; a first value determining unit, configured to determine a corresponding energy value of the associated signal data in the acquired current frame, Recorded as a test energy value; a second value determining unit, configured to determine a total energy value corresponding to the input signal data in the current frame, recorded as a current frame total energy value; and a test signal determining unit configured to use the test energy value If the ratio of the test energy value to the total energy value of the current frame is greater than the set parameter value, determining that the input signal data of the current frame includes the associated signal data, and recording the current frame
  • the frame number is the second time information; otherwise, the input signal data of the next frame is read and the determination and judgment operations of the energy value and the energy ratio are repeatedly performed.
  • the device further includes:
  • a signal setting and storage module for setting a sine signal and a cosine signal of the same frequency, and respectively stored in the set sine buffer area and the cosine buffer area; wherein the sine signal and/or the cosine signal are used As the audio test signal.
  • the first value determining unit is specifically configured to:
  • P F represents the power of the frequency F corresponding to the associated signal data
  • micBuf[i] represents the amplitude corresponding to the i+1th sampling point in the input buffer in the unit frame
  • sinBuf[i] represents the i+ in the unit frame.
  • cosBuf[i] represents the amplitude corresponding to the i+1th sampling point in the unit frame in the cosine buffer area
  • N represents the number of sampling points in the unit frame;
  • the energy value of the associated signal data in the current frame is determined based on the power value, and is recorded as a test energy value.
  • the second value determining unit is specifically configured to:
  • E T represents the total power value of the signal data in the unit frame T
  • micBuf[i] represents the amplitude corresponding to the i+1th sampling point in the input buffer in the unit frame
  • N represents the number of sampling points in the unit frame
  • a total energy value of the signal data in the current frame is determined based on the total power value, and is recorded as a current frame total energy value.
  • the device further includes:
  • a signal data deleting module configured to delete a set frame before the signal data corresponding to the second time information in the input buffer area after determining the echo delay based on the first time information and the second time information a plurality of signal data such that the input signal data in the input buffer and the audio signal data in the reference buffer simultaneously enter the echo cancellation module; wherein the set number of frames is equal to the echo delay corresponding to The number of time frames.
  • the embodiment of the invention provides an intelligent conference device, which integrates an apparatus for determining echo delay provided by the fourth embodiment.
  • the determination of the echo delay can be performed by the method for determining the echo delay provided in the above-described first to third embodiments.
  • the smart conference device belongs to a type of electronic device with a call function, and the smart conference device integrates a voice call system, and also has an audio input device and audio.
  • the device is output, and the voice call system includes an echo cancellation module for echo cancellation.
  • the determining device for determining the echo delay provided by the foregoing embodiment of the present invention is integrated in the smart conference device, it can be determined simply and effectively by setting the test signal before establishing a call connection with other electronic devices having a call function.
  • the echo delay in the smart conference device provides a basis for shortening or eliminating the echo delay and improving the stability of the echo cancellation module, and further improves the user experience of the smart conference device.

Landscapes

  • Engineering & Computer Science (AREA)
  • Signal Processing (AREA)
  • Telephone Function (AREA)

Abstract

本发明公开了一种回声时延的确定方法、装置及智能会议设备。该方法包括：在基于音频输出设备播放包含音频测试信号的音频信号数据时，获取所述音频信号数据并存放于参考缓存区中；获取由音频输入设备拾取的输入信号数据，并将输入信号数据存放于输入缓存区；确定在参考缓存区中搜索到音频测试信号时的第一时间信息；确定在输入缓存区中搜索到所述音频测试信号时的第二时间信息；基于第一时间信息以及第二时间信息确定回声时延。利用该方法，能够在具有通话功能的电子设备与其他具有通话功能的电子设备建立通话连接之前，通过设定的测试信号简单有效地确定电子设备中的回声时延，以便为消除回声时延以及提高回声消除性能的稳定性提供基础。

Description

一种回声时延的确定方法、装置及智能会议设备 技术领域
本发明实施例涉及声音信号处理技术领域，尤其涉及一种回声时延的确定方法、装置及智能会议设备。
背景技术
在具有通话功能的电子设备中，回声是指电子设备的扬声器播放出来的声音再次被电子设备的麦克风拾取后发回远端其他具有通话功能的电子设备，使远端的说话者能听到自己的声音。一般地，说话者很难适应在通话过程中听到自己的回声，因此，需要使用回声消除技术，或称自动回声消除(Automatic Echo Cancellation，简称AEC)来消除回声。具体地，基于AEC技术来消除回声的基本原理可描述为：从语音系统中获取扬声器当前所发出声音的参考信号，并对麦克风所拾取的麦克风输入信号进行采样，然后，将参考信号作为自适应滤波器的输入值，经过自适应滤波器的处理后其输出值就模拟了所采样的麦克风输入信号中的回声信号，之后，从所采样的麦克风输入信号中减去从自适应滤波器输出的回声信号，由此完成了回声消除，这样说话者就听不到自己的说话声。
一般地，在具有通话功能的电子设备中，基于语音通话系统进行通话时，需要把播放和采集两个通路上的信号数据先缓存到各自的缓冲区。这样当前要发送到扬声器的信号数据到实际发出声音会有一段延迟，同时麦克风采集到的麦克风输入信号回到采集通路也会存在一定延迟，由此从获取到参考信号到采集到麦克风输入信号之间便形成了时间延迟。电子设备中的回声消除模块在做 回声消除时，每次都同时从播放和采集线程对应的缓存区读取的信号数据并对所读取的信号数据进行回音消除，但因为存在时间延迟，回音消除模块从采集缓存区读取到信号数据的时间延后于从播放缓存区读取到相同信号数据的时间，所以上述时间时延也可称为回声时延。
在实际使用中，当上述回声时延太长时，自适应滤波器对信号数据进行处理时所需要的阶数、计算量以及数据量都会大幅度增大，最终影响其滤波效果。因此，在进行回音消除操作时，需要尽量的缩短回声时延。要缩短其回声时延，首先需要确定其回声时延的长度，但是现有电子设备的语音通话系统对声音信号的播放和采集是通过不同的硬件通路和不同的线程完成的，在最初启动播放和采集线程时，因不确定各自缓存区中是否已存在信号数据，所以很难直接确定回声时延的长度。
现有的回声时延的确定方法通常基于相关性来分析麦克风输入信号中相似于参考信号的信号数据，由此估计回声时延，但是该方法必须要等到通话设备与另一通话设备建立连接后，在另一方说话者讲话时才能开始测量，属于被动测量，其计算和测量的操作过程繁琐且量测结果的可靠性也较低。
发明内容
本发明提供了一种回声时延的确定方法、装置及智能会议设备，简单有效地确定了具有通话功能的电子设备中的回声时延，为提高回声消除性能的稳定性提供了基础。
本发明实施例采用以下技术方案：
第一方面，本发明实施例提供了一种回声时延的确定方法，该方法包括：
在基于音频输出设备播放包含音频测试信号的音频信号数据时，获取所述 音频信号数据并存放于设定的参考缓存区中，其中，所述音频测试信号为至少一路单频信号；
获取基于音频输入设备拾取的输入信号数据，并将所述输入信号数据存放于输入缓存区中，其中，所述输入信号数据中包含了所述音频测试信号的关联信号数据；
确定在所述参考缓存区中搜索到音频测试信号时对应的当前时间信息，记为第一时间信息；
确定在所述输入缓存区中搜索到所述关联信号数据时对应的当前时间信息，记为第二时间信息；
基于所述第一时间信息以及所述第二时间信息确定所述回声时延。
第二方面，本发明实施例还提供了一种回声时延的确定装置，该装置包括：
参考数据存放模块，用于在基于音频输出设备播放包含音频测试信号的音频信号数据时，获取所述音频信号数据并存放于设定的参考缓存区中，其中，所述音频测试信号为至少一路单频信号。
输入数据存放模块，用于获取基于音频输入设备拾取的输入信号数据，并将所述输入信号数据存放于输入缓存区中，其中，所述输入信号数据中包含了所述音频测试信号的关联信号数据；
第一信息确定模块，用于确定在所述参考缓存区中搜索到音频测试信号时对应的当前时间信息，记为第一时间信息；
第二信息确定模块，用于确定在所述输入缓存区中搜索到所述关联信号数据时对应的当前时间信息，记为第二时间信息；
回声时延确定模块，用于基于所述第一时间信息以及所述第二时间信息确定所述回声时延。
第三方面，本发明实施例又提供了一种智能会议设备，该智能会议设备集成了本发明实施例提供的一种回声时延的确定装置。
本发明提供了一种回声时延的确定方法、装置及智能会议设备，该确定方法首先在基于音频输出设备播放包含音频测试信号的音频信号数据时，获取所述音频信号数据并存放于设定的参考缓存区中；然后获取由音频输入设备拾取的输入信号数据，并将所述输入信号数据存放至输入缓存区；之后，分别在参考缓存区音频测试信号以及输入缓存区中搜索关联信号数据，并确定搜索到音频测试信号及关联信号数据时的时间信息，分别记为第一时间信息和第二时间信息，由此基于第一时间信息和第二时间信息确定回声时延。利用该确定方法，能够在具有通话功能的电子设备与其他具有通话功能的电子设备建立通话连接之前，通过设定的测试信号简单有效地确定电子设备中的回声时延，以便为消除回声时延以及提高回声消除性能的稳定性提供基础。
附图说明
图1为本发明实施例一提供的一种回声时延的确定方法的流程图；
图2为本发明实施例二提供的一种回声时延的确定方法的流程图；
图3为本发明实施例三提供的一种回声时延的确定方法的流程图；
图4为本发明实施例四提供的一种回声时延的确定装置的结构框图。
具体实施方式
下面结合附图和实施例对本发明作进一步地详细说明。可以理解的是，此处所描述的具体实施例仅仅用于解释本发明，而非对本发明的限定。另外还需要说明的是，为了便于描述，附图中仅示出了与本发明相关的部分而非全部内容。在更加详细地讨论示例性实施例之前应当提到的是，一些示例性实施例被描述成作为流程图描绘的处理或方法。虽然流程图将各项操作(或步骤)描述成顺序的处理，但是其中的许多操作可以被并行地、并发地或者同时实施。此外，各项操作的顺序可以被重新安排。当其操作完成时所述处理可以被终止，但是还可以具有未包括在附图中的附加步骤。所述处理可以对应于方法、函数、规程、子例程、子程序等等。
实施例一
图1为本发明实施例一提供的一种回声时延的确定方法的流程图，本实施例可适用于对具有通话功能的电子设备中的回声时延进行确定的情况，该方法可以由回声时延的确定装置来执行。该装置可通过硬件和/或软件的方式实现，并一般可集成于具有语音通话功能的电子设备中。
一般地，对于具有语音通话功能的电子设备而言，主要基于电子设备中的语音通话系统来实现与其他具有语音通话功能电子设备的语音通话。本发明实施例提供的回声时延的确定方法，具体在当前电子设备启动语音通话系统后，与其他电子设备建立通话连接前，基于设定的音频测试信号来确定电子设备产生的回声时延。在本实施例中，所述电子设备具体可指手机、电脑、智能会议设备等能够基于语音通话系统实现语音通话的设备。
如图1所示，本实施例提供的一种回声时延的确定方法，具体包括：
S101、在基于音频输出设备播放包含音频测试信号的音频信号数据时，获 取所述音频信号数据并存放于设定的参考缓存区中，其中，所述音频测试信号为至少一路单频信号。
在本实施例中，在检测到启动电子设备的语音通话系统后，会首先发出包含音频测试信号的音频信号数据，然后基于语音通话系统的通话原理，所述音频信号数据会存放于一个播放缓存中，之后会从该播放缓存中读取所述音频信号数据并通过电子设备的音频输出设备播放所述音频信号数据，其中，所述音频输出设备具体可指电子设备中的听筒和扬声器等用于播放音频数据的音频播放器件。需要说明的是，在基于所述音频输出设备播放所述音频信号数据的同时，还需要获取所述音频信号数据并存放至一个设定的参考缓存区中。所述参考缓存区存放的音频信号数据具体可以被回音消除模块读取并作为回声消除模块中自适应滤波器的输入值，以进行后续的回音消除操作。
在本实施例中，对于具有通话功能的电子设备而言，为了消除所述电子设备与其他电子设备建立通话时产生的回声，一般会在电子设备中设置回声消除模块，以用于消除通话时产生的回声，本发明实施例所基于的回声消除模块的工作原理已在上述背景技术中描述，这里不再详述。
在本实施例中，所述音频测试信号具体为预先设定的至少一路单频信号，所述单频信号可以是任一路单频率的音频信号并一般存放于设定的信号缓存中，由于所述单频信号具有周期短、波动简单等特点，所以适用于作为音频测试信号。所述音频信号数据具体可理解为确定所述回声时延所需的能够从音频输出设备播放的信号数据，需要说明的是，本发明实施例为了保证回声时延的有效量测，所述音频信号数据除包含所述音频测试信号外，还包括一段无声信号数据。
S102、获取基于音频输入设备拾取的输入信号数据，并将所述输入信号数据存放于输入缓存区中，其中，所述输入信号数据中包含了所述音频测试信号的关联信号数据。
在本实施例中，基于音频输出设备播放所述音频测试信号后，所述音频测试信号可能会形成相应的关联信号数据，所述关联信号数据具体可以是所述音频测试信号形成的关联信号数据，且该关联信号数据会被音频输入设备重新拾取并与音频输入设备所拾取的其他信号数据一起称为输入信号数据。此外，在拾取所述输入信号数据后，所述输入信号数据会首先存放于设定的输入缓存区中，之后会以设定的方式从所述输入缓存区中读取所述输入信号数据并传送至之后的回声消除模块。
S103、确定在所述参考缓存区中搜索到音频测试信号时对应的当前时间信息，记为第一时间信息。
S104、确定在所述输入缓存区中搜索到所述关联信号数据时对应的当前时间信息，记为第二时间信息。
一般地，所述回声时延具体可理解为参考缓存区中音频信号数据被回声消除模块所读取的时间与输入缓存区中的关联信号数据被回声消除模块所读取时的时间差，所述时间差可以基于本发明实施例提供的回声时延的确定方法来获取。具体地，由步骤S103以及S104可知，为了确定所述回声时延，首先需要在所述参考缓存区以及所述输入缓存区分别读取设定长度的信号数据；然后在所读取的信号数据中分别搜索所述音频测试信号及所述关联信号数据，如果在参考缓存区中搜索到所述音频测试信号，则记录搜索到所述音频测试信号时的时间信息；同样，如果在输入缓存区中搜索到所述关联信号数据，则记录搜索 到所述关联信号数据时的时间信息；最终，基于所记录的两个时间信息，就可以确定所述回声时延。
在本实施例中，并没有限制从所述参考缓存区以及所述输入缓存区对应搜索所述音频测试信号及所述关联信号数据的起始时间，优选地，可以在完成步骤S102之后开始所述音频测试信号或所述关联信号数据的搜索操作。由于在上述两缓存区中搜索各自信号数据时没有限定时间顺序，所以在确定所述回声时延时，需要将搜索到所述音频测试信号时记录的时间信息具体为一个相对的时间段信息，且所述时间段信息为在当前读取的信号数据中搜索到所述音频测试信号或所述关联信号数据的时刻与开始在缓存区(参考缓存区或输入缓存区)中读取信号数据的时刻的时间差。
示例性地，所述第一时间信息具体可以理解为从在所述参考缓存区中首次读取信号数据开始到在所读取的信号数据中搜索到音频测试信号结束所形成的时间段信息；所述第二时间信息具体可以理解为从在所述输入缓存区中首次读取信号数据开始到在所读取的信号数据中搜索到关联信号数据结束所形成的时间段信息。
需要注意的是，可以基于具体时刻来描述所述时间段信息，如时间段信息为搜索到音频测试信号的时刻与首次从缓存区(参考缓存区或输入缓存区)中读取信号数据的时刻之差；还可以基于时间帧(每帧时长为T，例如10ms)来描述所述时间段信息，如将首次从缓存区中读取信号数据记为第0帧，则所述时间段信息为搜索到音频测试信号时对应的当前帧号。
S105、基于所述第一时间信息以及所述第二时间信息确定所述回声时延。
在本实施例中，基于上述步骤S103以及S104分别确定所述第一时间信息 以及第二时间信息之后，可以确定所述第二时间信息与第一时间信息的时间差，示例性地，如，基于具体时刻来描述所述第一时间信息以及第二时间信息时，可以确定所述第一时间信息对应的时间段，以及所述第二时间信息对应的时间段，由此可将上述两个时间段之差确定为所述回声时延；又如，基于时间帧来描述所述第一时间信息以及第二时间信息时，可以确定所述第一时间信息对应的时间帧数，以及所述第二时间信息对应的时间帧数，由此可将上述两个时间帧数之差确定为所述回声时延。
本发明实施例提供的一种回声时延的确定方法，首先在设定的参考缓存区以及输入缓存区中存放了包含音频测试信号的信号数据，然后通过对所述参考缓存区以及输入缓存区中的信号数据进行搜索来确定搜索到音频测试信号时对应的时间信息，最终基于所确定的时间信息确定通话设备的回声时延。利用该确定方法，能够在具有通话功能的电子设备与其他具有通话功能的电子设备建立通话连接之前，通过设定的测试信号简单有效地确定电子设备中的回声时延，以便为消除回声时延以及提高回声消除性能的稳定性提供基础。
实施例二
图2为本发明实施例二提供的一种回声时延的确定方法的流程图，本实施例以上述实施例为基础进行优化，在本实施例中，将所述确定在所述参考缓存区中搜索到音频测试信号时对应的当前时间信息，记为第一时间信息，进一步优化为：以帧为单位读取所述参考缓存区中的音频信号数据；确定所获取的当前帧中音频信号数据的数值是否为0，若是，则读取下一帧的音频信号数据并重复执行数值的确定操作；若否，则确定当前帧的信号数据中包含了所述音频 测试信号，记所述当前帧的帧号为第一时间信息。
进一步地，将所述确定在所述输入缓存区中搜索到所述音频测试信号时对应的当前时间信息，记为第二时间信息，具体化为：以帧为单位读取所述输入缓存区中的输入信号数据；确定所述关联信号数据在所获取的当前帧中对应的能量值，记为测试能量值；确定所述当前帧中输入信号数据对应的总能量值，记为当前帧总能量值；如果所述测试能量值大于设定阈值且所述测试能量值与所述当前帧总能量值的比值大于设定参数值，则确定当前帧的输入信号数据中包含了所述关联信号数据，记所述当前帧的帧号为第二时间信息；否则，读取下一帧的输入信号数据并重复执行能量值及能量比值的确定及判断操作。
如图2所示，本发明实施例二提供的一种回声时延的确定方法，具体包括如下操作：
S201、在基于音频输出设备播放包含音频测试信号的音频信号数据时，获取所述音频信号数据并存放于设定的参考缓存区中，其中，所述音频测试信号为至少一路单频信号。
示例性地，为了保证电子设备的音频输入设备能够拾取到音频测试信号的关联信号数据，因此在启动语音通话系统后，与其他电子设备建立通话连接前，首先发出一段无声信号(该无声信号具体可理解为数值为0的信号数据)，然后再发出已构造好的音频测试信号(为至少一路单频信号)，由此形成包含无声信号以及音频测试信号的音频信号数据并基于音频输出设备播放，且在播放的同时存放至参考缓存区。
S202、获取基于音频输入设备拾取的输入信号数据，并将所述输入信号数据存放于输入缓存区中，其中，所述输入信号数据中包含了所述音频测试信号 的关联信号数据。
示例性地，在音频输出设备播放所述音频信号数据后，所述音频输入设备首先拾取到音频信号数据中的无声信号，然后拾取到所述音频测试信号的关联信号数据。在所述音频输入设备拾取所述音频信号数据的时候，所述音频输入设备还会拾取一些噪声信号，本实施例将所拾取的无声信号、关联信号数据以及其他噪声信号等称为输入信号数据。
在本实施例中，步骤S203～步骤S206为所述第一时间信息的确定过程，所述步骤S207～步骤S210为所述第二时间信息的确定过程。对于上述两个时间信息的确定，并不存在先后顺序的限定，只要基于步骤S201确定所述参考缓存区中存放了音频信号数据就可开始第一时间信息的确定，同样只要基于步骤S202确定所述输入缓存区中存放了输入信号数据就可开始第二时间信息的确定，优选地，本实施例在完成上述步骤S201和S202之后开始进行时间信息的确定操作。
S203、以帧为单位读取所述参考缓存区中的音频信号数据。
在本实施例中，将所述音频信号数据存入所述参考缓存区之前，并不清楚所述参考缓存区中是否已存在数据，因此，在进行第一时间信息确定时，可以以帧为单位读取所述参考缓存区中的信号数据(包括音频信号数据和之前可能存在的缓存数据)。
在本实施例中，所述以帧为单位读取所述参考缓存区中的音频信号数据具体可理解为每次读取所述参考缓存区中的单位帧数据。示例性地，在已知采样率的前提下可以确定单位帧所包含的采样点数，假设采用率为16KHz，且设定单位帧的时长为10ms，则单位帧所包含的采样点数为160，即每次从参考缓存 区中读取信号数据的个数为160。
需要注意的是，在缓存区(参考缓存区和输入缓存区)中读取信号数据(音频信号数据或输入信号数据)时，如果缓存区中当前存放的信号数据个数不足N，则需要等待所述缓存区中的信号数据个数达到N，只有在达到N之后才能再次读取，其中，N表示单位帧中包含的信号数据个数。
S204、确定所获取的当前帧中音频信号数据的数值是否为0，若是，则执行步骤S205；若否，则执行步骤S206。
在本实施例中，在获取当前帧的音频信号数据后，可以判定所获取的当前帧中音频信号数据的数值是否为0，如果为0则说明当前帧中的音频信号数据为无声信号，需要执行步骤S205；否则可认为当前帧中包含了所述音频测试信号，需要执行步骤S206。
S205、读取所述参考缓存区中下一帧的音频信号数据，并返回步骤S204。
在本实施例中，如果所获取的当前帧中信号数据的数值为0，则可接着在所述参考缓存区中读取下一帧的音频信号数据，之后再返回步骤S204对下一帧的音频信号数据进行判断。
S206、确定当前帧的音频信号数据中包含了所述音频测试信号，记所述当前帧的帧号为第一时间信息。
在本实施例中，如果所获取的当前帧中音频信号数据的数值不为0，则可确定在当前帧搜索到了所述参考缓存中的音频测试信号，由此可将当前帧的帧号记为第一时间信息。
S207、以帧为单位读取所述输入缓存区中的输入信号数据。
同上述步骤S203，在本实施例中，将所述输入信号数据存入所述输入缓存 区之前，并不清楚所述输入缓存区中是否已存在数据，因此，在进行第二时间信息确定时，可以以帧为单位读取所述输入缓存区中的信号数据(包括输入信号数据以及之前可能存在的缓存数据)。
S208、确定所述关联信号数据在所获取的当前帧中对应的能量值，记为测试能量值。
在本实施例中，由于所述音频输入设备拾取的信号除了关联信号数据之外，还存在电路噪声、环境噪声、人声等干扰信号，所以并不能对所获取的输入缓存区中的输入信号数据简单地通过判断信号数据的数值是否为0来搜索所述关联信号数据。
在本实施例中，要判断当前帧的信号数据中是否存在所述关联信号数据，首先可以确定所述关联信号数据对应的能量值。一般情况下，信号数据的能量值的确定可以通过傅里叶变换或者快速傅里叶变换的方式实现，且对于单频信号而言，还可以基于特定的计算公式来简单求取。本实施例将所求取的关联信号数据的能量值称作测试能量值。
S209、确定所述当前帧中输入信号数据对应的总能量值，记为当前帧总能量值。
在本实施例中，在基于步骤S208确定所述关联信号数据对应的测试能量值之后，还可以确定当前帧中所包含输入信号数据对应的总能量值。一般地，所述总能量值可以通过相应的卷积公式获取。
S210、如果所述测试能量值大于设定阈值且所述测试能量值与所述当前帧总能量值的比值大于设定参数值，则确定当前帧的输入信号数据中包含了所述关联信号数据，记所述当前帧的帧号为第二时间信息；否则，读取下一帧的输 入信号数据并返回步骤S208。
在本实施例中，要确定所述当前帧中是否包含所述关联信号数据，则可以基于步骤S210设定的判定条件实现，如果符合所述判定条件，则认为所述当前帧中包含所述关联信号数据，此时，可以记所述当前帧的帧号为第二时间信息；否则，需要获取下一帧的输入信号数据并返回S208再次进行判定操作。
S211、基于所述第一时间信息以及所述第二时间信息确定所述回声时延。
示例性地，基于步骤S203～步骤S206以及步骤S207～步骤S210分别确定出的第一时间信息和第二时间信息，分别相当于在所述参考缓存区中搜索到音频测试信号时对应的当前帧信息，以及在所述输入缓存区中搜索到关联信号数据时对应的当前帧信息，由于首次从所述参考缓存区以及所述输入缓存区中读取信号数据均记为第0帧，所以所述回声时延为上述第一时间信息与所述第二时间信息的差值所对应的时间长度，其中，所述差值为相差时间帧的帧数。
本发明实施例二提供的一种回声时延的确定方法，具体化了第一时间信息以及第二时间信息的确定过程，利用该确定方法，能够在具有通话功能的电子设备与其他具有通话功能的电子设备建立通话连接之前，通过设定的测试信号简单有效地确定电子设备中的回声时延，以便为消除回声时延以及提高回声消除性能的稳定性提供基础。
实施例三
图3为本发明实施例三提供的一种回声时延的确定方法的流程图，本发明实施例以上述实施例为基础进行优化，在本实施例中，该确定方法还优化包括了：设定相同频率的正弦信号和余弦信号，并分别存放于设定的正弦缓存区和 余弦缓存区中；其中，所述正弦信号和/或所述余弦信号用于作为所述音频测试信号。
进一步地，在上述优化的基础上，将确定所述关联信号数据在所获取的当前帧中对应的能量值，记为测试能量值，具体优化为：基于设定的功率计算公式确定所述关联信号数据在所述当前帧中的功率值，其中，所述功率计算公式设定为：
Figure PCTCN2016113262-appb-000001
PF表示关联信号数据所对应频率F的功率，micBuf[i]表示单位帧中第i+1个采样点在所述输入缓存区对应的幅值，sinBuf[i]表示单位帧中第i+1个采样点在所述正弦缓存区对应的幅值，cosBuf[i]表示单位帧中第i+1个采样点在所述余弦缓存区对应的幅值，N表示单位帧中的采样点数；基于所述功率值确定所述关联信号数据在所述当前帧中的能量值，记为测试能量值。
进一步地，在上述优化的基础上，还将确定所述当前帧中输入信号数据对应的总能量值，记为当前帧总能量值，具体优化为：基于设定的总功率计算公式确定所述当前帧中输入信号数据的总功率值，其中，单位帧中输入信号数据的总功率计算公式设定为：
Figure PCTCN2016113262-appb-000002
ET表示输入信号数据在单位帧T内的总功率值，micBuf[i]表示单位帧中第i+1个采样点在所述输入缓存区对应的幅值，N表示单位帧中的采样点数；基于所述总功率值确定所述当前帧中输入信号数据的总能量值，记为当前帧总能量值。
此外，本发明实施例在将基于所述第一时间信息以及所述第二时间信息确定所述回声时延之后，还优化增加了：删除所述输入缓存区中第二时间信息所 对应信号数据之前的设定帧数的信号数据，以使所述输入缓存区中的输入信号数据与所述参考缓存区中的音频信号数据同时进入回声消除模块；其中，所述设定帧数等于所述回声时延对应的时间帧数。
如图3所示，本发明实施例三提供的一种回声时延的确定方法，具体包括如下操作：
S301、设定相同频率的正弦信号和余弦信号，并分别存放于设定的正弦缓存区和余弦缓存区中。
在本实施例中，步骤S301相当于一个预处理操作，即，用来预先设定用于测试的单频信号，本实施例将所述单频信号设定为正弦信号或余弦信号，并预先存放在设定好的正弦缓存区和余弦缓存区中。
需要说明的是，由于所述语音测试信号为至少一路单频信号，即，存在由多路单频信号组成语音测试信号的可能，且上述多路单频信号有可能是正弦信号也有可能是余弦信号，所以为了保证音频测试信号的单频特性，需要设定所述正弦信号和余弦信号的频率相同。
S302、在基于音频输出设备播放包含音频测试信号的音频信号数据时，获取所述音频信号数据并存放于设定的参考缓存区中，其中，所述音频测试信号为至少一路单频信号。
S303、获取基于音频输入设备拾取的输入信号数据，并将所述输入信号数据存放于输入缓存区中，其中，所述输入信号数据中包含了所述音频测试信号的关联信号数据。
S304、确定在所述参考缓存区中搜索到音频测试信号时对应的当前时间信息，记为第一时间信息。
在本实施例中，步骤S302～S304已在上述实施例中具体阐述，这里不再详述。需要说明的是，本实施例中的第一时间信息为搜索到所述音频测试信号时对应的当前帧号。
下述步骤S305～S310具体阐述了所述第二时间信息在本实施例中的确定过程。
S305、以帧为单位读取所述输入缓存区中的输入信号数据。
S306、基于设定的功率计算公式确定所述关联信号数据在所述当前帧中的功率值。
在本实施例中，要计算所述关联信号数据在所述当前帧中的能量值，首先计算所述关联信号数据在所述当前帧中的功率值。具体地，由于本实施例中构造的音频测试信号为单频信号，则所述音频测试信号的关联信号数据也对应为单频信号，且只需要考虑所述关联信号数据对应的能量值，所以无需基于复杂的傅里叶变换或快速傅里叶变换计算，可以直接采用设定的能量计算公式来简单求取。
示例性地，所述功率计算公式设定为：
Figure PCTCN2016113262-appb-000003
PF表示关联信号数据所对应频率F的功率，micBuf[i]表示单位帧中第i+1个采样点在所述输入缓存区对应的幅值，sinBuf[i]表示单位帧中第i+1个采样点在所述正弦缓存区对应的幅值，cosBuf[i]表示单位帧中第i+1个采样点在所述余弦缓存区对应的幅值，N表示单位帧中的采样点数。
S307、基于所述功率值确定所述关联信号数据在所述当前帧中的能量值，记为测试能量值。
在本实施例中，基于能量值的计算公式，即能量值等于功率值乘以时间，在计算出所述功率值之后，可以确定所述关联信号数据在所述当前帧中的能量值，记所述能量值为测试能量值。
S308、基于设定的总功率计算公式确定所述当前帧中输入信号数据的总功率值。
在本实施例中，还可以基于设定的总功率计算公式确定所述当前帧中信号数据的总功率值。具体地，基于总功率计算公式设定为：
Figure PCTCN2016113262-appb-000004
就可以确定所述当前帧中信号数据的总功率值，其中，ET表示信号数据在单位帧T内的总功率值，micBuf[i]表示单位帧中第i+1个采样点在所述输入缓存区对应的幅值，N表示单位帧中的采样点数。
S309、基于所述总功率值确定所述当前帧中输入信号数据的总能量值，记为当前帧总能量值。
在本实施例中，确定出所述总功率值之后，就可以确定所述所述当前帧中输入信号数据的总能量值，并记为当前帧总能量值。
S310、如果所述测试能量值大于设定阈值且所述测试能量值与所述当前帧总能量值的比值大于设定参数值，则确定当前帧的输入信号数据中包含了所述关联信号数据，记所述当前帧的帧号为第二时间信息；否则，读取下一帧的输入信号数据并返回步骤S306。
示例性地，假设基于S306和S307确定出的测试能量值为WF，基于S308和S309确定出的总能量值为WT，则确定所述当前帧的输入信号数据中是否包含关联信号数据的判定条件为：确定WF是否大于W0以及WF/WT是否大于设定参数值，其中，W0为设定的能量阈值，一般取单频信号的历史经验值，所述 设定参数值可优选为0.6。
S311、基于所述第一时间信息以及所述第二时间信息确定所述回声时延。
示例性地，假设确定所述第一时间信息为第6帧，所述第二时间信息为第10帧，则所述回声时延为4帧所对应的时间长度，假设1帧的时间长度为10ms，则4帧的时间长度为40ms。
S312、删除所述输入缓存区中第二时间信息所对应信号数据之前的设定帧数的信号数据，以使所述输入缓存区中的输入信号数据与所述参考缓存区中的音频信号数据同时进入回声消除模块。
在本实施例中，通过确定所述回声时延可知，所述音频测试信号的关联信号数据进入所述输入缓存区的时间延后于进入参考缓存区的时间，即，回音消除模块在参考缓存区中读取到所述音频测试信号时，还不能在所述输入缓存区中读取到所述关联信号数据。因此，需要在所述输入缓存区中删除关联信号数据前与回声时延长度相对应的信号数据，才能保证回音消除模块在参考缓存区中读取到所述音频测试信号的同时也在所述输入缓存区中读取到关联信号数据。
需要说明的是，由于语音通话系统启动后其播放线程和采集线程的时间差保持不变，所以在基于本实施例进行步骤S312的操作后，参考缓存区中缓存的信号数据长度将比输入缓存区中缓存的信号数据长度多出回声时延所对应的数据长度。这样可以保证在基于语音通话系统进行实际通话时，所述参考缓存区中的音频信号数据与所述输入缓存区中的与所述音频信号数据相关的输入信号数据能够同时进入回声消除模块，由此降低了自适应滤波器对音频信号数据的处理时间，进而保证了回音消除的工作效率。
本发明实施例三提供的一种回声时延的确定方法，在第二时间信息的确定 过程中，具体化了测试能量值以及当前帧总能量值的确定过程；同时还增加了信号数据的删除操作，能够在确定回声时延后，基于回声时延通过删除信号数据使得输入缓存区中的输入信号数据与所述参考缓存区中的音频信号数据同时进入回声消除模块，由此大大降低自适应滤波器处理音频信号数据的处理时间。利用该方法，能够在具有通话功能的电子设备与其他具有通话功能的电子设备建立通话连接之前，通过设定的测试信号简单有效地确定电子设备中的回声时延，以便为消除回声时延以及提高回声消除性能的稳定性提供基础，大大减少回音消除的处理时间。
实施例四
图4为本发明实施例四提供的一种回声时延的确定装置的结构框图，本实施例可适用于对具有通话功能的电子设备中的回声时延进行确定的情况，该装置可通过硬件和/或软件的方式实现，并一般可集成于具有语音通话功能的电子设备中。如图4所示，该确定装置包括：参考数据存放模块41，输入数据存放模块42，第一信息确定模块43，第二信息确定模块44以及回声时延确定模块45。
其中，参考数据存放模块41，用于在基于音频输出设备播放包含音频测试信号的音频信号数据时，获取所述音频信号数据并存放于设定的参考缓存区中，其中，所述音频测试信号为至少一路单频信号。
输入数据存放模块42，用于获取基于音频输入设备拾取的输入信号数据，并将所述输入信号数据存放于输入缓存区中，其中，所述输入信号数据中包含了所述音频测试信号的关联信号数据。
第一信息确定模块43，用于确定在所述参考缓存区中搜索到音频测试信号时对应的当前时间信息，记为第一时间信息。
第二信息确定模块44，用于确定在所述输入缓存区中搜索到所述关联信号数据时对应的当前时间信息，记为第二时间信息。
回声时延确定模块45，用于基于所述第一时间信息以及所述第二时间信息确定所述回声时延。
在本实施例中，该装置首先通过参考数据存放模块41在基于音频输出设备播放包含音频测试信号的音频信号数据时，获取所述音频信号数据并存放于设定的参考缓存区中；然后通过输入数据存放模块42获取基于音频输入设备拾取的输入信号数据，并将所述输入信号数据存放于输入缓存区中；之后又通过第一信息确定模块43确定在所述参考缓存区中搜索到音频测试信号时对应的当前时间信息，记为第一时间信息，以及第二信息确定模块44确定在所述输入缓存区中搜索到所述关联信号数据时对应的当前时间信息，记为第二时间信息；最终通过回声时延确定模块45基于所述第一时间信息以及所述第二时间信息确定所述回声时延。
本发明实施例四提供的一种回声时延的确定装置，能够在具有通话功能的电子设备与其他具有通话功能的电子设备建立通话连接之前，通过设定的测试信号简单有效地确定电子设备的回声时延，以便为消除回声时延以及提高回声消除性能的稳定性提供基础。
进一步地，所述第一信息确定模块43，具体用于：
以帧为单位读取所述参考缓存区中的音频信号数据；确定所获取的当前帧中音频信号数据的数值是否为0，若是，则读取下一帧的音频信号数据并重复 执行数值的确定操作；若否，则确定当前帧的信号数据中包含了所述音频测试信号，记所述当前帧的帧号为第一时间信息。
进一步地，所述第二信息确定模块44，具体包括：
信号读取单元，用于以帧为单位读取所述输入缓存区中的输入信号数据；第一数值确定单元，用于确定所述关联信号数据在所获取的当前帧中对应的能量值，记为测试能量值；第二数值确定单元，用于确定所述当前帧中输入信号数据对应的总能量值，记为当前帧总能量值；测试信号判定单元，用于当所述测试能量值大于设定阈值且所述测试能量值与所述当前帧总能量值的比值大于设定参数值时，则确定当前帧的输入信号数据中包含了所述关联信号数据，记所述当前帧的帧号为第二时间信息；否则，读取下一帧的输入信号数据并重复执行能量值及能量比值的确定及判断操作。
在上述实施例的基础上，该装置还包括：
信号设定与存放模块，用于设定相同频率的正弦信号和余弦信号，并分别存放于设定的正弦缓存区和余弦缓存区中；其中，所述正弦信号和/或所述余弦信号用于作为所述音频测试信号。
进一步地，所述第一数值确定单元，具体用于：
基于设定的功率计算公式确定所述关联信号数据在所述当前帧中的功率值，其中，所述功率计算公式设定为：
Figure PCTCN2016113262-appb-000005
PF表示关联信号数据所对应频率F的功率，micBuf[i]表示单位帧中第i+1个采样点在所述输入缓存区对应的幅值，sinBuf[i]表示单位帧中第i+1个采样点在所述正弦缓存区对应的幅值，cosBuf[i]表示单位帧中第i+1个采样点在所述余弦缓存区对应的幅 值，N表示单位帧中的采样点数；基于所述功率值确定所述关联信号数据在所述当前帧中的能量值，记为测试能量值。
进一步地，所述第二数值确定单元，具体用于：
基于设定的总功率计算公式确定所述当前帧中信号数据的总功率值，其中，单位帧中信号数据的总功率计算公式设定为：
Figure PCTCN2016113262-appb-000006
ET表示信号数据在单位帧T内的总功率值，micBuf[i]表示单位帧中第i+1个采样点在所述输入缓存区对应的幅值，N表示单位帧中的采样点数；基于所述总功率值确定所述当前帧中信号数据的总能量值，记为当前帧总能量值。
在上述实施例的基础上，该装置还包括：
信号数据删除模块，用于在基于所述第一时间信息以及所述第二时间信息确定所述回声时延之后，删除所述输入缓存区中第二时间信息所对应信号数据之前的设定帧数的信号数据，以使所述输入缓存区中的输入信号数据与所述参考缓存区中的音频信号数据同时进入回声消除模块；其中，所述设定帧数等于所述回声时延对应的时间帧数。
实施例五
本发明实施例提供了一种智能会议设备，该智能会议设备集成了上述实施例四提供的一种回声时延的确定装置。可以通过上述实施例一至实施例三提供的回声时延的确定方法进行回声时延的确定。
在本实施例中，所述智能会议设备属于具有通话功能的电子设备的一种，所述智能会议设备中集成有语音通话系统，同时还具备音频输入设备以及音频 输出设备，且该语音通话系统中包括用于回音消除的回声消除模块。
在所述智能会议设备中集成本发明上述实施例提供的一种回声时延的确定装置之后，能够在与其他具有通话功能的电子设备建立通话连接之前，通过设定的测试信号简单有效地确定该智能会议设备中的回声时延，由此为缩短或消除回声时延以及提高回声消除模块工作性能的稳定性提供基础，同时也进一步提高了智能会议设备的用户体验。
注意，上述仅为本发明的较佳实施例及所运用技术原理。本领域技术人员会理解，本发明不限于这里所述的特定实施例，对本领域技术人员来说能够进行各种明显的变化、重新调整和替代而不会脱离本发明的保护范围。因此，虽然通过以上实施例对本发明进行了较为详细的说明，但是本发明不仅仅限于以上实施例，在不脱离本发明构思的情况下，还可以包括更多其他等效实施例，而本发明的范围由所附的权利要求范围决定。

Claims (10)

  1. 一种回声时延的确定方法，其特征在于，包括：
    在基于音频输出设备播放包含音频测试信号的音频信号数据时，获取所述音频信号数据并存放于设定的参考缓存区中，其中，所述音频测试信号为至少一路单频信号；
    获取基于音频输入设备拾取的输入信号数据，并将所述输入信号数据存放于输入缓存区中，其中，所述输入信号数据中包含了所述音频测试信号的关联信号数据；
    确定在所述参考缓存区中搜索到音频测试信号时对应的当前时间信息，记为第一时间信息；
    确定在所述输入缓存区中搜索到所述关联信号数据时对应的当前时间信息，记为第二时间信息；
    基于所述第一时间信息以及所述第二时间信息确定所述回声时延。
  2. 根据权利要求1所述的方法，其特征在于，所述确定在所述参考缓存区中搜索到音频测试信号时对应的当前时间信息，记为第一时间信息，具体包括：
    以帧为单位读取所述参考缓存区中的音频信号数据；
    确定所获取的当前帧中音频信号数据的数值是否为0，若是，则读取下一帧的音频信号数据并重复执行数值的确定操作；若否，则确定当前帧的信号数据中包含了所述音频测试信号，记所述当前帧的帧号为第一时间信息。
  3. 根据权利要求1所述的方法，其特征在于，所述确定在所述输入缓存区中搜索到所述音频测试信号时对应的当前时间信息，记为第二时间信息，具体包括：
    以帧为单位读取所述输入缓存区中的输入信号数据；
    确定所述关联信号数据在所获取的当前帧中对应的能量值，记为测试能量值；
    确定所述当前帧中输入信号数据对应的总能量值，记为当前帧总能量值；
    如果所述测试能量值大于设定阈值且所述测试能量值与所述当前帧总能量值的比值大于设定参数值，则确定当前帧的输入信号数据中包含了所述关联信号数据，记所述当前帧的帧号为第二时间信息；否则，读取下一帧的输入信号数据并重复执行能量值及能量比值的确定及判断操作。
  4. 根据权利要求3所述的方法，其特征在于，还包括：
    设定相同频率的正弦信号和余弦信号，并分别存放于设定的正弦缓存区和余弦缓存区中；
    其中，所述正弦信号和/或所述余弦信号用于作为所述音频测试信号。
  5. 根据权利要求4所述的方法，其特征在于，所述确定所述关联信号数据在所获取的当前帧中对应的能量值，记为测试能量值，具体包括：
    基于设定的功率计算公式确定所述关联信号数据在所述当前帧中的功率值，其中，所述功率计算公式设定为：
    Figure PCTCN2016113262-appb-100001
    PF表示关联信号数据所对应频率F的功率，micBuf[i]表示单位帧中第i+1个采样点在所述输入缓存区对应的幅值，sinBuf[i]表示单位帧中第i+1个采样点在所述正弦缓存区对应的幅值，cosBuf[i]表示单位帧中第i+1个采样点在所述余弦缓存区对应的幅值，N表示单位帧中的采样点数；
    基于所述功率值确定所述关联信号数据在所述当前帧中的能量值，记为测试能量值。
  6. 根据权利要求3所述的方法，其特征在于，所述确定所述当前帧中输入信号数据对应的总能量值，记为当前帧总能量值，具体包括：
    基于设定的总功率计算公式确定所述当前帧中输入信号数据的总功率值，其中，单位帧中输入信号数据的总功率计算公式设定为：
    Figure PCTCN2016113262-appb-100002
    ET表示输入信号数据在单位帧T内的总功率值，micBuf[i]表示单位帧中第i+1个采样点在所述输入缓存区对应的幅值，N表示单位帧中的采样点数；
    基于所述总功率值确定所述当前帧中输入信号数据的总能量值，记为当前帧总能量值。
  7. 根据权利要求1-6任一所述的方法，其特征在于，在基于所述第一时间信息以及所述第二时间信息确定所述回声时延之后，还包括：
    删除所述输入缓存区中第二时间信息所对应信号数据之前的设定帧数的信号数据，以使所述输入缓存区中的输入信号数据与所述参考缓存区中的音频信号数据同时进入回声消除模块；
    其中，所述设定帧数等于所述回声时延对应的时间帧数。
  8. 一种回声时延的确定装置，其特征在于，包括：
    参考数据存放模块，用于在基于音频输出设备播放包含音频测试信号的音频信号数据时，获取所述音频信号数据并存放于设定的参考缓存区中，其中，所述音频测试信号为至少一路单频信号。
    输入数据存放模块，用于获取基于音频输入设备拾取的输入信号数据，并将所述输入信号数据存放于输入缓存区中，其中，所述输入信号数据中包含了所述音频测试信号的关联信号数据；
    第一信息确定模块，用于确定在所述参考缓存区中搜索到音频测试信号时对应的当前时间信息，记为第一时间信息；
    第二信息确定模块，用于确定在所述输入缓存区中搜索到所述关联信号数据时对应的当前时间信息，记为第二时间信息；
    回声时延确定模块，用于基于所述第一时间信息以及所述第二时间信息确定所述回声时延。
  9. 根据权利要求8所述的装置，其特征在于，所述第一信息确定模块，具体用于：
    以帧为单位读取所述参考缓存区中的音频信号数据；
    确定所获取的当前帧中音频信号数据的数值是否为0，若是，则读取下一帧的音频信号数据并重复执行数值的确定操作；若否，则确定当前帧的信号数据中包含了所述音频测试信号，记所述当前帧的帧号为第一时间信息。
  10. 一种智能会议设备，其特征在于，该智能会议设备集成了如权利要求8或9所述的回声时延的确定装置。
PCT/CN2016/113262 2016-08-31 2016-12-29 一种回声时延的确定方法、装置及智能会议设备 Ceased WO2018040430A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201610799055.2 2016-08-31
CN201610799055.2A CN106210371B (zh) 2016-08-31 2016-08-31 一种回声时延的确定方法、装置及智能会议设备

Publications (1)

Publication Number Publication Date
WO2018040430A1 true WO2018040430A1 (zh) 2018-03-08

Family

ID=58086046

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/113262 Ceased WO2018040430A1 (zh) 2016-08-31 2016-12-29 一种回声时延的确定方法、装置及智能会议设备

Country Status (2)

Country Link
CN (1) CN106210371B (zh)
WO (1) WO2018040430A1 (zh)

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106210371B (zh) * 2016-08-31 2018-09-18 广州视源电子科技股份有限公司 一种回声时延的确定方法、装置及智能会议设备
CN108702568B (zh) * 2016-12-30 2020-04-21 华为技术有限公司 一种测试音频回路时延的方法及设备
CN108630215B (zh) * 2017-09-21 2020-02-21 视联动力信息技术股份有限公司 一种基于视联网的回声抑制方法及装置
CN109658946A (zh) * 2017-10-12 2019-04-19 深圳前海黑鲸科技有限公司 一种回声处理方法、装置、存储介质及终端设备
CN109040501A (zh) * 2018-09-10 2018-12-18 成都擎天树科技有限公司 一种改善voip通话质量的回声消除方法
CN110166882B (zh) * 2018-09-29 2021-05-25 腾讯科技（深圳）有限公司 远场拾音设备、及远场拾音设备中采集人声信号的方法
CN109346098B (zh) * 2018-11-20 2022-06-07 网宿科技股份有限公司 一种回声消除方法及终端
CN110265048B (zh) * 2019-03-18 2021-11-02 阿波罗智联(北京)科技有限公司 回声消除方法、装置、设备及存储介质
CN110691301A (zh) * 2019-09-25 2020-01-14 晶晨半导体（深圳）有限公司 一种测试远场语音设备与外置喇叭之间延迟时间的方法
CN112423120B (zh) * 2020-01-22 2023-09-01 上海哔哩哔哩科技有限公司 音频时延的检测方法及系统
CN113593589B (zh) * 2020-04-30 2022-06-28 阿波罗智联(北京)科技有限公司 回声时延检测方法、装置及电子设备
CN112489671B (zh) * 2020-12-18 2024-04-16 深圳市潮流网络技术有限公司 回声数据消除方法及装置、终端设备、存储介质
TWI778502B (zh) * 2021-01-22 2022-09-21 威聯通科技股份有限公司 回聲延時估計方法及回聲延時估計系統
CN113707160B (zh) * 2021-03-05 2026-03-31 腾讯科技（深圳）有限公司 一种回声延迟确定方法、装置、设备及存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102780821A (zh) * 2012-07-06 2012-11-14 歌尔声学股份有限公司 一种送受话端采样率偏差纠正方法和系统
CN202949477U (zh) * 2012-07-06 2013-05-22 歌尔声学股份有限公司 一种送受话端采样率偏差纠正系统
CN105304093A (zh) * 2015-11-10 2016-02-03 百度在线网络技术（北京）有限公司 用于语音识别的信号前端处理方法及装置
CN105872156A (zh) * 2016-05-25 2016-08-17 腾讯科技（深圳）有限公司 一种回声时延跟踪方法及装置
CN106210371A (zh) * 2016-08-31 2016-12-07 广州视源电子科技股份有限公司 一种回声时延的确定方法、装置及智能会议设备

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102780821A (zh) * 2012-07-06 2012-11-14 歌尔声学股份有限公司 一种送受话端采样率偏差纠正方法和系统
CN202949477U (zh) * 2012-07-06 2013-05-22 歌尔声学股份有限公司 一种送受话端采样率偏差纠正系统
CN105304093A (zh) * 2015-11-10 2016-02-03 百度在线网络技术（北京）有限公司 用于语音识别的信号前端处理方法及装置
CN105872156A (zh) * 2016-05-25 2016-08-17 腾讯科技（深圳）有限公司 一种回声时延跟踪方法及装置
CN106210371A (zh) * 2016-08-31 2016-12-07 广州视源电子科技股份有限公司 一种回声时延的确定方法、装置及智能会议设备

Also Published As

Publication number Publication date
CN106210371A (zh) 2016-12-07
CN106210371B (zh) 2018-09-18

Similar Documents

Publication Publication Date Title
WO2018040430A1 (zh) 一种回声时延的确定方法、装置及智能会议设备
CN110602327B (zh) 语音通话方法、装置、电子设备及计算机可读存储介质
CN103152546B (zh) 基于模式识别和延迟前馈控制的视频会议回声抑制方法
CN103391347B (zh) 一种自动录音的方法及装置
CN108134863B (zh) 一种基于双统计量的改进型双端检测装置及检测方法
WO2016180100A1 (zh) 一种音频处理的性能提升方法及装置
WO2020097828A1 (zh) 回声消除方法、延时估计方法、装置、存储介质及设备
WO2017202292A1 (zh) 一种回声时延跟踪方法及装置
CN113823306A (zh) 语音的回声消除方法、装置、设备及存储介质
CN111883163A (zh) 音频啸叫抑制方法、设备和系统及神经网络训练方法
CN109920444B (zh) 回声时延的检测方法、装置以及计算机可读存储介质
CN110246516B (zh) 一种语音通信中小空间回声信号的处理方法
CN111883154A (zh) 回声消除方法及装置、计算机可读的存储介质、电子装置
WO2018040432A1 (zh) 一种音频通话的实现方法、系统及智能会议设备
WO2016078233A1 (zh) 录音方法、装置及终端
CN113870889B (zh) 一种回声消除中的时延估计方法、装置以及电子设备
CN111800552A (zh) 音频输出处理方法、装置、系统及电子设备
CN115174724A (zh) 通话降噪方法、装置、设备及可读存储介质
CN107452398B (zh) 回声获取方法、电子设备及计算机可读存储介质
WO2017045512A1 (zh) 一种语音识别的方法、装置、终端及语音识别设备
CN116935872A (zh) 残留回声估计方法、装置、系统、设备及存储介质
CN115410593A (zh) 音频信道的选择方法、装置、设备及存储介质
CN103312911B (zh) 数据处理方法和电子终端
CN116312545B (zh) 多噪声环境下的语音识别系统和方法
WO2018040431A1 (zh) 一种信号数据的处理方法、装置及智能会议设备

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16914991

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16914991

Country of ref document: EP

Kind code of ref document: A1