• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    Voice / Audio Signal Processing Method and Apparatus Embodiments of the present invention disclose a voice / audio signal processing method and apparatus. In one embodiment, the voice / audio signal processing method includes: when an alternating band voice / audio signal, obtaining an initial high frequency signal corresponding to a current voice / audio signal frame; obtaining a time domain global gain parameter of the initial high frequency signal; performing weighting processing on an energy ratio and the time domain world gain parameter, and using a weighted value obtained as a global gain parameter, where the energy ratio is a relationship between the energy of a historical signal frame in the domain of high frequency time and energy of a current high frequency initial signal frame; rectifies the initial high frequency signal using the predicted global gain parameter to obtain a corrected high frequency time domain signal; and synthesizing a current signal frame in the narrow frequency time domain and corrected high frequency time domain signal and outputting the synthesized signal.
    公开号:BR112014021407B1
    申请号:R112014021407-7
    申请日:2013-03-01
    公开日:2019-11-12
    发明作者:Miao Lei；Liu Zexin
    申请人:Huawei Tech Co Ltd；
    IPC主号:
    专利说明:

    [001] This application claims the priority of Chinese Patent Application No. 201,210,051,672.6, filed with the Chinese Patent Office on March 1, 2012, and entitled METHOD AND APPARATUS FOR PROCESSING VOICE / AUDIO SIGNAL, which is incorporated herein by reference in its entirety.
    TECHNICAL FIELD [002] The present invention relates to the field of digital signal processing technologies, and in particular, to a method and apparatus for processing voice / audio signals.
    FUNDAMENTALS [003] In the field of digital communication, transmission of voice, images, audio and videos is required in a wide range of applications, such as a mobile phone call, an audio / video conference, broadcast television and multimedia entertainment. The audio is digitized, and is transmitted from one terminal to another terminal using an audio communication network. The terminal here can be a cell phone, a digital phone terminal, or an audio terminal of any other type, where the digital phone terminal is, for example, a VOIP phone, an ISDN phone, a computer or a communication phone cable. To reduce the resources occupied by a voice / audio signal during storage or transmission, the voice / audio signal is compressed at a transmission terminal and then transmitted to a receiving terminal, and at the receiving terminal, the voice / audio signal is
    Petition 870180140087, of 10/10/2018, p. 7/53
    2/41 restored through decompression processing and is reproduced.
    [004] In today's multi-rate voice / audio encoding, because of different network states, a network truncates bit streams at different bit rates, where bit streams are transmitted from an encoder to the network, and in a decoder, the truncated bit streams are decoded into speech / audio signals of different bandwidths. As a result, the outgoing voice / audio signals switch between different bandwidths.
    [005] Sudden switching between signals of different bandwidth causes obvious audible discomfort to human ears. In addition, because updating filter states during time-frequency transform or frequency-time transform generally requires the use of a consecutive interframe parameter, when some proper processing is not performed during bandwidth switching, an error may occur during the update. of these states, which causes some phenomena of abrupt changes in energy and deterioration of sound quality.
    SUMMARY [006] An objective of the modalities of the present invention is to provide a method and apparatus for processing the voice / audio signal, in order to improve sound comfort during switching the bandwidth of voice / audio signals.
    [007] In accordance with an embodiment of the present invention, a method of processing the voice / audio signal includes:
    when a voice / audio signal switches from a
    Petition 870180140087, of 10/10/2018, p. 8/53
    3/41 wide frequency signal for a narrow frequency signal, obtain an initial high frequency signal corresponding to a current voice / audio signal frame;
    obtain a global time domain gain parameter of the high frequency signal according to a spectrum slope parameter of the current voice / audio signal frame and a correlation between a current narrow frequency signal frame and a history frame narrow frequency signal;
    correct the initial high frequency signal by using the global time domain gain parameter to obtain a corrected high frequency time domain signal; and synthesizing a current frame of the narrow frequency time domain signal and the corrected high frequency time domain signal and outputting the synthesized signal.
    [008] According to another embodiment of the present invention, a method of processing the voice / audio signal includes:
    when a voice / audio signal switches bandwidth, obtain an initial high frequency signal corresponding to a current voice / audio signal frame;
    obtain a global time domain gain parameter of the initial high frequency signal;
    perform weighting processing on an energy ratio and the global time domain gain parameter, and use a weighted value obtained as a predicted global gain parameter, where the energy ratio is a ratio between the energy of a history frame of high frequency and energy time domain signal from a current frame of
    Petition 870180140087, of 10/10/2018, p. 9/53
    4/41 initial high frequency signal;
    correct the initial high frequency signal by using the predicted global gain parameter, to obtain a corrected high frequency time domain signal; and synthesizing a current frame of the narrow frequency time domain signal and the corrected high frequency time domain signal and outputting the synthesized signal.
    [009] According to another embodiment of the present invention, a voice / audio signal processing apparatus includes:
    a prediction unit, configured for: when a voice / audio signal switches from a wide frequency signal to a narrow frequency signal, obtain an initial high frequency signal corresponding to a current voice / audio signal frame;
    a parameter obtaining unit, configured to obtain a global time domain gain parameter of the high frequency signal according to a spectrum inclination parameter of the current voice / audio signal frame and a correlation between a current frame of voice narrow frequency signal and a narrow frequency signal history chart;
    a correction unit, configured to correct the initial high frequency signal by using the predicted global gain parameter, to obtain a corrected high frequency time domain signal; and a synthesis unit, configured to synthesize a current frame of the narrow frequency time domain signal and the high frequency time domain signal
    Petition 870180140087, of 10/10/2018, p. 10/53
    5/41 corrected and output the synthesized signal.
    [0010] According to another embodiment of the present invention, a voice / audio signal processing apparatus includes:
    an acquisition unit, configured for: when a voice / audio signal switches bandwidth, obtain an initial high frequency signal corresponding to a current voice / audio signal frame;
    a parameter acquisition unit, configured to obtain a global time domain gain parameter corresponding to the initial high frequency signal;
    a weighting processing unit, configured to perform weighting processing on an energy ratio and the time domain global gain parameter, and use a weighted value obtained as a predicted global gain parameter, where the energy ratio is a ratio between the energy of a high frequency time domain signal history frame and the energy of a current initial high frequency signal frame;
    a correction unit, configured to correct the initial high frequency signal by using the predicted global gain parameter, to obtain a corrected high frequency time domain signal; and a synthesis unit, configured to synthesize a current frame of narrow frequency time domain signal and the corrected high frequency time domain signal to emit the synthesized signal.
    [0011] In the embodiments of the present invention, during switching between a wide frequency band and a narrow frequency band, a high frequency signal is
    Petition 870180140087, of 10/10/2018, p. 11/53
    6/41 corrected in order to implement a smooth transition of the high frequency signal between the wide frequency band and the narrow frequency band, thus effectively eliminating sound discomfort caused by switching between the wide frequency band and the narrow frequency band; furthermore, because a bandwidth switching algorithm and a high frequency signal encoding / decoding algorithm before switching are in the same signal domain, this not only ensures that no additional delay is added and the algorithm is simple, but also ensures performance of an output signal.
    BRIEF DESCRIPTION OF THE DRAWINGS [0012] In order to describe the technical solutions in the modalities of the present invention or in the state of the art more clearly, the following briefly presents the attached drawings necessary to describe the modalities or the prior art. Apparently, the accompanying drawings in the following description show only a few embodiments of the present invention, and a person with current knowledge in the art can still derive other designs from these attached drawings without creative efforts.
    [0013] Figure 1 is a schematic flowchart of an embodiment of a method of processing the voice / audio signal according to the present invention;
    [0014] Figure 2 is a schematic flowchart of another embodiment of a method of processing the voice / audio signal according to the present invention;
    [0015] Figure 3 is a schematic flowchart of another modality of a signal processing method
    Petition 870180140087, of 10/10/2018, p. 12/53
    7/41 voice / audio according to the present invention;
    [0016] Figure 4 is a schematic flowchart of another embodiment of a method of processing the voice / audio signal according to the present invention;
    [0017] Figure 5 is a schematic structural diagram of an embodiment of a voice / audio signal processing apparatus according to the present invention;
    [0018] Figure 6 is a schematic structural diagram of an embodiment of a voice / audio signal processing apparatus according to the present invention;
    [0019] The figure 7 is one diagram structural schematic of a modality of a unity of obtaining parameter according to gift invention;[0020] The figure 8 is one diagram structural schematic of a modality of a unity of obtaining parameter of global gain in a deal with the present invention; [0021] The figure 9 is one diagram structural schematic of a modality in an acquisition unit according to this invention; and [0022] The figure 10 is one diagram structural schematic of another modality of a Device of
    voice / audio signal processing according to the present invention.
    DESCRIPTION OF THE MODALITIES [0023] The following clearly and completely describes the technical solutions in the modalities of the present invention with reference to the attached drawings in the modalities of the present
    Petition 870180140087, of 10/10/2018, p. 13/53
    8/41 invention. Apparently, the described modalities are only a part, instead of all the modalities of the present invention. All other modalities obtained by one of ordinary skill in the art based on the modalities of the present invention without creating efforts must fall within the scope of protection of the present invention.
    [0024] In the area of digital signal processing, audio codecs and video codecs are widely used in various electronic devices, for example, a cell phone, a wireless device, a personal data assistant (PDA), a computer handheld or portable, a GPS receiver / navigator, a camera, an audio / video player, a video camera, video recorder, and a monitoring device. Typically, this type of electronic device includes an audio encoder or an audio decoder, where the audio encoder or decoder can be directly implemented by a digital circuit or a chip, for example, a DSP (digital signal processor), or be implemented by a software code triggering a processor to execute a process on the software code.
    [0025] In the state of the art, because bandwidth of voice / audio signals transmitted over a network are different, in a process of transmitting voice / audio signals, bandwidths of voice / audio signals change frequently, and switching phenomena from a narrow frequency voice / audio signal to a wide frequency voice / audio signal and switching from a wide frequency voice / audio signal to a narrow frequency voice / audio signal exist. Such a process of switching
    Petition 870180140087, of 10/10/2018, p. 14/53
    9/41 a voice / audio signal between high and low frequency bands is referred to as bandwidth switching. Bandwidth switching includes switching from a narrow frequency signal to a broad frequency signal and switching from a wide frequency signal to a narrow frequency signal. The narrow frequency signal mentioned in the present invention is a voice signal that has only one low frequency component and a high frequency component is empty after sampling above and low pass filtering, whereas the voice / audio signal of Wide frequency has both a low frequency signal component and a high frequency signal component. The narrow frequency signal and the broad frequency signal are relative. For example, for a narrowband signal, a broadband signal is a broadband signal; and for a broadband signal, a super wide band signal is a broad frequency signal. In general, a narrowband signal is a voice / audio signal with a sampling rate of 8 kHz; a broadband signal is a voice / audio signal that the sample rate is 16 kHz; and a super wide band signal is a voice / audio signal of which a sampling rate is 32 kHz.
    [0026] When a high frequency signal encoding / decoding algorithm before switching is selected from time domain and frequency domain encoding / decoding algorithms according to different signal types, or when a coding algorithm of the high frequency signal before switching is a time domain coding algorithm, in order to ensure continuity of
    Petition 870180140087, of 10/10/2018, p. 15/53
    10/41 output signals during switching, a switching algorithm is maintained in a signal domain for processing, where the signal domain is the same as that of the high frequency encoding / decoding algorithm before switching. That is, when the time domain encoding / decoding algorithm is used for the high frequency signal before switching, a time domain switching algorithm is used as a switching algorithm to be used; when the frequency domain encoding / decoding algorithm is used for the high frequency signal before switching, a frequency domain switching algorithm is used as a switching algorithm to be used. In the prior art, when a time domain frequency band extension algorithm is used before switching, a similar time domain switching technology is not used after switching.
    [0027] In voice / audio coding, processing is generally performed using a frame as a unit. A current input audio frame that needs to be processed is a current voice / audio signal frame. The current voice / audio signal frame includes a narrow frequency signal and a high frequency signal, that is, a current narrow frequency signal frame and a current high frequency signal frame. Any voice / audio signal frame before the current high frequency signal frame is a voice / audio signal history frame, which also includes a narrow frequency signal history frame and a signal signal history frame.
    Petition 870180140087, of 10/10/2018, p. 16/53
    11/41 high frequency. A voice / audio signal frame prior to the current voice / audio signal frame is a previous voice / audio signal frame.
    [0028] Referring to Figure 1, an embodiment of a voice / audio signal processing method of the present invention includes:
    [0029] S101: When a voice / audio signal switches bandwidth, obtain an initial high frequency signal corresponding to a current voice / audio signal frame.
    [0030] The current voice / audio signal frame includes a current narrow frequency signal frame and a current high frequency time domain signal frame. Bandwidth switching includes switching from a narrow frequency signal to a broad frequency signal and switching from a wide frequency signal to a narrow frequency signal. In the case of switching from a narrow frequency signal to a broad frequency signal, the current voice / audio signal frame is the current wide frequency signal frame, including a narrow frequency signal and a high frequency signal, and the initial high frequency signal of the current voice / audio signal frame is a real signal and can be obtained directly from the current voice / audio signal frame. In the case of switching from a wide frequency signal to a narrow frequency signal, the current voice / audio signal frame is the current narrow frequency signal frame from which a current high frequency time domain signal frame is empty, the initial high frequency signal of the current voice / audio signal frame is a predicted signal, and a high frequency signal
    Petition 870180140087, of 10/10/2018, p. 17/53
    12/41 corresponding to the current narrow frequency signal frame needs to be predicted and used as the initial high frequency signal.
    [0031] S102: Obtain a global time domain gain parameter corresponding to the initial high frequency signal.
    [0032] In the case of switching from a narrow frequency signal to a broad frequency signal, the global time domain gain parameter of the high frequency signal can be obtained by decoding. In the case of switching from a wide frequency signal to a narrow frequency signal, the global time domain gain parameter of the high frequency signal can be obtained according to the current signal frame: the global domain gain parameter The time frequency of the high frequency signal is obtained according to a spectrum slope parameter of the narrow frequency signal and a correlation between a current narrow frequency signal frame and a narrow frequency signal history frame.
    [0033] S103: Perform weighting processing on an energy ratio and the global time domain gain parameter, and use a weighted value obtained as a predicted global gain parameter, where the energy ratio is a ratio between energy of a high frequency time domain signal from a voice / audio signal history frame and initial high frequency signal energy from the current voice / audio signal frame.
    [0034] A history frame of the final output signal history frame is used as the
    Petition 870180140087, of 10/10/2018, p. 18/53
    13/41 voice / audio signal history is used, and the initial high frequency signal is used as the current voice / audio signal frame. The energy ratio Reason = Esyn (-1) / Esyn_tmp, where Esyn (-1) represents the energy of the high frequency time domain signal syn from the history frame, and Esyn_tmp represents the energy of the domain signal of initial syn high frequency time corresponding to the current frame.
    [0035] The predicted global gain parameter gain = alpha * Reason + beta * gain ', where gain' is the global time domain gain parameter, alpha + beta = 1, and alpha and beta values are different from according to different types of signals.
    [0036] S104: Correct the initial high frequency signal by using the predicted global gain parameter, to obtain a corrected high frequency time domain signal.
    [0037] The correction refers to the signal being multiplied, that is, the initial high frequency signal is multiplied by the predicted global gain parameter. In another embodiment, in step S102, a time domain envelope parameter and the global time domain gain parameter that correspond to the initial high frequency signal are obtained; therefore, in step S104, the initial high frequency signal is corrected using the time domain envelope parameter and the predicted global gain parameter, to obtain the corrected high frequency time domain signal; that is, the predicted high frequency signal is multiplied by the time domain envelope parameter and the
    Petition 870180140087, of 10/10/2018, p. 19/53
    14/41 global gain of expected time domain, to obtain the corrected high frequency time domain signal.
    [0038] In the case of switching, from a narrow frequency signal to a broad frequency signal, the time domain envelope parameter of the high frequency signal can be obtained through decoding. In the case of switching from a broad frequency signal to a narrow frequency signal, the time domain envelope parameter of the high frequency signal can be obtained according to the current signal frame: a series of predetermined values or a high frequency time domain envelope parameter of the history frame can be used as the high frequency time domain envelope parameter of the current voice / audio signal frame.
    [0039] S105: Synthesize a current frame of narrow frequency time domain signal and corrected high frequency time domain signal and output the synthesized signal.
    [0040] In the previous modality, during switching between a wide frequency band and a narrow frequency band, a high frequency signal is corrected, in order to implement a smooth transition of the high frequency signal between the wide frequency band and the frequency band narrow, thus effectively eliminating sound discomfort caused by switching between the wide frequency band and the narrow frequency band; furthermore, because of a bandwidth switching algorithm and a high frequency signal encoding / decoding algorithm before
    Petition 870180140087, of 10/10/2018, p. 20/53
    15/41 switching are in the same signal domain, this not only ensures that no additional delay is added and the algorithm is simple, but also guarantees the performance of an output signal.
    [0041] Referring to Figure 2, another embodiment of a voice / audio signal processing method of the present invention includes:
    [0042] S201: When a wide frequency signal switches to a narrow frequency signal, provide for a predicted high frequency signal corresponding to a current narrow frequency signal frame.
    [0043] When a wide frequency signal switches to a narrow frequency signal, a previous frame is the wide frequency signal, and a current frame is the narrow frequency signal. The step of predicting a predicted high frequency signal corresponding to a current narrow frequency signal frame includes: predicting a high frequency signal excitation signal from the current voice / audio signal frame according to the current signal signal frame. narrow frequency; predict a LPC coefficient (Linear Predictive Coding) of the high frequency signal of the current voice / audio signal frame; and synthesizing the predicted high-frequency excitation signal and the LPC coefficient, to obtain the predicted high-frequency signal syn_tmp.
    [0044] In one embodiment, parameters such as a step period, an algebraic coding book, and a gain can be extracted from the narrow frequency signal, and the high frequency excitation signal is predicted by resampling and filtering.
    Petition 870180140087, of 10/10/2018, p. 21/53
    16/41 [0045] In another mode, operations such as sampling above, low pass, and obtaining an absolute value or a square can be performed on the narrow frequency time domain signal or a time domain excitation signal narrow frequency in order to predict the high frequency excitation signal.
    [0046] To assume the LPC coefficient of the high frequency signal, a high frequency LPC coefficient of a history frame or a series of predefined values can be used as the LPC coefficient of the current frame; or different prediction modes can be used for different types of signal.
    [0047] S202: Obtain a time domain envelope parameter and a global time domain gain parameter that correspond to the predicted high frequency signal.
    [0048] A series of predetermined values can be used as the high frequency time domain envelope parameter of the current frame. Narrowband signals can generally be classified into several types, a series of values can be predefined for each type, and a group of predefined time domain envelope parameters can be selected according to the current frame types of band signals narrow; or a group of time domain envelope values can be defined, for example, when the number of time domain envelopes is M, the default values can be M 0.3536s. In this mode, obtaining a time domain envelope parameter is an optional step, but not a necessary one.
    Petition 870180140087, of 10/10/2018, p. 22/53
    17/41 [0049] The global time domain gain parameter of the high frequency signal is obtained according to a spectrum slope parameter of the narrow frequency signal and a correlation between a current narrow frequency signal frame and a narrow frequency signal history chart, which includes the following steps in one mode:
    [0050] S2021: Classify the current voice / audio signal frame as a first type of signal or a second type of signal according to the spectrum inclination parameter of the current voice / audio signal frame and the correlation between the current narrow frequency signal frame and the narrow frequency signal history frame, in which in one embodiment, the first type of signal is a fricative signal, and the second type of signal is a non-fricative signal; and when the spectrum slope parameter> 5 and a correlation parameter color is less than a certain value, classify the narrow frequency signal as a fricative, and the rest as non-fricative.
    [0051] The parameter color showing the correlation between the current narrow frequency signal frame and the narrow frequency signal history frame can be determined according to an energy magnitude relationship between the signals of the same frequency band , or it can be determined according to an energy relationship between several same frequency bands, or it can be calculated according to a formula showing an autocorrelation or a cross correlation between time domain signals or showing an autocorrelation or a
    Petition 870180140087, of 10/10/2018, p. 23/53
    18/41 cross correlation between time domain excitation signals.
    [0052] S2022: When the current voice / audio signal frame is a first type of signal, limit the spectrum slope parameter to less than or equal to a first predetermined value, to obtain a parameter limit value of spectrum skew, and use the spectrum skew parameter threshold value as the global time domain gain parameter of the high frequency signal. That is, when the spectrum tilt parameter of the current voice / audio signal frame is less than or equal to the first predetermined value, an original value of the spectrum tilt parameter is maintained as the tilt parameter threshold value. spectrum; when the spectrum tilt parameter of the current voice / audio signal frame is greater than the first predetermined value, the first predetermined value is used as the spectrum tilt parameter threshold value.
    [0053] The time gain domain global gain parameter 'is obtained according to the following formula:
    where tilt is the spectrum slope parameter, and dl is the first predetermined value.
    [0054] S2023: When the current voice / audio signal frame is a second type of signal, limit the spectrum tilt parameter to a value in a first interval, to obtain a threshold parameter value of
    Petition 870180140087, of 10/10/2018, p. 24/53
    19/41 spectrum skew, and use the spectrum skew parameter threshold value as the global time domain gain parameter of the high frequency signal. That is, when the spectrum tilt parameter of the current voice / audio signal frame belonging to the first interval, an original value of the spectrum tilt parameter is maintained as the limit value of the spectrum tilt parameter; when the spectrum pitch parameter of the current voice / audio signal frame is greater than an upper limit of the first interval, the upper limit of the first interval is used as the spectrum pitch parameter limit value; when the spectrum slope parameter of the current voice / audio signal frame is less than the lower limit of the first interval, the lower limit of the first interval is used as the spectrum slope parameter limit value.
    [0055] The time gain domain global gain parameter 'is obtained according to the following formula:
    tilt> tilt and [a t / ] gain '= a, tilt <ab, tilt> b where tilt is the spectrum slope parameter and [a, b] is the first interval.
    [0056] In one embodiment, a tilt spectrum slope parameter of a narrow frequency signal and a color parameter showing a correlation between a current narrow frequency signal frame and a narrow frequency signal history frame are
    Petition 870180140087, of 10/10/2018, p. 25/53
    Obtained 20/41; current frame of signs are classified into two types, fricative and non-fricative, according to the slope and color; when the slope spectrum parameter> 5 and the color of the correlation parameter is less than a certain value, the narrow frequency signal is classified as a fricative, the rest being non-fricative; slope is limited within a range of 0.5 <= slope <= 1.0 and is used as a global time domain gain parameter of a non-fricative, and slope is limited to a slope value range <= 8.0 and is used as a parameter of the global time domain gain of a fricative. For a fricative, a spectrum skew parameter can be any value greater than 5, and for a non-fricative, a spectrum skew parameter can be any value less than or equal to 5, or it can be greater than 5. A In order to ensure that a tilt spectrum slope parameter can be used as an estimated time domain global gain parameter, slope is limited within a range of values and then used as a global domain gain parameter time. That is, when slope> 8, it is determined that slope = 8 is used as a global gain parameter of a fricative's time domain; when slope <0.5, slope = 0.5, or when slope> 1.0, which slope = 1.0, and 0.5 or 1.0 is used as a global gain parameter time domain of a non-fricative.
    [0057] S203: Perform weighting processing on an energy ratio and the global gain parameter of
    Petition 870180140087, of 10/10/2018, p. 26/53
    21/41 time domain, and use a weighted value obtained as a predicted global gain parameter, where the energy ratio is a ratio between the energy of a high frequency time domain signal from a signal history frame. voice / audio and initial high-frequency signal energy of the current voice / audio signal frame.
    [0058] Calculation is performed on the energy ratio Ratio = Esyn (-1) / Esyn_tmp, and the weighted value of slope and Ratio is used as the predicted global gain parameter gain of the current frame, ie gain = alpha * Ratio + beta * gain ', where gain' is the global time domain gain parameter, alpha + beta = 1, alpha and beta values are different according to the different types of signal, Esyn (-1) represents the energy of the high frequency time domain signal finally emitted syn from the history frame, and Esyn_tmp represents the energy of the predicted high frequency time domain signal syn from the current frame.
    [0059] S204: Correct the predicted high frequency signal using the time domain envelope parameter and the predicted global gain parameter, to obtain a corrected high frequency time domain signal.
    [0060] The predicted high frequency signal is multiplied by the time domain envelope parameter and the predicted global time domain gain parameter, to obtain the high frequency time domain signal.
    [0061] In this mode, the time domain envelope parameter is optional. When only the time domain global gain parameter is included, the predicted high frequency signal can be corrected using the
    Petition 870180140087, of 10/10/2018, p. 27/53
    22/41 predicted global gain parameter, to obtain the corrected high frequency time domain signal. That is, the predicted high frequency signal is multiplied by the predicted global gain parameter to obtain the corrected high frequency time domain signal.
    [0062] S205: Synthesize the current frame of the narrow frequency time domain signal and the corrected high frequency time domain signal and output the synthesized signal.
    [0063] The Esyn energy of the syn high frequency time domain signal is used to predict a global time domain gain parameter for a next frame. That is, an Esyn value is assigned to Esyn (-1).
    [0064] In the previous modality, a high frequency band of a narrow frequency signal following a broad frequency signal is corrected, in order to implement a smooth transition of the high frequency part between a broad frequency band and a frequency band narrow, thus effectively eliminating sound discomfort caused by switching between the wide frequency band and the narrow frequency band; in addition, because the corresponding processing is carried out on the board during switching, a problem that occurs during parameter and status update is indirectly eliminated. By maintaining, a bandwidth switching algorithm and a high frequency signal encoding / decoding algorithm before switching, in the same signal domain, this not only ensures that no additional delay is added and the algorithm is simple, but it also guarantees the performance of an output signal.
    Petition 870180140087, of 10/10/2018, p. 28/53
    [0065] Referring to Figure 3, another embodiment of a voice / audio signal processing method of the present invention includes:
    [0066] S301: When a narrow frequency signal switches to a broad frequency signal, obtain a current high frequency signal frame.
    [0067] When a narrow frequency signal switches to a broad frequency signal, a previous frame is a narrow frequency signal, and a current frame is a wide frequency signal.
    [0068] S302: Obtain a time domain envelope parameter and a global time domain gain parameter that correspond to the high frequency signal.
    [0069] The time domain envelope parameter and the global time domain gain parameter can be obtained directly from the current high frequency signal frame. Obtaining a time domain envelope parameter is an optional step.
    [0070] S303: Perform weighting processing on an energy ratio and the global time domain gain parameter, and use a weighted value obtained as a predicted global gain parameter, where the energy ratio is a ratio between energy from a high frequency time domain signal from a voice / audio signal history frame and energy from an initial high frequency signal from a current voice / audio signal frame.
    [0071] Since the current frame is a broad frequency signal, parameters of the high frequency signal can all be obtained through decoding. In order to
    Petition 870180140087, of 10/10/2018, p. 29/53
    24/41 To ensure a smooth transition during switching, the global time domain gain parameter is smoothed as follows:
    [0072] Calculation is performed on the energy ratio Ratio = Esyn (-1) / Esyn_tmp, where Esyn (-1) represents the energy of a high frequency time domain signal finally emitted syn from a history frame, and Esyn_tmp represents high frequency syn domain time signal energy of the current frame.
    [0073] The weighted value of the time gain global gain parameter and Reason that are obtained through decoding is used as the predicted global gain parameter of the current frame, that is, gain = alpha * Reason + beta * gain ', where gain' is the global time domain gain parameter, alpha + beta = 1, and alpha and beta values are different according to different types of signals.
    [0074] When narrowband signals from the current audio frame and a previous voice / audio signal frame have a predetermined correlation, a value obtained by attenuation, according to a given step size, a weighting factor alpha of the energy ratio corresponding to the previous voice / audio signal frame is used as a weighting factor of the energy ratio corresponding to the current audio frame, in which the attenuation is performed frame by frame, until alpha is 0.
    [0075] When narrow frequency signals of consecutive frames are of the same type of signal, or a correlation between narrow frequency signals of consecutive frames satisfies a certain condition, that is, the
    Petition 870180140087, of 10/10/2018, p. 30/53
    25/41 consecutive frames have a certain correlation or signal types of consecutive frames are similar, alpha is attenuated frame by frame according to a certain step size until alpha is attenuated to 0; when the narrow frequency signals of consecutive frames have no correlation, alpha is directly attenuated to 0, that is, a current decoding result is maintained without weighting or correcting it.
    [0076] S304: Correct the high frequency signal by using the time domain envelope parameter and the predicted global gain parameter, to obtain a corrected high frequency time domain signal.
    [0077] The correction refers to the high frequency signal being multiplied by the time domain envelope parameter and the expected global time domain gain parameter, to obtain the corrected high frequency time domain signal.
    [0078] In this mode, the time domain envelope parameter is optional. When only the time domain global gain parameter is included, the high frequency signal can be corrected using the predicted global gain parameter to obtain the corrected high frequency time domain signal. That is, the high frequency signal is multiplied by the predicted global gain parameter, to obtain the corrected high frequency time domain signal.
    [0079] S305: Synthesize a current frame of narrow frequency time domain signal and corrected high frequency time domain signal and output the synthesized signal.
    Petition 870180140087, of 10/10/2018, p. 31/53
    26/41 [0080] In the previous modality, a high frequency band of a broad frequency signal following a narrow frequency signal is corrected, in order to implement a smooth transition of the high frequency part between a broad frequency band and a narrow frequency band, thus effectively eliminating sound discomfort caused by switching between the wide frequency band and the narrow frequency band; in addition, because the corresponding processing is carried out on the board during switching, a problem that occurs during parameter and status update is indirectly eliminated. By maintaining, a bandwidth switching algorithm and a high frequency signal encoding / decoding algorithm before switching, in the same signal domain, this not only ensures that no additional delay is added and the algorithm is simple, but it also guarantees the performance of an output signal.
    [0081] Referring to Figure 4, another embodiment of a voice / audio signal processing method of the present invention includes:
    [0082] S401: When a voice / audio signal switches from a wide frequency signal to a narrow frequency signal, obtain an initial high frequency signal corresponding to a current voice / audio signal frame.
    [0083] When a wide frequency signal switches to a narrow frequency signal, a previous frame is the wide frequency signal, and a current frame is the narrow frequency signal. The step of predicting an initial high frequency signal corresponding to a current frame of
    Petition 870180140087, of 10/10/2018, p. 32/53
    27/41 narrow frequency signal includes: providing a high frequency signal excitation signal from the current voice / audio signal frame according to the current narrow frequency signal frame; predict an LPC coefficient of the high frequency signal of the current voice / audio signal frame; and synthesizing the predicted high-frequency excitation signal and the LPC coefficient, to obtain the predicted high-frequency signal syn_tmp.
    [0084] In one embodiment, parameters such as a step period, an algebraic coding book, and a gain can be extracted from the narrow frequency signal, and the high frequency excitation signal is predicted by resampling and filtering.
    [0085] In another mode, operations such as sampling above, low pass, and obtaining an absolute value or a square can be performed on the narrow frequency time domain signal or a narrow frequency time domain excitation signal , in order to predict the high frequency excitation signal.
    [0086] To assume the LPC coefficient of the high frequency signal, a high frequency LPC coefficient of a history frame or a series of predefined values can be used as the LPC coefficient of the current frame; or different prediction modes can be used for different types of signal.
    [0087] S402: Obtaining a global time domain gain parameter of the high frequency signal according to a spectrum inclination parameter of the current voice / audio signal frame and a correlation between a current narrow frequency signal frame and a history chart of
    Petition 870180140087, of 10/10/2018, p. 33/53
    28/41 narrow frequency signal.
    [0088] In one embodiment, the following steps are included:
    [0089] S2021: Classify the current voice / audio signal frame as a first type of signal or a second type of signal according to the spectrum inclination parameter of the current voice / audio signal frame and the correlation between the current narrow frequency signal frame and the narrow frequency signal history frame, in which in one embodiment, the first type of signal is a fricative signal, and the second type of signal is a non-fricative signal.
    [0090] In one mode, when the slope spectrum parameter> 5, and a correlation parameter color is less than a certain value, the narrow frequency signal is classified as a fricative, the rest being non-fricative. The parameter color showing the correlation between the current narrow frequency signal frame and the narrow frequency signal history frame can be determined according to an energy magnitude relationship between the signals in the same frequency band, or can be determined according to an energy relationship between several same frequency bands, or it can be calculated according to a formula showing an autocorrelation or a cross correlation between time domain signals or showing an autocorrelation or a cross correlation between the time domain excitement.
    [0091] S2022: When the current voice / audio signal frame is a first type of signal, limit the parameter of
    Petition 870180140087, of 10/10/2018, p. 34/53
    29/41 spectrum skew to less than or equal to a first predetermined value, to obtain a spectrum skew parameter threshold value, and use the spectrum skew parameter threshold value as the global gain parameter of time domain of the high frequency signal. That is, when the spectrum tilt parameter of the current voice / audio signal frame is less than or equal to the first predetermined value, an original value of the spectrum tilt parameter is maintained as the tilt parameter threshold value. spectrum; when the spectrum tilt parameter of the current voice / audio signal frame is greater than the first predetermined value, the first predetermined value is used as the spectrum tilt parameter threshold value.
    [0092] When the current voice / audio signal frame is a fricative signal, the global gain parameter of the time domain gain 'is obtained according to the following formula:
    where tilt is the spectrum slope parameter, and dl is the first predetermined value.
    [0093] S2023: When the current voice / audio signal frame is a second type of signal, limit the spectrum tilt parameter to a value in a first interval, to obtain a spectrum tilt parameter limit value, and use the spectrum tilt parameter threshold value as the
    Petition 870180140087, of 10/10/2018, p. 35/53
    30/41 global gain of time domain of the high frequency signal. That is, when the spectrum skew parameter of the current voice / audio signal frame belonging to the first interval, an original spectrum skew parameter value is maintained as the spectrum skew parameter threshold value; when the spectrum pitch parameter of the current voice / audio signal frame is greater than an upper limit of the first interval, the upper limit of the first interval is used as the spectrum pitch parameter limit value; when the spectrum slope parameter of the current voice / audio signal frame is less than the lower limit of the first interval, the lower limit of the first interval is used as the spectrum slope parameter limit value.
    [0094] When the current voice / audio signal frame is a non-fricative signal, the global gain parameter of the time domain gain 'is obtained according to the following formula:
    gain ’= J a, tilt <a where tilt is the spectrum slope parameter, and [a, b] is the first range.
    [0095] In one embodiment, a tilt spectrum slope parameter of a narrow frequency signal and a color parameter showing a correlation between a current narrow frequency signal frame and a narrow frequency signal history frame are
    Petition 870180140087, of 10/10/2018, p. 36/53
    Obtained 31/41; current frame of signs are classified into two types, fricative and non-fricative, according to the slope and color; when the slope spectrum parameter> 5 and the color of the correlation parameter is less than a certain value, the narrow frequency signal is classified as a fricative, the rest being non-fricative; slope is limited within a range of 0.5 <= slope <= 1.0 and is used as a global time domain gain parameter of a non-fricative, and slope is limited to a slope value range <= 8.0 and is used as a parameter of the global time domain gain of a fricative. For a fricative, a spectrum skew parameter can be any value greater than 5, and for a non-fricative, a spectrum skew parameter can be any value less than or equal to 5, or it can be greater than 5. A In order to ensure that a tilt spectrum slope parameter can be used as a predicted global gain parameter, slope is limited within a range of values and then used as a parameter in the global gain time domain. That is, when the slope> 8, the slope = 8 and 8 is determined to be used as a global time domain gain parameter of a fricative signal; when slope <0.5, slope = 0.5, or when slope> 1.0, slope = 1.0, 0.5 or 1.0 is determined and is used as a gain parameter global time domain of a non-fricative signal.
    [0096] S403: Correct the initial high frequency signal by means of the global domain gain parameter
    Petition 870180140087, of 10/10/2018, p. 37/53
    32/41 time, to obtain a corrected high frequency time domain signal.
    [0097] In one mode, the initial high frequency signal is multiplied by the global time domain gain parameter, to obtain the corrected high frequency time domain signal.
    [0098] In another embodiment, step S403 can include:
    perform weighting processing on an energy ratio and the time domain global gain parameter, and use a weighted value obtained as a predicted global gain parameter, where the energy ratio is a ratio between the energy of a frame of history of high frequency time and energy domain signal of a current initial high frequency signal frame; and correcting the initial high frequency signal by using the predicted global gain parameter, to obtain a corrected high frequency time domain signal; that is, the initial high frequency signal is multiplied by the predicted global gain parameter, to obtain a corrected high frequency time domain signal.
    [0099] Optionally, before step S403, the method can also include:
    obtaining a time domain envelope parameter corresponding to the initial high frequency signal, and the correction of the initial high frequency signal by using the predicted global gain parameter includes:
    correct the initial high frequency signal using the time domain envelope parameter and the global time domain gain parameter.
    Petition 870180140087, of 10/10/2018, p. 38/53
    33/41 [00100] S404: Synthesize a current frame of the narrow frequency time domain signal and the corrected high frequency time domain signal and output the synthesized signal.
    [00101] In the previous modality, when a wide frequency band switches to a narrow frequency band, a global time domain gain parameter of a high frequency signal is obtained according to a spectrum slope parameter and a correlation interframe. When using the narrow frequency spectrum slope parameter, an energy relationship between a narrow frequency signal and a high frequency signal can be estimated correctly, thus to better estimate energy of the high frequency signal. Using interframe correlation, an interframe correlation between high frequency signals can be estimated, making good use of the correlation between narrow frequency frames. In this way, when weighting is performed to obtain an overall high frequency gain, the previous real information can be used as well, and an undesirable noise is not introduced. The high frequency signal is corrected using the global time domain gain parameter, in order to implement a smooth transition of the high frequency part between the wide frequency band and the narrow frequency band, thus effectively eliminating sound discomfort caused by switching between the wide frequency band and the narrow frequency band.
    [00102] In association with the previous method modalities, the present invention further provides a voice / audio signal processing apparatus. The device can
    Petition 870180140087, of 10/10/2018, p. 39/53
    34/41 be located on a delivery device, a network device, or a test device. The voice / audio signal processing apparatus may be implemented by a hardware circuit, or may be implemented by software in combination with hardware. For example, referring to Figure 5, a processor calls the voice / audio signal processing apparatus to implement voice / audio signal processing. The voice / audio signal processing apparatus can perform the methods and processes in the previous method modalities.
    [00103] Referring to Figure 6, a modality of a voice / audio signal processing apparatus includes:
    an acquisition unit 601, configured for: when a voice / audio signal switches bandwidth, obtain an initial high frequency signal corresponding to a current voice / audio signal frame;
    a parameter obtaining unit 602, configured to obtain a global time domain gain parameter corresponding to the initial high frequency signal;
    a weighting processing unit 603, configured to perform weighting processing on an energy ratio and the time domain global gain parameter, and use a weighted value obtained as a predicted global gain parameter, where the energy ratio is a ratio between the energy of a high frequency time domain signal history frame and the energy of a current initial high frequency signal frame;
    a 604 correction unit, configured to correct
    Petition 870180140087, of 10/10/2018, p. 40/53
    35/41 the initial high frequency signal for using the predicted global gain parameter, to obtain a corrected high frequency time domain signal; and a synthesis unit 605, configured to synthesize a current frame of narrow frequency time domain signal and the corrected high frequency time domain signal and output the synthesized signal.
    [00104] In one embodiment, switching bandwidth is switching from a wide frequency signal to a narrow frequency signal, and the parameter obtaining unit 602 includes:
    a unit for obtaining the global gain parameter, configured to obtain the time domain global gain parameter of the high frequency signal according to a spectrum inclination parameter of the current voice / audio signal frame and a correlation between a current voice / audio signal frame and a narrow frequency signal history frame.
    [00105] Referring to Figure 7, in another embodiment, the switching of bandwidth is switching from a broad frequency signal to a narrow frequency signal, and the parameter obtaining unit 602 includes:
    a time domain envelope obtaining unit 701, configured to use a series of predefined values as a time domain high frequency envelope parameter of the current voice / audio signal frame; and a unit for obtaining the global gain parameter702, configured to obtain the global time parameter gain of the high frequency signal of
    Petition 870180140087, of 10/10/2018, p. 41/53
    36/41 according to a spectrum skew parameter of the current voice / audio signal frame and a correlation between a current voice / audio signal frame and a narrow frequency signal history frame.
    [00106] Therefore, the correction unit 604 is configured to correct the initial high frequency signal by using the time domain envelope parameter and the predicted global gain parameter, to obtain the corrected high frequency time domain signal. .
    [00107] Referring to Figure 8, in addition, a modality of the unit for obtaining global gain parameter 702 includes:
    a rating unit 801, configured to classify the current voice / audio signal frame as a first type of signal or a second type of signal according to the spectrum pitch parameter of the current voice / audio signal frame and the correlation between the current voice / audio signal frame and the narrow frequency signal history frame;
    a first 802 limiting unit, configured for: when the current voice / audio signal frame is a first type of signal, limit the spectrum slope parameter to less than or equal to a first predetermined value, to obtain a value spectrum tilt parameter limit, and use the spectrum tilt parameter limit value as the time domain global gain parameter of the high frequency signal; and a second limiting unit 803, configured for: when the current voice / audio signal frame is a second
    Petition 870180140087, of 10/10/2018, p. 42/53
    37/41 signal type, limit the spectrum tilt parameter to a value in a first interval, to obtain a spectrum tilt parameter limit value, and use the spectrum tilt parameter limit value as the parameter global gain of time domain of the high frequency signal.
    [00108] Furthermore, in one embodiment, the first type of signal is a fricative signal, and the second type of signal is a non-fricative signal; when the spectrum slope parameter> 5 and a correlation parameter color is less than a certain value, the narrow frequency signal is classified as a fricative, the rest being non-fricative; the first predetermined value is 8; and the first predefined interval is [0.5, 1].
    [00109] Referring to Figure 9, in one modality, the acquisition unit 601 includes:
    an excitation signal acquisition unit 901, configured to predict a high frequency signal excitation signal according to the current voice / audio signal frame;
    an LPC coefficient obtaining unit 902, configured to provide an LPC coefficient of the high frequency signal; and a generation unit 903, configured to synthesize the excitation signal of the high frequency signal and the LPC coefficient of the high frequency signal, to obtain the predicted high frequency signal.
    [00110] In one embodiment, bandwidth switching is switching from a narrow frequency signal to
    Petition 870180140087, of 10/10/2018, p. 43/53
    38/41 a broad frequency signal, and the voice / audio signal processing apparatus further includes:
    a weighting factor setting unit, configured for: when narrowband signals from the current voice / audio signal audio frame and a previous voice / audio signal frame have a predetermined correlation, use a value obtained by attenuation , according to a given step size, an alpha of the energy ratio weighting factor corresponding to the previous frame of the voice / audio signal as a weighting factor of the energy ratio corresponding to the current audio frame, in which the attenuation is executed frame by frame, until alpha is 0.
    [00111] Referring to Figure 10, another modality of a voice / audio signal processing apparatus includes:
    a prediction unit 1001, configured for: when a voice / audio signal switches from a wide frequency signal to a narrow frequency signal, obtain an initial high frequency signal corresponding to a current voice / audio signal frame ;
    a parameter acquisition unit 1002, configured to obtain a global time domain gain parameter of the high frequency signal according to a spectrum slope parameter of the current voice / audio signal frame and a correlation between a current frame narrow frequency signal and a narrow frequency signal history chart;
    a correction unit 1003, configured to correct the initial high frequency signal by using the parameter
    Petition 870180140087, of 10/10/2018, p. 44/53
    39/41 expected global gain, to obtain a corrected high frequency time domain signal; and a synthesis unit 1004, configured to synthesize the current frame of the narrow frequency time domain signal and the corrected high frequency time domain signal and output the synthesized signal.
    [00112] Referring to Figure 8, the unit for obtaining parameter 1002 includes:
    a rating unit 801, configured to classify the current voice / audio signal frame as a first type of signal or a second type of signal according to the spectrum pitch parameter of the current voice / audio signal frame and the correlation between the current voice / audio signal frame and the narrow frequency signal history frame;
    a first 802 limiting unit, configured for: when the current voice / audio signal frame is a first type of signal, limit the spectrum slope parameter to less than or equal to a first predetermined value, to obtain a value spectrum tilt parameter limit, and use the spectrum tilt parameter limit value as the time domain global gain parameter of the high frequency signal; and a second limiting unit 803, configured for: when the current voice / audio signal frame is a second type of signal, limit the spectrum inclination parameter to a value in a first interval, to obtain a limit value of spectrum slope parameter, and use the slope parameter limit value of
    Petition 870180140087, of 10/10/2018, p. 45/53
    40/41 spectrum as the time domain global gain parameter of the high frequency signal.
    [00113] Furthermore, in one embodiment, the first type of sign is a fricative sign, and the second type of sign is a non-fricative sign; when the spectrum slope parameter> 5 and a correlation parameter color is less than a certain value, the narrow frequency signal is classified as a fricative, the rest being non-fricative; the first predetermined value is 8; and the first predefined interval is [0.5, 1].
    [00114] Optionally, in a modality, the voice / audio signal processing device also includes:
    a weighting processing unit, configured to perform weighting processing on an energy ratio and the time domain global gain parameter, and use a weighted value obtained as a predicted global gain parameter, where the energy ratio is a ratio between the energy of a high frequency time domain signal history frame and the energy of a current initial high frequency signal frame; and the correction unit is configured to correct the initial high frequency signal by using the predicted global gain parameter, to obtain the corrected high frequency time domain signal.
    [00115] In another mode, the parameter obtaining unit is further configured to obtain a time domain parameter envelope corresponding to the initial high frequency signal; and the correction unit is configured to correct the high frequency signal
    Petition 870180140087, of 10/10/2018, p. 46/53
    41/41 initial for using the time domain envelope parameter and the global time domain gain parameter.
    [00116] A person with current knowledge in the art can understand that all or part of the processes of the methods in the modalities can be implemented by a computer program that instructs relevant hardware. The program can be stored on a computer-readable storage medium. When the program is executed, the method processes in the modalities are carried out. The storage medium can include: a magnetic disk, an optical disk, a read-only memory (Read-Only Memory, ROM), or a random access memory (Random Access Memory, RAM).
    [00117] The above are merely exemplary modalities to illustrate the present invention, but the scope of the present invention is not limited to them. Modifications or variations are readily apparent to persons skilled in the prior art, without departing from the spirit and scope of the present invention.
    权利要求:
    Claims (23)
    [1]
    1. Method of processing voice / audio signals, characterized by the fact that it comprises:
    when a signal / audio speech changes from a wide frequency signal to a narrow frequency signal, obtaining an initial high frequency signal corresponding to a current voice / audio signal frame;
    obtaining a parameter in the time domain of global high frequency gain of the signal according to a parameter of slope spectrum of the current frame of speech / audio signal and a correlation between a current frame of narrow frequency signal and a historical frame of the narrow frequency signal;
    rectifies the initial high frequency signal, using the time domain global gain parameter, to obtain a corrected high frequency time domain signal; and synthesizing a current signal frame in the narrow frequency time domain and correcting the signal in the high frequency time domain and outputting the synthesized signal.
    [2]
    2. Method, according to claim 1, characterized by the fact that obtaining a time domain global high frequency gain parameter of the signal according to a spectrum inclination parameter of the current speech / audio signal frame and a correlation between a current narrow frequency signal frame and a historical narrow frequency signal frame comprises:
    classifying the current frame of the voice / audio signal as a first type of signal or a second type of signal according to the parameter of the frame's spectrum tilt
    2/13 current signal / audio speech and the correlation between the current narrow frequency signal frame and the historic narrow frequency signal frame;
    when the current frame of the speech / audio signal is a first type of signal, which limits the slope of the parameter spectrum to less than or equal to a first predetermined value, to obtain a slope parameter limit value of the spectrum;
    when the current frame of the speech / audio signal is a second type of signal, which limits the slope of the parameter spectrum to a value of a first range, to obtain a limit value of the spectrum slope parameter; and using the limit value of the spectrum slope parameter as the domain time high frequency global gain parameter of the signal.
    [3]
    3. Method according to claim 2, characterized by the fact that the first type of sign · is a fricative sign, and the second type of sign is a non-fricative sign; when the slope parameter of the slope spectrum 5 and a color correlation parameter is less than a certain value, the narrow frequency signal is classified as a fricative, the remainder being non-fricative; the first predetermined value is 8; and the first predefined interval is [0.5; 1].
    [4]
    4. Method according to any one of claims 1 to 3, characterized by the fact that the correction of the initial high frequency signal, using the global time domain gain parameter, to obtain
    3/13 a high frequency time domain signal comprises corrected:
    performing weighting processing with an energy value and the time domain global gain parameter, and using a weighted value obtained as a global gain parameter, predicted, where the energy ratio is a ratio between the energy of a historical signal frame in the time domain and high energy frequency of a current high frequency signal frame; and rectifies the initial high-frequency signal, using the global gain parameter, he predicted.
    [5]
    5. Method according to any one of claims 1 to 3, characterized by the fact that it further comprises:
    obtaining a time domain envelope parameter that corresponds to the initial high frequency signal, in which to correct the initial high frequency signal, using the global parameter gain time domain comprises:
    rectifies the initial high frequency signal, using the envelope parameter in the time domain and the time domain global gain parameter.
    [6]
    6. Voice / audio signal processing method characterized by the fact that it comprises:
    when a voice / audio signal changes bands, obtaining an initial high-frequency signal corresponding to a current speech / audio signal frame;
    obtaining a global gain parameter time domain of the initial high frequency signal;
    4/13 performing weighting processing with an energy value and the global time domain gain parameter, and using a weighted value obtained as a global gain parameter, predicted, where the energy ratio is a ratio between the energy of a historical signal frame in the domain of time and high frequency energy of a current frame of initial high frequency signal;
    rectifies the initial high frequency signal, using the predicted global gain parameter, to obtain a corrected high frequency time domain signal; and synthesizing a current signal frame in the narrow frequency time domain and correcting the signal in the high frequency time domain and outputting the synthesized signal.
    [7]
    7. Method according to claim 6, characterized by the fact that switching the bandwidth is changing a wide frequency signal to a narrow frequency signal, and obtaining an overall gain parameter corresponding to the first high frequency signal comprises:
    obtaining a parameter in the time domain of global high frequency gain of the signal according to a parameter of slope spectrum of the current frame of speech / audio signal and a correlation between a current frame of narrow frequency signal and a historical frame of the narrow frequency signal.
    [8]
    8. Method according to claim 7, characterized by the fact that obtaining a time domain global high frequency gain parameter of the signal according to a spectrum inclination parameter of a current frame of the speech / audio signal and a correlation
    5/13 between a current frame of a narrow frequency signal and a historical frame of the narrow frequency signal comprises:
    classifying the current frame of the voice / audio signal as a first type of signal or a second type of signal according to the spectrum slope parameter of the current frame of the signal / audio speech and the correlation between the current frame of the narrow frequency signal and the historic narrow frequency signal frame;
    when the current frame of the speech / audio signal is a first type of signal, which limits the slope of the parameter spectrum to less than or equal to a first predetermined value, to obtain a slope parameter limit value of the spectrum;
    when the current frame of the speech / audio signal is a second type of signal, which limits the slope of the parameter spectrum to a value of a first range, to obtain a limit value of the spectrum slope parameter; and using the limit value of the spectrum slope parameter as the domain time high frequency global gain parameter of the signal.
    [9]
    9. Method according to claim 8, characterized by the fact that the first type of signal is a fricative signal, and the second type of signal is a non-fricative signal; when the slope parameter of the slope spectrum> 5 and a color correlation parameter is less than a certain value, the narrow frequency signal is classified as a fricative, the
    6/13 remaining non-fricatives; the first predetermined value is 8; and the first predefined interval is [0.5; 1].
    [10]
    10. Method according to claim 6, characterized by the fact that switching the bandwidth is changing a wide frequency signal to a narrow frequency signal, and obtaining a corresponding initial high frequency signal a current speech / audio frame of the signal comprises:
    predict a high frequency excitation signal according to the current voice / audio signal frame;
    predict an LPC coefficient of the high frequency signal; and synthesizing the high frequency excitation signal and the LPC coefficient of the high frequency signal, to obtain the predicted high frequency signal.
    [11]
    11. Method according to claim 6, characterized in that the switching of the bandwidth is the change from a narrow frequency signal to a wide frequency signal, and the method further comprises:
    when the narrowband signals of the current speech / audio signal frame and a previous speech / audio signal frame have a predetermined correlation, using a value obtained by attenuation, according to a given step size, an alpha factor weighting of the energy ratio corresponding to the previous voice / audio signal frame as a weighting factor of the energy index corresponding to the current audio frame, where the attenuation is made frame by frame until alpha is 0.
    7/13
    [12]
    12. Voice / audio signal processing method characterized by the fact that it comprises:
    a prediction unit, configured for: when a signal / audio speech changes from a wide frequency signal to a narrow frequency signal, obtain an initial high frequency signal corresponding to a current voice / audio signal frame;
    obtaining a parameter unit, configured to obtain a time domain parameter global high frequency gain of the signal according to a spectrum slope parameter of the current speech / audio signal frame and a correlation between a current signal frame of narrow frequency and a historic narrow frequency signal picture;
    a correction unit, configured to correct the initial high frequency signal, using the predicted global gain parameter, to obtain a corrected high frequency time domain signal; and a synthesis unit, configured to synthesize a current signal frame in the narrow frequency time domain and the signal in the corrected high frequency time domain and output of the synthesized signal.
    [13]
    13. Apparatus, according to claim 12, characterized by the fact that the unit comprises obtaining parameters:
    a classification unit, configured to classify the current frame of the voice / audio signal as a first type of signal or a second type of signal according to the spectrum inclination parameter of the current frame of the speech / audio speech and the correlation between the
    8/13 current frame of signal / audio speech and historical frame of the narrow frequency signal;
    a first limiting unit, configured for: when the current speech / audio frame is a signal of a first type of signal, limit the spectrum inclination parameter to less than or equal to a first predetermined value, to obtain a limit value of the spectrum inclination parameter, and use the limit value of the spectrum inclination parameter as the domain time global gain parameter of the high frequency signal; and a second limiting unit, configured for: when the current speech / audio frame is a signal of a second type of signal, limit the inclination of the parameter spectrum to a value in a first range, to obtain a limit value of the inclination parameter of the spectrum, and use the spectrum tilt parameter limit value as the time domain parameter global gain of the high frequency signal.
    [14]
    Apparatus according to claim 13, characterized by the fact that the first type of signal is a fricative signal, and the second type of signal is a non-fricative signal; when the slope parameter of the slope spectrum> 5 and a color correlation parameter is less than a certain value, the narrow frequency signal is classified as a fricative, the remainder being non-fricatives; the first predetermined value is 8; and the first predefined interval is [0.5, 1].
    [15]
    15. Apparatus according to any one of claims 12 to 14, characterized by the fact that it further comprises:
    9/13 a weighting processing unit, configured to perform weighting processing with an energy value and the time domain global gain parameter, and use a weighted value obtained as a global gain parameter, he predicted, where the energy ratio is a ratio between the energy of a frame of historical signal in the time domain and the energy of a current frame of initial high frequency high frequency signal, which the device is configured to correct to correct the high signal initial frequency, using the predicted global gain parameter, to obtain the corrected high frequency time domain signal.
    [16]
    16. Apparatus according to any one of claims 12 to 14, characterized by the fact that the parameter obtaining unit is further configured to obtain a time domain envelope parameter corresponding to the first high frequency signal; and the device is configured for correction to correct the initial high frequency signal, using the envelope parameter in the time domain and the time domain global gain parameter.
    [17]
    17. Voice / audio signal processing method characterized by the fact that it comprises:
    an acquisition unit, configured for: when a voice / audio signal changes band, obtain an initial high frequency signal corresponding to a current voice / audio signal frame;
    obtaining a parameter unit, configured to obtain a global time domain gain parameter corresponding to the first high frequency signal;
    10/13 a weighting processing unit, configured to perform weighting processing with an energy value and the time domain global gain parameter, and use a weighted value obtained as a global gain parameter, he predicted, where the energy ratio is a ratio between the energy of a historical signal frame in the time domain and the energy of a current high frequency high frequency initial signal frame;
    a correction unit, configured to correct the initial high frequency signal, using the predicted global gain parameter, to obtain a corrected high frequency time domain signal; and a synthesis unit, configured to synthesize a current signal frame in the narrow frequency time domain and the signal in the corrected high frequency time domain and output of the synthesized signal.
    [18]
    18. Apparatus according to claim 17, characterized by the fact that switching the bandwidth is changing a wide frequency signal to a narrow frequency signal, and obtaining unit parameters comprises:
    a global gain parameter gain unit, configured to obtain the high frequency global gain parameters time domain of the signal according to a spectrum slope parameter of the current frame of the speech / audio signal and a correlation between a frame current voice / audio signal and a historic narrow frequency signal frame.
    11/13
    [19]
    19, Apparatus, according to claim 18, characterized by the fact that the global parameter gain unit comprises:
    a classification unit, configured to classify the current frame of the voice / audio signal as a first type of signal or a second type of signal according to the spectrum inclination parameter of the current frame of the speech / audio speech and the correlation between the current signal / audio speech frame and the historical frame of the narrow frequency signal;
    a first limiting unit, configured for: when the current speech / audio frame is a signal of a first type of signal, limit the spectrum inclination parameter to less than or equal to a first predetermined value, to obtain a limit value of the spectrum slope parameter, and use the limit value of the spectrum slope parameter as the domain time global gain parameter of the high frequency signal; and a second limiting unit, configured for: when the current speech / audio frame is a signal of a second type of signal, limit the inclination of the parameter spectrum to a value in a first range, to obtain a limit value of the inclination parameter of the spectrum, and use the spectrum tilt parameter limit value as the time domain parameter global gain of the high frequency signal.
    [20]
    20. Apparatus according to claim 19, characterized by the fact that the first type of signal is a fricative signal, and the second type of signal is a non-fricative signal; when the spectrum slope parameter
    12/13 slope> 5 and a color correlation parameter is less than a certain value, the narrow frequency signal is classified as a fricative, the rest being non-fricative; the first predetermined value is 8; and the first predefined interval is [0.5, 1].
    [21]
    21. Apparatus according to any one of claims 17 to 20, characterized in that the switching of the bandwidth is switching from a narrow frequency signal to a wide frequency signal, and the apparatus further comprises:
    an envelope unit in the time domain obtaining, configured to use a series of predefined values as a high frequency time domain domain parameter of the current frame of the voice / audio signal; and the apparatus is configured for correction to correct the initial high frequency signal, using the time domain envelope parameter and the predicted global gain parameter, to obtain the corrected high frequency time domain signal.
    [22]
    22. Apparatus according to any one of claims 17 to 20, characterized by the fact that the acquisition unit comprises:
    an excitation signal obtaining unit, configured to provide a high frequency signal excitation signal according to the current voice / audio signal frame;
    an LPC coefficient obtaining unit, configured to predict an LPC coefficient of the high frequency signal; and a synthesis unit, configured to synthesize the excitation signal of the high frequency signal and the
    13/13 LPC coefficient of the high frequency signal, to obtain the expected high frequency signal.
    [23]
    23. Apparatus according to any one of claims 17 to 20, characterized in that the switching of the bandwidth is switching from a narrow frequency signal to a wide frequency signal, and the apparatus further comprises:
    a clamping unit weighting factor, configured for: when the narrowband signals of the current speech / audio signal frame and a previous speech / audio signal frame have a predetermined correlation, use a value obtained by attenuation, according to a given step size, an alpha weighting factor of the energy ratio corresponding to the previous voice / audio signal frame as a weighting factor of the energy index corresponding to the current audio frame, in which the attenuation is made frame by frame until alpha is 0.
    类似技术:
    公开号 | 公开日 | 专利标题
    BR112014021407B1|2019-11-12|Voice / audio signal processing method and handset
    RU2628197C2|2017-08-15|Masking errors in pictures
    BR112014023577B1|2021-08-24|AUDIO SIGNAL ENCODING METHOD AND DEVICE AND AUDIO SIGNAL DECODING METHOD AND DEVICE
    US20130317831A1|2013-11-28|Bandwidth expansion method and apparatus
    BR112015018064B1|2020-12-01|method for predicting high frequency band signal, decoding device and encoding device
    JP6612808B2|2019-11-27|Conversation / voice signal processing method and encoding apparatus
    CN105761724B|2021-02-09|Voice frequency signal processing method and device
    同族专利:
    公开号 | 公开日
    EP3193331B1|2019-05-15|
    EP2821993A1|2015-01-07|
    CN105469805A|2016-04-06|
    ES2867537T3|2021-10-20|
    RU2616557C1|2017-04-17|
    JP2017027068A|2017-02-02|
    ES2741849T3|2020-02-12|
    HUE053834T2|2021-07-28|
    CN105469805B|2018-01-12|
    US9691396B2|2017-06-27|
    PL3534365T3|2021-07-12|
    TR201911006T4|2019-08-21|
    CN103295578B|2016-05-18|
    US10559313B2|2020-02-11|
    KR20160121612A|2016-10-19|
    US20150006163A1|2015-01-01|
    RU2585987C2|2016-06-10|
    JP2018197869A|2018-12-13|
    US20190318747A1|2019-10-17|
    JP2015512060A|2015-04-23|
    MX2014010376A|2014-12-05|
    MX364202B|2019-04-16|
    SG11201404954WA|2014-10-30|
    MX345604B|2017-02-03|
    KR101702281B1|2017-02-03|
    ES2629135T3|2017-08-07|
    CN103295578A|2013-09-11|
    ZA201406248B|2016-01-27|
    KR101844199B1|2018-03-30|
    RU2014139605A|2016-04-20|
    WO2013127364A1|2013-09-06|
    CA2865533C|2017-11-07|
    JP6378274B2|2018-08-22|
    DK3534365T3|2021-04-12|
    EP3534365A1|2019-09-04|
    KR20170013405A|2017-02-06|
    KR101667865B1|2016-10-19|
    EP3193331A1|2017-07-19|
    MY162423A|2017-06-15|
    EP2821993A4|2015-02-25|
    JP6558748B2|2019-08-14|
    US20180374488A1|2018-12-27|
    CA2865533A1|2013-09-06|
    JP6010141B2|2016-10-19|
    EP2821993B1|2017-05-10|
    IN2014KN01739A|2015-10-23|
    BR112014021407A2|2019-04-16|
    EP3534365B1|2021-01-27|
    US10013987B2|2018-07-03|
    US20170270933A1|2017-09-21|
    SG10201608440XA|2016-11-29|
    PT2821993T|2017-07-13|
    US10360917B2|2019-07-23|
    KR20140124004A|2014-10-23|
    PT3193331T|2019-08-27|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    CA2252170A1|1998-10-27|2000-04-27|Bruno Bessette|A method and device for high quality coding of wideband speech and audio signals|
    WO2000065866A1|1999-04-26|2000-11-02|Lucent Technologies Inc.|Path switching according to transmission requirements|
    CA2290037A1|1999-11-18|2001-05-18|Voiceage Corporation|Gain-smoothing amplifier device and method in codecs for wideband speech and audio signals|
    US6606591B1|2000-04-13|2003-08-12|Conexant Systems, Inc.|Speech coding employing hybrid linear prediction coding|
    US7113522B2|2001-01-24|2006-09-26|Qualcomm, Incorporated|Enhanced conversion of wideband signals to narrowband signals|
    JP2003044098A|2001-07-26|2003-02-14|Nec Corp|Device and method for expanding voice band|
    CN101010730B|2004-09-06|2011-07-27|松下电器产业株式会社|Scalable decoding device and signal loss compensation method|
    JP5100380B2|2005-06-29|2012-12-19|パナソニック株式会社|Scalable decoding apparatus and lost data interpolation method|
    RU2414009C2|2006-01-18|2011-03-10|ЭлДжи ЭЛЕКТРОНИКС ИНК.|Signal encoding and decoding device and method|
    US20110057818A1|2006-01-18|2011-03-10|Lg Electronics, Inc.|Apparatus and Method for Encoding and Decoding Signal|
    US9454974B2|2006-07-31|2016-09-27|Qualcomm Incorporated|Systems, methods, and apparatus for gain factor limiting|
    GB2444757B|2006-12-13|2009-04-22|Motorola Inc|Code excited linear prediction speech coding|
    JP4733727B2|2007-10-30|2011-07-27|日本電信電話株式会社|Voice musical tone pseudo-wideband device, voice musical tone pseudo-bandwidth method, program thereof, and recording medium thereof|
    RU2449386C2|2007-11-02|2012-04-27|Хуавэй Текнолоджиз Ко., Лтд.|Audio decoding method and apparatus|
    CN100585699C|2007-11-02|2010-01-27|华为技术有限公司|A kind of method and apparatus of audio decoder|
    KR100930061B1|2008-01-22|2009-12-08|성균관대학교산학협력단|Signal detection method and apparatus|
    CN101499278B|2008-02-01|2011-12-28|华为技术有限公司|Audio signal switching and processing method and apparatus|
    CN101751925B|2008-12-10|2011-12-21|华为技术有限公司|Tone decoding method and device|
    JP5448657B2|2009-09-04|2014-03-19|三菱重工業株式会社|Air conditioner outdoor unit|
    CN102044250B|2009-10-23|2012-06-27|华为技术有限公司|Band spreading method and apparatus|
    US8484020B2|2009-10-23|2013-07-09|Qualcomm Incorporated|Determining an upperband signal from a narrowband signal|
    JP5287685B2|2009-11-30|2013-09-11|ダイキン工業株式会社|Air conditioner outdoor unit|
    CN101964189B|2010-04-28|2012-08-08|华为技术有限公司|Audio signal switching method and device|
    EP2676262B1|2011-02-14|2018-04-25|Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V.|Noise generation in audio codecs|
    US8000968B1|2011-04-26|2011-08-16|Huawei Technologies Co., Ltd.|Method and apparatus for switching speech or audio signals|
    CN103295578B|2012-03-01|2016-05-18|华为技术有限公司|A kind of voice frequency signal processing method and device|CN103295578B|2012-03-01|2016-05-18|华为技术有限公司|A kind of voice frequency signal processing method and device|
    CN108364657B|2013-07-16|2020-10-30|超清编解码有限公司|Method and decoder for processing lost frame|
    CN104517610B|2013-09-26|2018-03-06|华为技术有限公司|The method and device of bandspreading|
    WO2015055532A1|2013-10-18|2015-04-23|Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V.|Concept for encoding an audio signal and decoding an audio signal using deterministic and noise like information|
    MX355091B|2013-10-18|2018-04-04|Fraunhofer Ges Forschung|Concept for encoding an audio signal and decoding an audio signal using speech related spectral shaping information.|
    US20150170655A1|2013-12-15|2015-06-18|Qualcomm Incorporated|Systems and methods of blind bandwidth extension|
    KR101864122B1|2014-02-20|2018-06-05|삼성전자주식회사|Electronic apparatus and controlling method thereof|
    CN106683681B|2014-06-25|2020-09-25|华为技术有限公司|Method and device for processing lost frame|
    WO2019002831A1|2017-06-27|2019-01-03|Cirrus Logic International Semiconductor Limited|Detection of replay attack|
    GB2563953A|2017-06-28|2019-01-02|Cirrus Logic Int Semiconductor Ltd|Detection of replay attack|
    GB201713697D0|2017-06-28|2017-10-11|Cirrus Logic Int Semiconductor Ltd|Magnetic detection of replay attack|
    GB201801527D0|2017-07-07|2018-03-14|Cirrus Logic Int Semiconductor Ltd|Method, apparatus and systems for biometric processes|
    GB201801530D0|2017-07-07|2018-03-14|Cirrus Logic Int Semiconductor Ltd|Methods, apparatus and systems for authentication|
    GB201801528D0|2017-07-07|2018-03-14|Cirrus Logic Int Semiconductor Ltd|Method, apparatus and systems for biometric processes|
    GB201803570D0|2017-10-13|2018-04-18|Cirrus Logic Int Semiconductor Ltd|Detection of replay attack|
    GB201801874D0|2017-10-13|2018-03-21|Cirrus Logic Int Semiconductor Ltd|Improving robustness of speech processing system against ultrasound and dolphin attacks|
    GB201801663D0|2017-10-13|2018-03-21|Cirrus Logic Int Semiconductor Ltd|Detection of liveness|
    GB201801664D0|2017-10-13|2018-03-21|Cirrus Logic Int Semiconductor Ltd|Detection of liveness|
    GB2567503A|2017-10-13|2019-04-17|Cirrus Logic Int Semiconductor Ltd|Analysing speech signals|
    GB201719734D0|2017-10-30|2018-01-10|Cirrus Logic Int Semiconductor Ltd|Speaker identification|
    GB201801659D0|2017-11-14|2018-03-21|Cirrus Logic Int Semiconductor Ltd|Detection of loudspeaker playback|
    US11264037B2|2018-01-23|2022-03-01|Cirrus Logic, Inc.|Speaker identification|
    US10692490B2|2018-07-31|2020-06-23|Cirrus Logic, Inc.|Detection of replay attack|
    US10915614B2|2018-08-31|2021-02-09|Cirrus Logic, Inc.|Biometric authentication|
    US11037574B2|2018-09-05|2021-06-15|Cirrus Logic, Inc.|Speaker recognition and speaker change detection|
    法律状态:
    2017-12-26| B08F| Application dismissed because of non-payment of annual fees [chapter 8.6 patent gazette]|Free format text: REFERENTE A 5A ANUIDADE. |
    2018-02-20| B08G| Application fees: restoration [chapter 8.7 patent gazette]|
    2019-03-26| B15K| Others concerning applications: alteration of classification|Free format text: A CLASSIFICACAO ANTERIOR ERA: G10L 19/00 Ipc: G10L 19/083 (2013.01), G10L 19/02 (2000.01) |
    2019-04-09| B06A| Patent application procedure suspended [chapter 6.1 patent gazette]|
    2019-10-22| B09A| Decision: intention to grant [chapter 9.1 patent gazette]|
    2019-11-12| B16A| Patent or certificate of addition of invention granted [chapter 16.1 patent gazette]|Free format text: PRAZO DE VALIDADE: 20 (VINTE) ANOS CONTADOS A PARTIR DE 01/03/2013, OBSERVADAS AS CONDICOES LEGAIS. (CO) 20 (VINTE) ANOS CONTADOS A PARTIR DE 01/03/2013, OBSERVADAS AS CONDICOES LEGAIS |
    优先权:
    申请号 | 申请日 | 专利标题
    CN201210051672.6|2012-03-01|
    CN201210051672.6A|CN103295578B|2012-03-01|2012-03-01|A kind of voice frequency signal processing method and device|
    PCT/CN2013/072075|WO2013127364A1|2012-03-01|2013-03-01|Voice frequency signal processing method and device|
    [返回顶部]