Time-domain noise shaping method, apparatus, terminal, base station and chip

Abstract

The application relates to a time-domain noise shaping method, device, terminal, base station and chip. The method comprises the following steps: determining a signal detection result of a to-be-processed audio signal according to near-Nyquist frequency band energy and middle-low frequency band energy of the to-be-processed audio signal with a preset sampling rate; determining a shaping start decision result of a high frequency band spectrum of the to-be-processed audio signal according to the signal detection result and prediction gain information of the high frequency band spectrum; and performing corresponding time-domain noise shaping processing on the high frequency band spectrum according to the shaping start decision result, so as to obtain a processed high frequency band spectrum of the to-be-processed audio signal. The method can improve the signal processing quality of time-domain noise shaping.

Description

Technical Field

[0001] This application relates to the field of signal processing technology, and in particular to a time-domain noise shaping method, apparatus, terminal, base station, and chip. Background Technology

[0002] Currently, in order to improve the quality of audio signals, it is crucial to accurately perform time-domain noise shaping on audio signals with a sampling rate of 48kHz (44.1kHz).

[0003] In traditional techniques, when performing time-domain noise shaping on audio signals with a sampling rate of 48kHz (44.1kHz), a method of uniform processing of high and low frequencies is generally used; however, this method is prone to introducing noise, resulting in low signal processing quality of time-domain noise shaping. Summary of the Invention

[0004] Therefore, it is necessary to provide a time-domain noise shaping method, apparatus, terminal, base station, and chip that can improve the signal processing quality of time-domain noise shaping in response to the above-mentioned technical problems.

[0005] In a first aspect, this application provides a time-domain noise shaping method, including:

[0006] The signal detection result of the audio signal to be processed is determined based on the near Nyquist frequency band energy and the mid-low frequency band energy of the audio signal to be processed at a preset sampling rate.

[0007] Based on the signal detection results and the predicted gain information of the high-frequency band spectrum of the audio signal to be processed, the shaping start decision result of the high-frequency band spectrum is determined.

[0008] Based on the shaping start decision result, the high-frequency band spectrum is subjected to corresponding time-domain noise shaping processing to obtain the processed high-frequency band spectrum of the audio signal to be processed.

[0009] In one embodiment, the signal detection result includes a first signal detection result and a second signal detection result;

[0010] The step of determining the shaping start decision result of the high-frequency band spectrum based on the signal detection result and the predicted gain information of the high-frequency band spectrum of the audio signal to be processed includes:

[0011] When the signal detection result is the first signal detection result, or when the predicted gain information of the high-frequency band spectrum of the audio signal to be processed meets the first preset condition, the shaping start decision result of the high-frequency band spectrum is determined as the first shaping start decision result.

[0012] If the signal detection result is the second signal detection result and the prediction gain information meets the second preset condition, the shaping start decision result is determined to be the second shaping start decision result.

[0013] If the signal detection result is the second signal detection result and the predicted gain information meets the third preset condition, the shaping start decision result is determined to be the third shaping start decision result.

[0014] In one embodiment, the step of performing corresponding time-domain noise shaping on the high-frequency band spectrum based on the shaping initiation decision result to obtain the processed high-frequency band spectrum of the audio signal to be processed includes:

[0015] If the shaping initiation decision result is the first shaping initiation decision result, the processed high-frequency band spectrum of the audio signal to be processed is obtained based on the high-frequency band spectrum.

[0016] When the shaping start decision result is the second shaping start decision result, the filter coefficients are adjusted according to the weights of the filter coefficients of the high-frequency band spectrum to obtain the adjusted filter coefficients of the high-frequency band spectrum. Based on the adjusted filter coefficients, the high-frequency band spectrum is subjected to corresponding time-domain noise shaping processing to obtain the processed high-frequency band spectrum of the audio signal to be processed.

[0017] When the shaping start decision result is the third shaping start decision result, the high-frequency band spectrum is subjected to corresponding time-domain noise shaping processing according to the filter coefficients to obtain the processed high-frequency band spectrum of the audio signal to be processed.

[0018] In one embodiment, determining the signal detection result of the audio signal to be processed based on the near-Nyquist frequency band energy and mid-low frequency band energy of the audio signal to be processed at a preset sampling rate includes:

[0019] Determine the energy ratio between the near-Nyquist frequency band energy and the mid-low frequency band energy of the audio signal to be processed at the preset sampling rate;

[0020] The near-Nyquist detection result of the audio signal to be processed is determined based on the difference between the energy ratio and the preset energy ratio.

[0021] Based on the near-Nyquist detection results, the signal detection results of the audio signal to be processed are determined.

[0022] In one embodiment, determining the near-Nyquist detection result of the audio signal to be processed based on the difference between the energy ratio and a preset energy ratio includes:

[0023] If the energy ratio is greater than the preset energy ratio, the near-Nyquist detection result of the audio signal to be processed is determined as the first signal detection result;

[0024] If the energy ratio is less than or equal to the preset energy ratio, the near-Nyquist detection result of the audio signal to be processed is determined as the second signal detection result.

[0025] In one embodiment, before determining the signal detection result of the audio signal to be processed based on the near-Nyquist frequency band energy and mid-low frequency band energy of the audio signal to be processed at a preset sampling rate, the method further includes:

[0026] Based on the frame length information and number of bands of the audio signal to be processed at a preset sampling rate, the target band index identifier of the audio signal to be processed is determined.

[0027] According to the target band index identifier, the audio signal to be processed is divided into near Nyquist frequency band and mid-low frequency band of the audio signal to be processed.

[0028] The near-Nyquist frequency band energy of the audio signal to be processed is determined based on the band energy of the near-Nyquist frequency band, and the mid-low frequency band energy of the audio signal to be processed is determined based on the band energy of the mid-low frequency band.

[0029] Secondly, this application also provides a time-domain noise shaping apparatus, comprising:

[0030] The detection result determination module is used to determine the signal detection result of the audio signal to be processed based on the near Nyquist frequency band energy and mid-low frequency band energy of the audio signal to be processed at a preset sampling rate.

[0031] The decision result determination module is used to determine the shaping start decision result of the high-frequency band spectrum based on the signal detection result and the predicted gain information of the high-frequency band spectrum of the audio signal to be processed;

[0032] The time-domain noise shaping module is used to perform corresponding time-domain noise shaping processing on the high-frequency band spectrum according to the shaping start decision result, so as to obtain the processed high-frequency band spectrum of the audio signal to be processed.

[0033] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0034] The signal detection result of the audio signal to be processed is determined based on the near Nyquist frequency band energy and the mid-low frequency band energy of the audio signal to be processed at a preset sampling rate.

[0035] Based on the signal detection results and the predicted gain information of the high-frequency band spectrum of the audio signal to be processed, the shaping start decision result of the high-frequency band spectrum is determined.

[0036] Based on the shaping start decision result, the high-frequency band spectrum is subjected to corresponding time-domain noise shaping processing to obtain the processed high-frequency band spectrum of the audio signal to be processed.

[0037] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0038] The signal detection result of the audio signal to be processed is determined based on the near Nyquist frequency band energy and the mid-low frequency band energy of the audio signal to be processed at a preset sampling rate.

[0039] Based on the signal detection results and the predicted gain information of the high-frequency band spectrum of the audio signal to be processed, the shaping start decision result of the high-frequency band spectrum is determined.

[0040] Based on the shaping start decision result, the high-frequency band spectrum is subjected to corresponding time-domain noise shaping processing to obtain the processed high-frequency band spectrum of the audio signal to be processed.

[0041] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0042] The signal detection result of the audio signal to be processed is determined based on the near Nyquist frequency band energy and the mid-low frequency band energy of the audio signal to be processed at a preset sampling rate.

[0043] Based on the signal detection results and the predicted gain information of the high-frequency band spectrum of the audio signal to be processed, the shaping start decision result of the high-frequency band spectrum is determined.

[0044] Based on the shaping start decision result, the high-frequency band spectrum is subjected to corresponding time-domain noise shaping processing to obtain the processed high-frequency band spectrum of the audio signal to be processed.

[0045] The aforementioned time-domain noise shaping method, device, terminal, base station, and chip first determine the signal detection result of the audio signal to be processed based on the near-Nyquist frequency band energy and mid-low frequency band energy of the audio signal to be processed at a preset sampling rate. Then, based on the signal detection result and the predicted gain information of the high-frequency band spectrum of the audio signal to be processed, the shaping start decision result of the high-frequency band spectrum is determined. Finally, based on the shaping start decision result, the high-frequency band spectrum is subjected to corresponding time-domain noise shaping processing to obtain the processed high-frequency band spectrum of the audio signal to be processed. In this way, when performing time-domain noise shaping on audio signals, the signal detection result of the audio signal can be more accurately determined by using the near-Nyquist frequency band energy and mid-to-low frequency band energy of the audio signal. This allows for a more accurate determination of the high-frequency band spectrum shaping initiation decision based on the predicted gain information of the high-frequency band spectrum. Furthermore, based on this decision, the corresponding time-domain noise shaping process can be performed more accurately on the high-frequency band spectrum, resulting in a more accurate processed high-frequency band spectrum and thus improving the signal processing quality of time-domain noise shaping. Moreover, since the entire process only processes the high-frequency band spectrum, it avoids the drawback of using a unified high- and low-frequency processing approach, which easily introduces noise and leads to lower signal processing quality in time-domain noise shaping, further improving the overall signal processing quality of time-domain noise shaping. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 This is a flowchart illustrating a temporal noise shaping method in one embodiment;

[0048] Figure 2 This is a flowchart illustrating a time-domain noise shaping method in another embodiment;

[0049] Figure 3 This is a schematic diagram of the input signal in one embodiment;

[0050] Figure 4 This is a schematic diagram of the output signal in one embodiment;

[0051] Figure 5 This is a schematic diagram comparing the performance of the TNS (Temporal Noise Shaping) module with and without TNS enabled at a frame length of 7.5ms in one embodiment.

[0052] Figure 6 This is a schematic diagram comparing the performance of the TNS module being turned off and on under a frame length of 10ms in one embodiment;

[0053] Figure 7 This is a schematic diagram comparing the performance of the LC3 (Low Complexity Communication Codec) standard specification and optimization scheme under a frame length of 7.5ms in one embodiment;

[0054] Figure 8 This is a schematic diagram comparing the performance of the LC3 standard specification and optimization scheme under a frame length of 10ms in one embodiment;

[0055] Figure 9 This is a structural block diagram of a time-domain noise shaping device in one embodiment;

[0056] Figure 10 This is a diagram of the internal structure of a terminal in one embodiment. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0058] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0059] In one exemplary embodiment, such as Figure 1 As shown, a time-domain noise shaping method is provided. This embodiment illustrates the application of this method to a server. It is understood that this method can also be applied to terminals, systems including terminals and servers, and implemented through interaction between the terminal and server. It can also be applied to base stations, chips, and chip modules. The terminal can be, but is not limited to, various personal computers, laptops, smartphones, and tablets; the server can be a standalone server or a server cluster composed of multiple servers. In this embodiment, the method includes the following steps:

[0060] Step S101: Determine the signal detection result of the audio signal to be processed based on the near Nyquist frequency band energy and mid-low frequency band energy of the audio signal to be processed at the preset sampling rate.

[0061] The preset sampling rate refers to a pre-set sampling rate. In practical scenarios, the preset sampling rate refers to a sampling rate of 48kHz or 44.1kHz.

[0062] The audio signal to be processed refers to the audio signal that needs to undergo time-domain noise shaping.

[0063] Among them, the energy of the near-Nyquist frequency band refers to the total energy of the signal components in the high-frequency sub-interval with frequencies close to the Nyquist frequency.

[0064] Among them, the energy of the mid-low frequency band refers to the sum of the energy corresponding to the signal components in the mid-frequency and low-frequency ranges.

[0065] The signal detection result is used to quantify the frequency distribution characteristics of the audio signal to be processed, specifically to identify whether the audio signal contains components close to the Nyquist frequency and with relatively high energy. In practical scenarios, the signal detection result refers to the near-Nyquist detection result, including flag=1 and flag=0.

[0066] For example, the server determines the signal type (analog or digital) of the initial audio signal at a preset sampling rate. If the initial audio signal is an analog signal, it inputs the initial audio signal to an analog-to-digital converter for signal format conversion, obtaining the format-converted initial audio signal, which is used as the audio signal to be processed at the preset sampling rate. If the initial audio signal is a digital signal, the initial audio signal at the preset sampling rate is used as the audio signal to be processed at the preset sampling rate. Next, the server uses a light low-pass filter (such as a resistor-capacitor filter with a cutoff frequency of 1 / 2 of the sampling rate) to remove high-frequency noise (such as ultra-high-frequency noise generated by circuit interference) from the audio signal to be processed, obtaining a denoised audio signal. Then, the server determines the near-Nyquist frequency band energy and mid-low frequency band energy of the denoised audio signal, and determines the signal detection result of the denoised audio signal based on the near-Nyquist frequency band energy and mid-low frequency band energy of the denoised audio signal, which is used as the signal detection result of the audio signal to be processed.

[0067] Step S102: Based on the signal detection results and the predicted gain information of the high-frequency band spectrum of the audio signal to be processed, determine the shaping start decision result of the high-frequency band spectrum.

[0068] The high-frequency band spectrum refers to the set of all frequency spectral coefficients (amplitude and phase) in the audio signal to be processed that are within a preset high-frequency range.

[0069] The prediction gain information refers to the prediction gain value of the high-frequency band spectrum, which is used to represent the predictability of the high-frequency band spectrum, also known as predgain. It should be noted that the larger the prediction gain value of the high-frequency band spectrum, the stronger the predictability of the high-frequency band spectrum, indicating that the high-frequency band signal is closer to the transient signal, and the higher the benefit of starting the shaping process (it can more accurately suppress noise and preserve details).

[0070] Among them, the shaping start decision result is used to represent the decision result of performing time-domain noise shaping on the high-frequency band spectrum, including prohibiting shaping (i.e. closing the subsequent shaping process), dynamic weighted shaping (i.e. starting shaping and performing filter coefficient weighting), and fixed coefficient shaping (i.e. starting shaping and maintaining the filter coefficients).

[0071] For example, the server queries the preset correspondence between the signal detection results, the predicted gain information and the shaping start decision results based on the signal detection results and the predicted gain information of the high-frequency band spectrum of the audio signal to be processed, and obtains the shaping start decision results corresponding to the signal detection results and the predicted gain information, which are used as the shaping start decision results of the high-frequency band spectrum.

[0072] Furthermore, the server can also perform feature extraction processing on the signal detection result and the predicted gain information of the high-frequency band spectrum of the audio signal to be processed, respectively, to obtain the first feature vector of the signal detection result and the second feature vector of the predicted gain information; then, the server fuses the first feature vector and the second feature vector to obtain the fused feature vector; then, the server inputs the fused feature vector into the trained shaping start decision prediction model to obtain the prediction probability of the high-frequency band spectrum under each preset shaping start decision result; then, the server selects the preset shaping start decision result with the highest prediction probability from each preset shaping start decision result as the shaping start decision result of the high-frequency band spectrum.

[0073] It should be noted that, for the low-frequency band spectrum of the audio signal to be processed, the server can determine the shaping start decision result of the low-frequency band spectrum based on the predicted gain information of the low-frequency band spectrum of the audio signal to be processed.

[0074] It should be noted that in traditional techniques, the shaping initiation decision for both the high-frequency and low-frequency spectrum of the audio signal to be processed is determined using the predicted gain information of the spectrum. However, compared to the low-frequency band, the high-frequency band, especially near the Nyquist frequency, is more prone to introducing noise if the signal energy is high. Therefore, this application combines the signal detection results of the audio signal to be processed with the predicted gain information of the high-frequency spectrum. This combination covers the core dimensions of high-frequency shaping decision, avoids misjudgments caused by single information, and ultimately achieves a more accurate decision, enabling a more precise determination of the high-frequency spectrum shaping initiation decision.

[0075] Step S103: Based on the shaping start decision result, perform corresponding time-domain noise shaping on the high-frequency band spectrum to obtain the processed high-frequency band spectrum of the audio signal to be processed.

[0076] Among them, the processed high-frequency band spectrum refers to the high-frequency band spectrum after time-domain noise shaping.

[0077] For example, the server generates a time-domain noise shaping instruction corresponding to the shaping start decision result; then, the server performs corresponding time-domain noise shaping processing on the high-frequency band spectrum according to the time-domain noise shaping instruction to obtain the processed high-frequency band spectrum of the audio signal to be processed.

[0078] In the above-mentioned time-domain noise shaping method, the signal detection result of the audio signal to be processed is first determined based on the near-Nyquist frequency band energy and mid-low frequency band energy of the audio signal to be processed at a preset sampling rate. Then, based on the signal detection result and the predicted gain information of the high-frequency band spectrum of the audio signal to be processed, the shaping start decision result of the high-frequency band spectrum is determined. Finally, based on the shaping start decision result, the high-frequency band spectrum is subjected to corresponding time-domain noise shaping processing to obtain the processed high-frequency band spectrum of the audio signal to be processed. In this way, when performing time-domain noise shaping on audio signals, the signal detection result of the audio signal can be more accurately determined by using the near-Nyquist frequency band energy and mid-to-low frequency band energy of the audio signal. This allows for a more accurate determination of the high-frequency band spectrum shaping initiation decision based on the predicted gain information of the high-frequency band spectrum. Furthermore, based on this decision, the corresponding time-domain noise shaping process can be performed more accurately on the high-frequency band spectrum, resulting in a more accurate processed high-frequency band spectrum and thus improving the signal processing quality of time-domain noise shaping. Moreover, since the entire process only processes the high-frequency band spectrum, it avoids the drawback of using a unified high- and low-frequency processing approach, which easily introduces noise and leads to lower signal processing quality in time-domain noise shaping, further improving the overall signal processing quality of time-domain noise shaping.

[0079] In one exemplary embodiment, the signal detection result includes a first signal detection result and a second signal detection result.

[0080] Therefore, step S102 above, based on the signal detection result and the predicted gain information of the high-frequency band spectrum of the audio signal to be processed, determines the shaping start decision result of the high-frequency band spectrum, specifically including the following: if the signal detection result is a first signal detection result, or if the predicted gain information of the high-frequency band spectrum of the audio signal to be processed meets a first preset condition, the shaping start decision result of the high-frequency band spectrum is determined to be a first shaping start decision result; if the signal detection result is a second signal detection result, and the predicted gain information meets a second preset condition, the shaping start decision result is determined to be a second shaping start decision result; if the signal detection result is a second signal detection result, and the predicted gain information meets a third preset condition, the shaping start decision result is determined to be a third shaping start decision result.

[0081] The first signal detection result, also known as the first near-Nyquist detection result, indicates that the energy of the near-Nyquist frequency band of the audio signal to be processed is significantly higher than the energy of the mid-low frequency band of the audio signal to be processed.

[0082] The second signal detection result, also known as the second near-Nyquist detection result, indicates that the energy of the near-Nyquist frequency band of the audio signal to be processed is significantly higher than that of the mid-low frequency band of the audio signal to be processed.

[0083] The first preset condition refers to the judgment condition of whether the predicted gain value corresponding to the predicted gain information is within the first preset range (e.g., less than 1.5).

[0084] The first ruling initiating plastic surgery proceedings prohibits plastic surgery.

[0085] The second preset condition refers to the judgment condition of whether the predicted gain value corresponding to the predicted gain information is within the second preset range (e.g., greater than or equal to 1.5 and less than 2).

[0086] Among them, the second plastic surgery initiation judgment result refers to dynamic weighted plastic surgery.

[0087] The third preset condition refers to the judgment condition of whether the predicted gain value corresponding to the predicted gain information is within the third preset range (such as greater than or equal to 2).

[0088] Among them, the third plastic surgery initiation decision result refers to fixed coefficient plastic surgery.

[0089] For example, the server simultaneously judges the signal detection result and the prediction gain information of the high-frequency band spectrum of the audio signal to be processed; when the signal detection result is the first signal detection result, or the prediction gain information of the high-frequency band spectrum of the audio signal to be processed satisfies a first preset condition, it is determined that the shaping activation decision result of the high-frequency band spectrum is shaping prohibition; when the signal detection result is the second signal detection result and the prediction gain information satisfies a second preset condition, it is determined that the shaping activation decision result is dynamic weighted shaping; when the signal detection result is the second signal detection result and the prediction gain information satisfies a third preset condition, it is determined that the shaping activation decision result is fixed coefficient shaping.

[0090] For example, when predgain<1.5 or the near-Nyquist flag flag=1, the subsequent TNS process is disabled; when 1.5≤predgain<2 and the near-Nyquist flag flag=0, filter coefficient weighting is performed; when predgain≥2 and the near-Nyquist flag flag=0, the coefficient weighting step is skipped.

[0091] In this embodiment, multi-dimensional combined judgment is performed by combining the signal detection result and the prediction gain information, so that a more accurate shaping activation decision result can be obtained, which is beneficial to improving the determination accuracy of the shaping activation decision result, further enhancing the scientificity and adaptability of high-frequency signal processing, and realizing precise adaptation of the high-frequency shaping strategy.

[0092] In an exemplary embodiment, the above step S103, performing corresponding time-domain noise shaping processing on the high-frequency band spectrum according to the shaping activation decision result to obtain a processed high-frequency band spectrum of the audio signal to be processed, specifically includes the following content: when the shaping activation decision result is the first shaping activation decision result, obtaining the processed high-frequency band spectrum of the audio signal to be processed based on the high-frequency band spectrum; when the shaping activation decision result is the second shaping activation decision result, adjusting the filter coefficients according to the weights of the filter coefficients of the high-frequency band spectrum to obtain adjusted filter coefficients of the high-frequency band spectrum, and performing corresponding time-domain noise shaping processing on the high-frequency band spectrum according to the adjusted filter coefficients to obtain the processed high-frequency band spectrum of the audio signal to be processed; when the shaping activation decision result is the third shaping activation decision result, performing corresponding time-domain noise shaping processing on the high-frequency band spectrum according to the filter coefficients to obtain the processed high-frequency band spectrum of the audio signal to be processed.

[0093] The filter coefficients specifically refer to LP (Linear Prediction) filter coefficients.

[0094] The adjusted filter coefficients refer to the filter coefficients after adjustment based on the weights of the filter coefficients.

[0095] For example, the server determines the shaping start decision result; if the shaping start decision result is the first shaping start decision result, the high-frequency band spectrum is used as the processed high-frequency band spectrum of the audio signal to be processed; if the shaping start decision result is the second shaping start decision result, the server queries the correspondence between the predicted gain information and the weights to obtain the weights corresponding to the predicted gain information, which are used as the weights of the filter coefficients of the high-frequency band spectrum, and adjusts the filter coefficients according to the weights of the filter coefficients of the high-frequency band spectrum to obtain the adjusted filter coefficients of the high-frequency band spectrum. Based on the adjusted filter coefficients, the high-frequency band spectrum is subjected to corresponding time-domain noise shaping processing to obtain the processed high-frequency band spectrum of the audio signal to be processed; if the shaping start decision result is the third shaping start decision result, the high-frequency band spectrum is subjected to corresponding time-domain noise shaping processing according to the filter coefficients to obtain the processed high-frequency band spectrum of the audio signal to be processed.

[0096] In this embodiment, a differentiated high-frequency band spectrum processing strategy is adapted to different shaping start-up decision results. This avoids meaningless computational consumption and enables adaptive processing for different characteristics of high-frequency signals. It effectively balances sound quality optimization, processing efficiency, and signal adaptability, significantly improving the scientific nature of high-frequency band spectrum processing.

[0097] In an exemplary embodiment, step S101, which determines the signal detection result of the audio signal to be processed based on the near-Nyquist frequency band energy and the mid-low frequency band energy of the audio signal to be processed at a preset sampling rate, specifically includes the following: determining the energy ratio between the near-Nyquist frequency band energy and the mid-low frequency band energy of the audio signal to be processed at the preset sampling rate; determining the near-Nyquist detection result of the audio signal to be processed based on the difference between the energy ratio and the preset energy ratio; and determining the signal detection result of the audio signal to be processed based on the near-Nyquist detection result.

[0098] The energy ratio is used to represent the ratio between the energy in the near-Nyquist frequency band and the energy in the mid-to-low frequency band.

[0099] The preset energy ratio refers to a pre-set energy ratio, such as 30. It should be noted that the preset energy ratio depends on the specific circumstances.

[0100] Specifically, the near-Nyquist test result refers to the near-Nyquist identifier.

[0101] For example, the server performs a quotient process on the near-Nyquist frequency band energy and the mid-low frequency band energy of the audio signal to be processed at a preset sampling rate to obtain an energy ratio; then, the server determines the near-Nyquist detection result of the audio signal to be processed based on the difference between the energy ratio and the preset energy ratio; then, the server uses the near-Nyquist detection result as the signal detection result of the audio signal to be processed.

[0102] For example, the near-Nyquist detection result of the audio signal to be processed can be obtained using the following formula:

[0103] Equation (1)

[0104] Where n refers to the band index identifier of the audio signal to be processed. This refers to the number of bands (total number of bands) of the audio signal to be processed. It refers to the band index identifier used to divide the high-frequency band and low-frequency band of the audio signal to be processed. It refers to the energy of the band whose band index is n in the audio signal to be processed.

[0105] In this embodiment, the distribution characteristics of high and low frequency signals can be quickly quantified by energy statistics alone, thereby determining the near-Nyquist characteristics. This enables the accurate capture of energy anomalies in high-frequency components close to the Nyquist frequency. While ensuring the accuracy of near-Nyquist characteristic detection, the processing flow is simplified, providing reliable pre-judgment support for high-frequency optimization of audio signals.

[0106] In an exemplary embodiment, the near-Nyquist detection result of the audio signal to be processed is determined based on the difference between the energy ratio and a preset energy ratio. Specifically, this includes: if the energy ratio is greater than the preset energy ratio, the near-Nyquist detection result of the audio signal to be processed is determined as a first signal detection result; if the energy ratio is less than or equal to the preset energy ratio, the near-Nyquist detection result of the audio signal to be processed is determined as a second signal detection result.

[0107] For example, the server determines the energy ratio based on a preset energy ratio; if the energy ratio is greater than the preset energy ratio, the near-Nyquist detection result of the audio signal to be processed is determined as the first signal detection result; if the energy ratio is less than or equal to the preset energy ratio, the near-Nyquist detection result of the audio signal to be processed is determined as the second signal detection result.

[0108] In this embodiment, the detection results can be output simply by comparing the high and low frequency energy statistics, which greatly reduces the computational complexity and is suitable for scenarios with limited computing power, such as mobile devices and embedded devices. At the same time, it reduces processing latency. While ensuring the specificity of near-Nyquist detection, it also takes into account computational efficiency and scenario adaptability, providing a simple and reliable preliminary judgment basis for high-frequency optimization of audio signals.

[0109] In an exemplary embodiment, step S101, before determining the signal detection result of the audio signal to be processed based on the near-Nyquist frequency band energy and mid-low frequency band energy of the audio signal to be processed at a preset sampling rate, specifically includes the following: determining the target band index identifier of the audio signal to be processed based on the frame length information and number of bands of the audio signal to be processed at a preset sampling rate; dividing the audio signal to be processed according to the target band index identifier to obtain the near-Nyquist frequency band and mid-low frequency band of the audio signal to be processed; determining the near-Nyquist frequency band energy of the audio signal to be processed based on the band energy of the near-Nyquist frequency band, and determining the mid-low frequency band energy of the audio signal to be processed based on the band energy of the mid-low frequency band.

[0110] Among them, frame length information refers to the duration value of discrete data blocks (frames) into which the audio signal to be processed is divided according to the time dimension, such as 7.5ms or 10ms.

[0111] The number of bands is used to represent the total number of bands in the audio signal to be processed, also known as... .

[0112] The target band index identifier refers to the band index identifier used to divide the high-frequency and low-frequency bands of the audio signal to be processed, also known as... .

[0113] Among them, the near-Nyquist frequency band is used to represent the set of all sub-frequency bands (bands) corresponding to the high-frequency sub-intervals in the audio signal to be processed that are close to the Nyquist frequency.

[0114] Among them, the mid-low frequency band is used to represent the set of all sub-frequency bands (bands) corresponding to the mid-frequency and low-frequency ranges in the audio signal to be processed.

[0115] Among them, band energy is used to represent the sum of the squares of the spectral coefficient amplitudes of all frequency points within a single sub-band (band).

[0116] For example, the server inputs the audio signal to be processed at a preset sampling rate into the trained signal feature extraction model to obtain the signal features of the audio signal to be processed at the preset sampling rate, and extracts the frame length information and the number of bands of the audio signal to be processed at the preset sampling rate from the signal features; then, the server queries the correspondence between the frame length information, the number of bands and the band index identifier according to the frame length information and the number of bands of the audio signal to be processed at the preset sampling rate, and obtains the band index identifier corresponding to the frame length information and the number of bands, which is used as the target band index identifier of the audio signal to be processed; for example, when the frame length information is 10ms... When the frame length information is 7.5ms Then, the server divides the audio signal to be processed according to the target band index identifier to obtain the high-frequency band and low-frequency band of the audio signal to be processed. Next, the server uses the band energy of the high-frequency band as the near Nyquist frequency band energy of the audio signal to be processed, and uses the band energy of the low-frequency band as the mid-low frequency band energy of the audio signal to be processed.

[0117] In this embodiment, high and low frequency bands are divided by using the target band index identifier determined by combining the frame length information and the number of bands of the audio signal. Without the need for complex manual matching, the target band can be quickly located by simply filtering the index, which greatly reduces the complexity of operation and helps to improve the efficiency of band division. This lays a data foundation for the high-frequency characteristic analysis and subsequent optimization processing of the audio signal.

[0118] In one exemplary embodiment, such as Figure 2 As shown, another time-domain noise shaping method is provided. Taking the application of this method to a server as an example, the specific steps include:

[0119] Step S201: Determine the target band index identifier of the audio signal to be processed based on the frame length information and number of bands of the audio signal to be processed at the preset sampling rate; divide the audio signal to be processed according to the target band index identifier to obtain the high-frequency band and low-frequency band of the audio signal to be processed.

[0120] Step S202: Determine the near-Nyquist frequency band energy of the audio signal to be processed based on the band energy of the near-Nyquist frequency band, and determine the mid-low frequency band energy of the audio signal to be processed based on the band energy of the mid-low frequency band.

[0121] Step S203: Determine the energy ratio between the near-Nyquist frequency band energy and the mid-low frequency band energy of the audio signal to be processed at the preset sampling rate; determine the near-Nyquist detection result of the audio signal to be processed based on the difference between the energy ratio and the preset energy ratio.

[0122] Step S204: Based on the near-Nyquist detection results, determine the signal detection results of the audio signal to be processed; the signal detection results include the first signal detection result and the second signal detection result.

[0123] Step S205: If the signal detection result is the first signal detection result, or if the predicted gain information of the high-frequency band spectrum of the audio signal to be processed meets the first preset condition, the shaping start decision result of the high-frequency band spectrum is determined as the first shaping start decision result.

[0124] Step S206: If the signal detection result is the second signal detection result and the prediction gain information meets the second preset condition, the shaping start decision result is determined to be the second shaping start decision result.

[0125] Step S207: If the signal detection result is the second signal detection result and the prediction gain information meets the third preset condition, the shaping start decision result is determined to be the third shaping start decision result.

[0126] Step S208: If the shaping start decision result is the first shaping start decision result, the processed high-frequency band spectrum of the audio signal to be processed is obtained based on the high-frequency band spectrum.

[0127] Step S209: If the shaping start decision result is the second shaping start decision result, the filter coefficients are adjusted according to the weight of the filter coefficients of the high-frequency band spectrum to obtain the adjusted filter coefficients of the high-frequency band spectrum. Based on the adjusted filter coefficients, the high-frequency band spectrum is subjected to corresponding time-domain noise shaping processing to obtain the processed high-frequency band spectrum of the audio signal to be processed.

[0128] Step S210: If the shaping start decision result is the third shaping start decision result, perform corresponding time-domain noise shaping processing on the high-frequency band spectrum according to the filter coefficients to obtain the processed high-frequency band spectrum of the audio signal to be processed.

[0129] In the aforementioned time-domain noise shaping method, when performing time-domain noise shaping on audio signals, the signal detection result of the audio signal can be more accurately determined by using the near-Nyquist frequency band energy and mid-to-low frequency band energy of the audio signal. This allows for a more accurate determination of the high-frequency band spectrum shaping initiation decision based on the predicted gain information of the high-frequency band spectrum. Furthermore, based on this decision, the corresponding time-domain noise shaping process can be performed more accurately on the high-frequency band spectrum, resulting in a more accurate processed high-frequency band spectrum and thus improving the signal processing quality of the time-domain noise shaping. Moreover, since the entire process only processes the high-frequency band spectrum, it avoids the drawback of using a unified high- and low-frequency processing approach, which easily introduces noise and leads to lower signal processing quality in time-domain noise shaping, further improving the overall signal processing quality of the time-domain noise shaping.

[0130] In an exemplary embodiment, to more clearly illustrate the time-domain noise shaping method provided by the embodiments of this application, the following specific embodiment will be used to describe the time-domain noise shaping method in detail. In one embodiment, this application also provides a time-domain noise shaping initiation decision design method. Specifically, it includes the following:

[0131] This solution enables near Nyquist detection for a sampling rate of 48kHz (44.1kHz). When the near Nyquist detection flag=1, some TNS filtering functions are disabled to avoid signal disturbances caused by abnormal TNS activation.

[0132] Specifically, the basic process of near-Nyquist detection and TNS implementation is as follows:

[0133] 1. Near-Nyquist detection identifies signals with relatively high energy in the near-Nyquist frequency range by comparing the energy of high-frequency and low-frequency bands.

[0134] Equation (1)

[0135] in =64, To divide the spectrum into The energy of each band is determined by defining the nn_idx~N_B band as the near Nyquist band and the 0~nn_idx band as the mid-low frequency band. By comparing the energy of the two bands, the signal with high energy in the near Nyquist band is selected and flag=1 is set; otherwise, flag=0. This flag is given to the TNS module.

[0136] 2. First, calculate the normalized function value, and perform windowing processing on the normalized function value using the order of the LP filter to obtain the processed normalized function value. Based on the processed normalized function value, estimate the LP filter coefficients and the prediction error using the Levinson-Durbin recursive procedure, and calculate the prediction gain based on the processed normalized function value and the prediction error. A low prediction gain indicates a tendency towards a flat time envelope, while a high prediction gain indicates a very uneven time envelope, meaning that TNS is more required for time domain noise shaping.

[0137] 3. Perform a TNS activation decision according to the value of the prediction error and the near-Nyquist detection result. When predgain<1.5 or the near-Nyquist flag=1, the subsequent TNS process is disabled. When 1.5≤predgain<2 and the near-Nyquist flag=0, filter coefficient weighting is performed. When predgain≥2 and the near-Nyquist flag=0, the coefficient weighting step is skipped. The coefficient weighting step is as follows: when predGain is between 1.5 and 2, TNS is prone to causing artifacts. Therefore, this problem is overcome by reducing filtering in the range of 1.5<predGain<2. By intelligently weighting the original filter coefficients, an intermediate state between full filtering and no filtering can be obtained. This intermediate state can reduce the disadvantages of TNS while maintaining its positive characteristics.

[0138] 4. Quantize the coefficients in the arcsine domain, encode the quantization index into the bitstream, and feed the quantized coefficients to the subsequent filtering module.

[0139] 5. The input spectrum is filtered by an FIR (Finite Impulse Response) filter and then output.

[0140] In the existing LC3 standard specification, at a sampling rate of 48kHz (44.1kHz), the signals input to the TNS and near-Nyquist detection modules are processed by the pre-module, so the effective frequency band is only 20kHz, and there is no signal with relatively high energy near the Nyquist frequency (24kHz). Thus, it is considered unnecessary to enable near-Nyquist detection, and the TNS module is enabled for all input signals. However, according to Figure 3 and Figure 4 the simulation results shown, when a sine wave signal with an input frequency close to 20kHz is filtered by TNS, signal disturbance will also occur, leading to degradation of codec performance.

[0141] A sine wave signal with a frequency close to 20kHz is used to test the effect of TNS filtering on codec performance. The effects of the TNS module being disabled and enabled on codec performance under different bit rates and frame lengths are as shown in Figure 5 and Figure 6Enabling TNS causes a significant performance drop at low bit rates (96kbps), with a maximum drop of 20dB. However, TNS can still function at high bit rates (>160kbps), and performance decreases when TNS is disabled.

[0142] Firstly, to address the issue of signal disturbance caused by TNS filtering signals with relatively high energy around 20kHz at a low bit rate of 48kHz (44.1kHz), this patent proposes enabling near Nyquist detection at a sampling rate of 48kHz (44.1kHz). Since the effective frequency band of the input signal spectrum is 20kHz, signals with concentrated energy around 20kHz will also satisfy the condition in formula (1), making the near Nyquist detection flag=1. Input signals that may cause disturbance can be filtered based on the detection result, so as to optimize the TNS calculation process for such signals.

[0143] Furthermore, when the near-Nyquist detection flag=1, this patent proposes to disable only the TNS's high-frequency filtering function, eliminating signal disturbances while ensuring no degradation in encoding and decoding performance at high bit rates. For signals with energy concentrated around 20kHz at a sampling rate of 48kHz (44.1kHz), the bandwidth is 0 to 20kHz. For this type of signal, TNS divides the frequency band into two segments and performs the analysis and filtering processes described above separately: first, it analyzes and filters the first filter coefficients for the 12th-200th frequency points (low frequencies); then, it analyzes and filters the second filter coefficients for the 200th-400th frequency points (high frequencies). Since the analysis and filtering processes of the two filters are completely independent, when the Nyquist flag=1, the main components of the signal are located within the effective frequency range of the second filter. The simulation results of the SNR (Signal to Noise Ratio) output of the encoder and decoder under different on and off conditions of the first and second filters are shown in Table 1 below (encoding and decoding parameters: 7.5ms frame length, 96kbps bit rate, 48kHz sampling rate, input 18.5kHz sine wave).

[0144] Table 1. Simulation results of SNR output for encoding and decoding under different on / off conditions of the first and second filters.

[0145]

[0146] The following conclusions can be drawn: For signals with main frequency components close to 20kHz, the first filter for low frequencies will not cause signal disturbance and will still have the function of suppressing quantization noise and improving codec performance. However, the second filter for high frequencies will cause signal disturbance, resulting in a decrease in codec performance.

[0147] This patent proposes an optimized solution, and its performance is compared with existing LC3 standards at different bit rates and frame lengths, for example... Figure 7 and Figure 8 Performance is generally improved across all bit rates, with a more significant improvement at low bit rates, up to 20dB (frame length 7.5ms, bit rate 96kbps).

[0148] The above optimization scheme, without adding a completely new detection module and detection identifier, and without affecting the encoded bitstream structure, uses the original near-Nyquist detection module to detect whether the TNS second filter needs to be turned off, while retaining the first filter that can suppress quantization noise. It achieves better performance than simply turning off all TNS functions, and has good compatibility with the original encoding and decoding system. It does not require classification processing for bit rate and frame length, and the performance optimization effect is effective for any frame length and bit rate, and has extremely high practical value.

[0149] In the above embodiments, when performing time-domain noise shaping on audio signals, the signal detection result of the audio signal can be more accurately determined by using the near-Nyquist frequency band energy and mid-to-low frequency band energy of the audio signal. This allows for a more accurate determination of the high-frequency band spectrum shaping initiation decision based on the predicted gain information of the high-frequency band spectrum. Furthermore, based on this decision, the corresponding time-domain noise shaping process can be performed more accurately on the high-frequency band spectrum, resulting in a more accurate processed high-frequency band spectrum and thus improving the signal processing quality of the time-domain noise shaping. Moreover, since the entire process only processes the high-frequency band spectrum, it avoids the drawback of using a unified high- and low-frequency processing method, which easily introduces noise and leads to lower signal processing quality in time-domain noise shaping, further improving the overall signal processing quality of time-domain noise shaping.

[0150] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0151] Based on the same inventive concept, this application also provides a time-domain noise shaping apparatus for implementing the time-domain noise shaping method described above. The solution provided by this apparatus is similar to the implementation described in the above method; therefore, the specific limitations in one or more time-domain noise shaping apparatus embodiments provided below can be found in the limitations of the time-domain noise shaping method described above, and will not be repeated here.

[0152] In one exemplary embodiment, such as Figure 9 As shown, a time-domain noise shaping device is provided, comprising: a detection result determination module 901, a decision result determination module 902, and a time-domain noise shaping module 903, wherein:

[0153] The detection result determination module 901 is used to determine the signal detection result of the audio signal to be processed based on the near Nyquist frequency band energy and mid-low frequency band energy of the audio signal to be processed at a preset sampling rate.

[0154] The decision determination module 902 is used to determine the shaping start decision result of the high-frequency band spectrum based on the signal detection result and the predicted gain information of the high-frequency band spectrum of the audio signal to be processed.

[0155] The time-domain noise shaping module 903 is used to perform corresponding time-domain noise shaping on the high-frequency band spectrum according to the shaping start decision result, so as to obtain the processed high-frequency band spectrum of the audio signal to be processed.

[0156] In an exemplary embodiment, the decision result determination module 902 is further configured to: determine the shaping start decision result of the high-frequency band spectrum as the first shaping start decision result when the signal detection result is the first signal detection result or the predicted gain information of the high-frequency band spectrum of the audio signal to be processed meets the first preset condition; determine the shaping start decision result as the second shaping start decision result when the signal detection result is the second signal detection result and the predicted gain information meets the second preset condition; and determine the shaping start decision result as the third shaping start decision result when the signal detection result is the second signal detection result and the predicted gain information meets the third preset condition.

[0157] In an exemplary embodiment, the time-domain noise shaping module 903 is further configured to: when the shaping start decision result is a first shaping start decision result, obtain the processed high-frequency band spectrum of the audio signal to be processed based on the high-frequency band spectrum; when the shaping start decision result is a second shaping start decision result, adjust the filter coefficients according to the weights of the filter coefficients of the high-frequency band spectrum to obtain the adjusted filter coefficients of the high-frequency band spectrum, and perform corresponding time-domain noise shaping processing on the high-frequency band spectrum according to the adjusted filter coefficients to obtain the processed high-frequency band spectrum of the audio signal to be processed; when the shaping start decision result is a third shaping start decision result, perform corresponding time-domain noise shaping processing on the high-frequency band spectrum according to the filter coefficients to obtain the processed high-frequency band spectrum of the audio signal to be processed.

[0158] In an exemplary embodiment, the detection result determination module 901 is further configured to determine the energy ratio between the near-Nyquist frequency band energy and the mid-low frequency band energy of the audio signal to be processed at a preset sampling rate; determine the near-Nyquist detection result of the audio signal to be processed based on the difference between the energy ratio and the preset energy ratio; and determine the signal detection result of the audio signal to be processed based on the near-Nyquist detection result.

[0159] In an exemplary embodiment, the detection result determination module 901 is further configured to determine the near-Nyquist detection result of the audio signal to be processed as a first signal detection result when the energy ratio is greater than a preset energy ratio; and to determine the near-Nyquist detection result of the audio signal to be processed as a second signal detection result when the energy ratio is less than or equal to the preset energy ratio.

[0160] In an exemplary embodiment, the time-domain noise shaping device further includes a band energy determination module, configured to determine the target band index identifier of the audio signal to be processed based on the frame length information and the number of bands of the audio signal to be processed at a preset sampling rate; divide the audio signal to be processed according to the target band index identifier to obtain the near-Nyquist frequency band and the mid-low frequency band of the audio signal to be processed; determine the near-Nyquist frequency band energy of the audio signal to be processed based on the band energy of the near-Nyquist frequency band, and determine the mid-low frequency band energy of the audio signal to be processed based on the band energy of the mid-low frequency band.

[0161] Each module in the aforementioned time-domain noise shaping device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0162] In one exemplary embodiment, a terminal is provided, the internal structure of which can be as follows: Figure 10 As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computational and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage medium. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a time-domain noise shaping method.

[0163] Those skilled in the art will understand that Figure 10 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0164] This embodiment also provides a base station, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described time-domain noise shaping method.

[0165] This embodiment also provides a chip, which includes at least one processor. When the processor executes a computer program, it implements the steps of the above-described time-domain noise shaping method.

[0166] This embodiment also provides a chip module coupled to a memory, used to implement the steps of the above-described time-domain noise shaping method when executing a computer program stored in the memory.

[0167] This embodiment also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the above-described time-domain noise shaping method.

[0168] This embodiment also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described time-domain noise shaping method.

[0169] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to mutually. In addition, the different parts between embodiments can also be combined with each other, and this invention does not limit this.

[0170] The temporal noise shaping method, apparatus, terminal, base station, and chip provided in this embodiment include: determining the signal detection result of the audio signal to be processed based on the near-Nyquist frequency band energy and mid-low frequency band energy of the audio signal to be processed at a preset sampling rate; determining the shaping start decision result of the high-frequency band spectrum based on the signal detection result and the predicted gain information of the high-frequency band spectrum of the audio signal to be processed; and performing corresponding temporal noise shaping processing on the high-frequency band spectrum based on the shaping start decision result to obtain the processed high-frequency band spectrum of the audio signal to be processed. When performing time-domain noise shaping on audio signals, the signal detection result can be more accurately determined by using the near-Nyquist frequency band energy and mid-to-low frequency band energy of the audio signal. This allows for a more accurate determination of the high-frequency band spectrum's shaping initiation decision, which in turn enables more precise time-domain noise shaping of the high-frequency band spectrum. This results in a more accurate processed high-frequency band spectrum, thus improving the signal processing quality of time-domain noise shaping. Furthermore, since the entire process only processes the high-frequency band spectrum, it avoids the problem of introducing noise and resulting in lower signal processing quality when using a unified high- and low-frequency processing approach, further enhancing the overall signal processing quality of time-domain noise shaping.

[0171] Regarding the modules / units included in the various devices and products described in the above embodiments, they can be software modules / units, hardware modules / units, or a combination of both. For example, for various devices and products applied to or integrated into a chip, all of their modules / units can be implemented using hardware methods such as circuits, or at least some modules / units can be implemented using software programs that run on a processor integrated within the chip, while the remaining (if any) modules / units can be implemented using hardware methods such as circuits; for various devices and products applied to or integrated into a chip module, all of their modules / units can be implemented using hardware methods such as circuits, and different modules / units can be located in the same component (e.g., chip, circuit module, etc.) or different components of the chip module, or at least some modules / units can be implemented using hardware methods such as circuits. The components can be implemented using software programs that run on the processor integrated within the chip module. The remaining (if any) modules / units can be implemented using hardware methods such as circuits. For various devices and products applied to or integrated into the terminal, each of its components / units can be implemented using hardware methods such as circuits. Different modules / units can be located in the same component (e.g., chip, circuit module, etc.) or in different components within the terminal. Alternatively, at least some modules / units can be implemented using software programs that run on the processor integrated within the terminal, while the remaining (if any) modules / units can be implemented using hardware methods such as circuits.

[0172] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0173] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0174] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A time-domain noise shaping method, characterized in that, The method includes: Based on the near-Nyquist frequency band energy and mid-low frequency band energy of the audio signal to be processed at a preset sampling rate, the signal detection result of the audio signal to be processed is determined; the signal detection result includes a first signal detection result and a second signal detection result; the first signal detection result is used to indicate that the near-Nyquist frequency band energy is detected to be higher than the mid-low frequency band energy; the second signal detection result is used to indicate that the near-Nyquist frequency band energy is not detected to be higher than the mid-low frequency band energy; If the signal detection result is the first signal detection result, or if the predicted gain information of the high-frequency band spectrum of the audio signal to be processed meets the first preset condition, the shaping start decision result of the high-frequency band spectrum is determined as the first shaping start decision result; if the signal detection result is the second signal detection result, and the predicted gain information meets the second preset condition, the shaping start decision result is determined as the second shaping start decision result; if the signal detection result is the second signal detection result, and the predicted gain information meets the third preset condition, the shaping start decision result is determined as the third shaping start decision result. When the shaping start decision result is the first shaping start decision result, the processed high-frequency band spectrum of the audio signal to be processed is obtained based on the high-frequency band spectrum; when the shaping start decision result is the second shaping start decision result, the filter coefficients are adjusted according to the weights of the filter coefficients of the high-frequency band spectrum to obtain the adjusted filter coefficients of the high-frequency band spectrum; based on the adjusted filter coefficients, the high-frequency band spectrum is subjected to corresponding time-domain noise shaping processing to obtain the processed high-frequency band spectrum of the audio signal to be processed; when the shaping start decision result is the third shaping start decision result, the high-frequency band spectrum is subjected to corresponding time-domain noise shaping processing according to the filter coefficients to obtain the processed high-frequency band spectrum of the audio signal to be processed.

2. The method according to claim 1, characterized in that, The step of determining the signal detection result of the audio signal to be processed based on the near-Nyquist frequency band energy and mid-low frequency band energy of the audio signal to be processed at a preset sampling rate includes: Determine the energy ratio between the near-Nyquist frequency band energy and the mid-low frequency band energy of the audio signal to be processed at the preset sampling rate; The near-Nyquist detection result of the audio signal to be processed is determined based on the difference between the energy ratio and the preset energy ratio. Based on the near-Nyquist detection results, the signal detection results of the audio signal to be processed are determined.

3. The method according to claim 2, characterized in that, The step of determining the near-Nyquist detection result of the audio signal to be processed based on the difference between the energy ratio and the preset energy ratio includes: If the energy ratio is greater than the preset energy ratio, the near-Nyquist detection result of the audio signal to be processed is determined as the first signal detection result; If the energy ratio is less than or equal to the preset energy ratio, the near-Nyquist detection result of the audio signal to be processed is determined as the second signal detection result.

4. The method according to any one of claims 1 to 3, characterized in that, Before determining the signal detection result of the audio signal to be processed based on the near-Nyquist frequency band energy and mid-low frequency band energy of the audio signal to be processed at a preset sampling rate, the process further includes: Based on the frame length information and number of bands of the audio signal to be processed at a preset sampling rate, the target band index identifier of the audio signal to be processed is determined. According to the target band index identifier, the audio signal to be processed is divided into near Nyquist frequency band and mid-low frequency band of the audio signal to be processed. The near-Nyquist frequency band energy of the audio signal to be processed is determined based on the band energy of the near-Nyquist frequency band, and the mid-low frequency band energy of the audio signal to be processed is determined based on the band energy of the mid-low frequency band.

5. A time-domain noise shaping device, characterized in that, The device includes: The detection result determination module is used to determine the signal detection result of the audio signal to be processed based on the near-Nyquist frequency band energy and the mid-low frequency band energy of the audio signal to be processed at a preset sampling rate; the signal detection result includes a first signal detection result and a second signal detection result; the first signal detection result is used to indicate that the near-Nyquist frequency band energy is detected to be higher than the mid-low frequency band energy; the second signal detection result is used to indicate that the near-Nyquist frequency band energy is not detected to be higher than the mid-low frequency band energy; The decision result determination module is configured to: determine the shaping start decision result of the high-frequency band spectrum as a first shaping start decision result when the signal detection result is the first signal detection result, or when the predicted gain information of the high-frequency band spectrum of the audio signal to be processed meets a first preset condition; determine the shaping start decision result as a second shaping start decision result when the signal detection result is the second signal detection result and the predicted gain information meets a second preset condition; and determine the shaping start decision result as a third shaping start decision result when the signal detection result is the second signal detection result and the predicted gain information meets a third preset condition. The time-domain noise shaping module is used to: obtain the processed high-frequency band spectrum of the audio signal to be processed based on the high-frequency band spectrum when the shaping start decision result is the first shaping start decision result; adjust the filter coefficients according to the weights of the filter coefficients of the high-frequency band spectrum to obtain the adjusted filter coefficients of the high-frequency band spectrum; perform corresponding time-domain noise shaping processing on the high-frequency band spectrum according to the adjusted filter coefficients to obtain the processed high-frequency band spectrum of the audio signal to be processed; and perform corresponding time-domain noise shaping processing on the high-frequency band spectrum according to the filter coefficients when the shaping start decision result is the third shaping start decision result to obtain the processed high-frequency band spectrum of the audio signal to be processed.

6. The apparatus according to claim 5, characterized in that, The detection result determination module is further used to determine the energy ratio between the near-Nyquist frequency band energy and the mid-low frequency band energy of the audio signal to be processed at a preset sampling rate; and to determine the near-Nyquist detection result of the audio signal to be processed based on the difference between the energy ratio and the preset energy ratio. Based on the near-Nyquist detection results, the signal detection results of the audio signal to be processed are determined.

7. The apparatus according to claim 6, characterized in that, The detection result determination module is further configured to determine the near Nyquist detection result of the audio signal to be processed as the first signal detection result when the energy ratio is greater than the preset energy ratio. If the energy ratio is less than or equal to the preset energy ratio, the near-Nyquist detection result of the audio signal to be processed is determined as the second signal detection result.

8. A terminal comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 4.

9. A base station, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 4.

10. A chip, characterized in that, The chip includes at least one processor, which executes a computer program to implement the steps of the method according to any one of claims 1 to 4.

Images