WO2024082800A1 - 音频处理方法、装置及终端设备 - Google Patents

音频处理方法、装置及终端设备 Download PDF

Info

Publication number
WO2024082800A1
WO2024082800A1 PCT/CN2023/113597 CN2023113597W WO2024082800A1 WO 2024082800 A1 WO2024082800 A1 WO 2024082800A1 CN 2023113597 W CN2023113597 W CN 2023113597W WO 2024082800 A1 WO2024082800 A1 WO 2024082800A1
Authority
WO
WIPO (PCT)
Prior art keywords
audio
audios
target
weights
determining
Prior art date
Application number
PCT/CN2023/113597
Other languages
English (en)
French (fr)
Inventor
尚楚翔
向肖肖
赵成帅
黄传增
Original Assignee
抖音视界有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 抖音视界有限公司 filed Critical 抖音视界有限公司
Publication of WO2024082800A1 publication Critical patent/WO2024082800A1/zh

Links

Classifications

    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
    • G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
    • G10L21/0208—Noise filtering
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
    • G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
    • G10L21/0208—Noise filtering
    • G10L21/0216—Noise filtering characterised by the method used for estimating noise
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
    • G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
    • G10L21/0208—Noise filtering
    • G10L21/0216—Noise filtering characterised by the method used for estimating noise
    • G10L21/0232—Processing in the frequency domain
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
    • G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
    • G10L21/0208—Noise filtering
    • G10L21/0264—Noise filtering characterised by the type of parameter measurement, e.g. correlation techniques, zero crossing techniques or predictive techniques
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/03—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
    • G10L25/18—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters the extracted parameters being spectral information of each sub-band
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/27—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique
    • G10L25/30—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique using neural networks

Definitions

  • the embodiments of the present disclosure relate to the field of audio processing technology, and in particular, to an audio processing method, apparatus, and terminal device.
  • the terminal device can collect audio in the space and play the audio.
  • the microphone in the terminal device can collect the user's voice and send the voice to other terminal devices.
  • a terminal device can collect audio in a space through a microphone, and send the collected audio to other terminal devices so that other terminal devices can play the audio.
  • a microphone in a conference room can collect the user's voice, and a terminal device can obtain the voice from the microphone and send the voice to other terminal devices so that other terminal devices can play the voice, thereby achieving the effect of a remote conference call.
  • target audio and non-target audio for example, audio associated with non-target objects or interfering audio
  • the present invention provides an audio processing method, an apparatus and a terminal device, which are used to solve the technical problem of poor audio playback effect in the prior art.
  • the present disclosure provides an audio processing method, the method comprising:
  • a second audio associated with the target direction is determined, and the second audio is played.
  • the present disclosure provides an audio processing device, the audio processing device comprising an acquisition module, a first determination module, a second determination module and a playback module, wherein:
  • the acquisition module is used to acquire multiple first audios collected by multiple audio collection devices
  • the first determination module is used to determine an angle feature based on the multiple first audios and the target direction, where the angle feature is used to indicate the proportion of the sound source in the target direction in each first audio;
  • the second determination module is used to determine the second audio frequency associated with the target direction based on the multiple first audio frequencies and the angle feature;
  • the playing module is used for playing the second audio.
  • an embodiment of the present disclosure provides a terminal device, including: a processor and a memory;
  • the memory stores computer-executable instructions
  • the processor executes the computer-executable instructions stored in the memory, so that the at least one processor executes the audio processing method as described in the first aspect and various possible aspects of the first aspect.
  • an embodiment of the present disclosure provides a computer-readable storage medium, in which computer execution instructions are stored.
  • a processor executes the computer execution instructions, the audio processing method as described in the first aspect and various possible aspects of the first aspect are implemented.
  • an embodiment of the present disclosure provides a computer program product, including a computer program, which, when executed by a processor, implements the audio processing method as described in the first aspect and various possible aspects of the first aspect.
  • an embodiment of the present disclosure provides a computer program, which, when executed by a processor, implements the audio processing method as described in the first aspect above.
  • FIG1 is a schematic diagram of an application scenario provided by an embodiment of the present disclosure.
  • FIG2 is a schematic flow chart of an audio processing method provided by an embodiment of the present disclosure.
  • FIG3 is a schematic diagram of a process for determining a plurality of first phase differences provided by an embodiment of the present disclosure
  • FIG4 is a schematic diagram of a method for determining a second audio according to an embodiment of the present disclosure
  • FIG5 is a schematic diagram of a process of determining a first weight and a second weight provided by an embodiment of the present disclosure
  • FIG6 is a schematic diagram of a process for determining a second audio according to an embodiment of the present disclosure
  • FIG7 is a schematic diagram of a process of an audio processing method provided by an embodiment of the present disclosure.
  • FIG8 is a schematic diagram of the structure of an audio processing device provided by an embodiment of the present disclosure.
  • FIG. 9 is a schematic diagram of the structure of a terminal device provided in an embodiment of the present disclosure.
  • Terminal device a device with wireless transceiver function. Terminal devices can be deployed on land, including indoors or outdoors, handheld, wearable or vehicle-mounted; they can also be deployed on the water (such as ships, etc.).
  • the terminal device can be a mobile phone, a portable Android device (PAD), a computer with wireless transceiver function, a virtual reality (VR) terminal device, an augmented reality (AR) terminal device, a wireless terminal in industrial control, a vehicle-mounted terminal device, a wireless terminal in self-driving, a wireless terminal device in remote medical, a wireless terminal device in smart grid, transportation safety, etc.
  • PAD portable Android device
  • VR virtual reality
  • AR augmented reality
  • the terminal devices involved in the embodiments of the present disclosure may also be referred to as terminals, user equipment (UE), access terminal devices, vehicle-mounted terminals, industrial control terminals, UE units, UE stations, mobile stations, mobile stations, remote stations, remote terminal devices, mobile devices, UE terminal devices, wireless communication devices, UE agents or UE devices, etc.
  • the terminal devices may also be fixed or mobile.
  • a terminal device can collect audio in a space through a microphone, and send the collected audio to other terminal devices so that other terminal devices can play the audio.
  • the microphone in the conference room can collect the voices of the users participating in the conference
  • the terminal device in the conference room can obtain the voices of the users participating in the conference from the microphone, and send the voices to the remotely connected terminal device so that the remotely connected terminal device can play the voices of the users participating in the conference.
  • the noise of the audio collected by the microphone is relatively large, which leads to poor audio playback effect.
  • the embodiment of the present disclosure provides an audio processing method, wherein a terminal device obtains multiple first audios collected by multiple audio collection devices, determines an angle feature based on the multiple first audios and the target direction, wherein the angle feature is used to indicate the proportion of the sound source in the target direction in each first audio, determines the first target audio associated with each first audio in the target direction based on the multiple first audios and the angle feature, and determines the non-target audio associated with each first audio in the target direction, determines the second audio based on the multiple first audios, the multiple first target audios and the non-target audios, and plays the second audio.
  • the terminal device can enhance the first target audio effect in the first audio, suppress the non-target audio in the first audio, thereby reducing the noise in the second audio and improving the playback effect of the second audio.
  • FIG1 is a schematic diagram of an application scenario provided by an embodiment of the present disclosure.
  • the user sends voice A to the audio acquisition device, and the audio acquisition device can also collect noise A and noise B in the environment.
  • the audio acquisition device sends a mixed audio of voice A, noise A, and noise B to the terminal device, and the terminal device can determine the angle feature of the mixed audio, and determine the user's voice A based on the mixed audio and the angle feature, and play voice A.
  • the terminal device can amplify the signal of voice A, suppress the signals of noise A and noise B, and thereby improve the audio playback effect.
  • FIG. 1 is only an illustrative example of an application scenario of an embodiment of the present disclosure, and is not intended to limit the application scenario of the embodiment of the present disclosure.
  • FIG2 is a flow chart of an audio processing method provided by an embodiment of the present disclosure. Referring to FIG2 , the method may include:
  • the execution subject of the embodiment of the present disclosure is a terminal device, or an audio processing device set in the terminal device.
  • the audio processing device can be implemented by software, or by a combination of software and hardware, which is not limited in the embodiment of the present disclosure.
  • the audio acquisition device is used to collect audio in the space.
  • the audio acquisition device may be a microphone.
  • the audio acquisition device may be a microphone in a conference room, or the audio acquisition device may be a microphone in a terminal device.
  • the embodiments of the present disclosure are not limited to this.
  • the first audio is audio captured by an audio capture device.
  • multiple microphones may be set in a conference room, and the first audio may be audio captured by multiple microphones.
  • the microphone in the conference room may capture the user's voice and determine the user's voice as the first audio.
  • the terminal device may obtain multiple first audios through an audio collection device.
  • the terminal device may be connected to multiple audio collection devices, and when the multiple audio collection devices obtain multiple first audios, the multiple first audios may be sent to the terminal device.
  • the terminal device may obtain multiple first audios through other terminal devices.
  • the terminal device may be connected to other terminal devices for communication, and when a microphone in the other terminal device obtains multiple first audios, the other terminal device may send the multiple first audios to the terminal device.
  • one audio acquisition device can acquire one first audio. For example, when a space includes one sound source, if the space includes one audio acquisition device, then one audio acquisition device can acquire one first audio for the sound source, and if the space includes six audio acquisition devices, then six audio acquisition devices can acquire six first audios for the sound source.
  • S202 Determine an angle feature based on multiple first audios and target directions.
  • the angle feature is used to indicate the proportion of the sound source in the target direction in each first audio.
  • the angle feature is used to indicate the proportion of the audio associated with the target direction in the first audio A and the proportion of the audio associated with the target direction in the first audio B.
  • the terminal device can determine the angle feature based on the following feasible implementation method: determine the phase difference between multiple first audios to obtain multiple first phase differences, determine the second phase difference associated with the target direction, and determine the angle feature based on the second phase difference and multiple first phase differences.
  • determining the phase difference between multiple first audios specifically: the terminal device determines the initial audio from the multiple first audios.
  • the terminal device may determine any one of the multiple first audios as the initial audio. For example, if the multiple first audios acquired by the terminal device include first audio A, first audio B, and first audio C, the terminal device may determine first audio A as the initial audio, the terminal device may determine first audio B as the initial audio, and the terminal device may also determine first audio C as the initial audio.
  • the phase difference between the initial audio and each first audio is obtained to obtain multiple first phase differences.
  • the multiple first audios collected by the terminal device include the first audio A, the first audio B, and the first audio C, and the terminal device determines that the initial audio is the first audio A, then the terminal device can obtain the phase difference A between the first audio A and the first audio B, and the terminal device can obtain the phase difference B between the first audio A and the first audio C, and then determine the phase difference A and the phase difference B as the first phase difference.
  • the terminal device when the terminal device determines the angle feature through multiple first audios and target directions, the terminal device can first perform Fourier transform on the multiple first audios, and then obtain multiple spectrum graphs associated with the multiple first audios, and then determine multiple first phase differences through the multiple spectrum graphs. This can improve the accuracy of determining the first phase difference, thereby obtaining accurate angle features and improving the effect of audio processing.
  • FIG3 is a schematic diagram of a process for determining multiple first phase differences provided by an embodiment of the present disclosure.
  • a first audio A, a first audio B, and a first audio C are included.
  • the first audio A, the first audio B, and the first audio C are subjected to Fourier transform processing to obtain a spectrum A associated with the first audio A, a spectrum B associated with the first audio B, and a spectrum C associated with the first audio C.
  • the first audio A is determined as the initial audio
  • the phase difference between the spectrum A and the spectrum B is determined as the first phase difference A
  • the phase difference between the spectrum A and the spectrum C is determined as the first phase difference B.
  • the terminal device may determine the phase difference between the sound source in the target direction and the audio collection device as the second phase difference associated with the target direction.
  • the second phase difference associated with the target direction is a phase difference caused by the distance difference between the sound source in the target direction and the microphone, and the phase difference is the target phase difference.
  • the angle feature is determined based on the second phase difference and the multiple first phase differences, specifically: the cosine similarity between the second phase difference and each first phase difference is determined to obtain multiple cosine similarities.
  • the second phase difference associated with the target direction is phase difference A
  • the first phase difference includes phase difference B and phase difference C
  • the multiple cosine similarities include the cosine similarity between phase difference A and phase difference B and the cosine similarity between phase difference A and phase difference C.
  • the terminal device may concatenate multiple cosine similarities to obtain an angle feature, or may superimpose multiple cosine similarities to obtain an angle feature, which is not limited in the embodiments of the present disclosure.
  • the terminal device may concatenate cosine similarity A and cosine similarity B to obtain an angle feature.
  • the terminal device may determine the target direction in response to the user's touch operation. For example, the terminal device may determine the direction in space corresponding to the user's touch position as the target direction based on the user's touch operation on any position in the video conference display screen.
  • the terminal device may determine the target direction in response to the user's voice operation. For example, when user A issues a voice message "Please let user B speak", the terminal device may determine the location of user B as the target direction. It should be noted that the terminal device may also determine the target direction in other ways, which is not limited in the embodiments of the present disclosure.
  • S203 Determine a second audio frequency associated with the target direction based on the multiple first audio frequencies and angle features.
  • the second audio is the audio acquired by the terminal device in the target direction.
  • the second audio is the audio of the sound source in the east direction.
  • the terminal device may determine the second audio associated with the target direction based on the following feasible implementation methods: determine multiple first target audios and non-target audios based on multiple first audios and angle features, and determine the second audio based on multiple first audios, multiple first target audios, and non-target audios.
  • the first target audio is the audio associated with the first audio in the target direction
  • the non-target audio is the audio associated with the first audio in other directions.
  • the angle feature can indicate the proportion of the sound source in the target direction in each first audio
  • the first target audio of each first audio in the target direction and the non-target audio (audio in other directions) of each first audio in the target direction can be obtained through the angle feature and multiple first audios.
  • the terminal device performs fusion processing on multiple sub-audios to obtain a second audio, and then can play the second audio.
  • the terminal device may also send the second audio to other terminal devices or a server, which is not limited in the embodiments of the present disclosure.
  • the disclosed embodiment provides an audio processing method, in which a terminal device obtains multiple first audios collected by multiple audio collection devices, determines angle features based on the multiple first audios and target directions, determines first target audios associated with each first audio in the target direction based on the multiple first audios and the angle features, and determines non-target audios associated with each first audio in the target direction, processes the multiple first target audios and non-target audios based on a second model to obtain multiple target weights, determines multiple sub-audios based on the multiple target weights and the multiple first audios, fuses the multiple sub-audios to obtain a second audio, and plays the second audio.
  • the terminal device can enhance the first target audio effect in the first audio, suppress the non-target audio in the first audio, and then reduce the noise in the second audio and improve the playback effect of the second audio.
  • FIG4 is a schematic diagram of a method for determining a second audio according to an embodiment of the present disclosure. Referring to FIG4 , the method flow includes:
  • S401 Determine a plurality of first target audios and non-target audios based on a plurality of first audios and angle features.
  • the first target audio is the audio associated with the first audio in the target direction
  • the non-target audio is the audio associated with the first audio in other directions.
  • each first audio includes a first target audio and a non-target audio.
  • the terminal device can obtain 6 first target audios and 6 non-target audios.
  • the terminal device may determine multiple first target audios and non-target audios based on the following feasible implementations: based on multiple first audios and angle features, determine multiple first weights and multiple second weights associated with multiple first audios, and based on multiple first weights, multiple second weights and multiple first audios, determine multiple first target audios and non-target audios.
  • the first weight may be the weight of the audio in the target direction in the first audio
  • the second weight may be the weight of the audio in other directions in the first audio.
  • multiple first weights and multiple second weights associated with multiple first audio are determined, specifically: based on the first model, multiple first audio and angle features, the first weights and the second weights are determined.
  • the first model is obtained by training multiple groups of first samples, and the multiple groups of first samples include multiple sample audios, angle features associated with multiple sample audios, and sample first weights and sample second weights associated with each sample audio. For example, obtain sample audio 1, angle feature 1 associated with sample audio 1, and sample first weight 1 and sample second weight 1 associated with each sample audio, wherein the group of first samples includes sample audio 1, sample angle feature 1, sample first weight 1, and sample second weight 1. Multiple groups of first samples can be obtained by using this method.
  • the structure of the first model may include an encoder, a dual path recurrent neural network (DPRNN for short) and a decoder.
  • DPRNN can effectively model the correlation between frequencies and the correlation between frames of audio, and thus can effectively decompose the audio.
  • the encoder is used to encode angle features and features of multiple spectrograms.
  • DPRNN can be composed of stacked SA-DPRNN blocks and adaptive feature fusion blocks (AFFB for short).
  • the SA-DPRNN block includes an intra-block recurrent neural network (RNN for short) and an inter-block RNN.
  • RNN intra-block recurrent neural network
  • AFFB assigns different weights to the contribution of different intermediate features to the training task, and thus can adaptively aggregate different intermediate features.
  • An attention mechanism can be introduced in DPRNN, and thus the training effect of the model can be improved.
  • the multiple first audios can be Fourier transformed to obtain multiple spectrograms corresponding to the multiple first audios, and the fused spectrum features of the multiple spectrum features are extracted through the first model, and the fused spectrum features are spliced with the angle features, and then the spliced features are processed to obtain multiple first weights and multiple second weights associated with the multiple first audios.
  • the terminal device inputs 6 spectrograms and angle features corresponding to 6 first audios to the first model.
  • the first model can extract the fused spectrum features from the 6 spectrograms, splice the fused spectrum features and the angle features to obtain the spliced features, and perform convolution processing on the spliced features to obtain 6 first weights and 6 second weights.
  • FIG5 is a schematic diagram of a process for determining a first weight and a second weight provided by an embodiment of the present disclosure. Please refer to FIG5, which includes: a first audio A, a first audio B, a first audio C, a target direction, and a first model.
  • the first audio A, the first audio B, and the first audio C are subjected to Fourier transform processing to obtain a spectrum A associated with the first audio A, a spectrum B associated with the first audio B, and a spectrum C associated with the first audio C.
  • the angle feature is determined by spectrum A, spectrum B, spectrum C and target direction, and the angle feature, spectrum A, spectrum B and spectrum C are input to the first model.
  • the first model can output the first weight A, the first weight B, the first weight C, the second weight A, the second weight B and the second weight C.
  • the first weight A and the second weight A are weights associated with the first audio A
  • the first weight B and the second weight B are weights associated with the first audio B
  • the first weight C and the second weight C are weights associated with the first audio C.
  • multiple first target audios and non-target audios are determined, specifically: based on multiple first weights and multiple first audios, multiple first target audios are determined. For example, each first audio is multiplied by the corresponding first weight to obtain the first target audio corresponding to the first audio.
  • This method can be used to obtain multiple first target audios associated with multiple first audios. For example, if the first weight corresponding to the first audio A is weight A, and the first weight corresponding to the first audio B is weight B, then the product of the first audio A and weight A is determined as the first target audio associated with the first audio A, and the product of the first audio B and weight B is determined as the first target audio associated with the first audio B.
  • non-target audio is determined based on multiple second weights and multiple first audios. For example, each first audio is multiplied by the corresponding second weight to obtain the non-target audio corresponding to the first audio.
  • This method can be used to obtain non-target audio associated with multiple first audios. For example, if the second weight corresponding to the first audio 1 is weight 1, and the second weight corresponding to the first audio 2 is weight 2, then the product of the first audio 1 and weight 1 is determined as the non-target audio associated with the first audio 1, and the product of the first audio 2 and weight 2 is determined as the non-target audio associated with the first audio 2.
  • S402 Determine a second audio based on a plurality of first audios, a plurality of first target audios, and a plurality of non-target audios.
  • the terminal device may determine the second audio based on the following feasible implementation method: multiple first target audios and non-target audios are processed based on the second model to obtain multiple target weights.
  • the target weight is the proportion of the audio associated with the target direction in each frequency point of the first audio.
  • the target weight may be a weight vector. If the first audio includes 100 frequency points, the weight vector includes 100 sub-vectors. If the first audio includes 1000 frequency points, the weight vector includes 1000 sub-vectors.
  • the second model is obtained by training multiple groups of second samples, wherein the multiple groups of second samples include multiple sample first target audios, multiple sample non-target audios associated with the multiple sample first target audios, and multiple sample target weights.
  • sample first target audio 1 and sample non-target audio 1 and sample weight 1 associated with sample first target audio 1 are obtained to obtain a group of second samples, wherein the group of second samples includes sample first target audio 1, sample non-target audio 1, and sample weight 1.
  • multiple sub-audios are determined, and the multiple sub-audios are fused to obtain a second audio. For example, after the terminal device processes 6 first target audios and 6 non-target audios through the second model, 6 target weights are obtained, each first audio corresponds to 1 target weight, and 6 sub-audios are obtained by multiplying the 6 first audios with the associated 6 target weights, and the 6 sub-audios are merged to obtain the second audio.
  • FIG6 is a schematic diagram of a process for determining a second audio provided by an embodiment of the present disclosure.
  • the first target audio A, the first target audio B, the first target audio C, the non-target audio A, the non-target audio B, and the non-target audio C are input into the second model, and the second model can output the target weight A, the target weight B, and the target weight C.
  • the product of the first audio A and the target weight A is determined as sub-audio A
  • the product of the first audio B and the target weight B is determined as sub-audio B
  • the product of the first audio C and the target weight C is determined as sub-audio C
  • sub-audio A, sub-audio B and sub-audio C are fused to obtain the second audio.
  • the disclosed embodiment provides a method for determining a second audio, which determines a plurality of first weights and a plurality of second weights associated with a plurality of first audios based on a plurality of first audios and angle features, determines a plurality of first target audios and non-target audios based on the plurality of first weights, the plurality of second weights and the plurality of first audios, and determines a second audio based on the plurality of first audios, the plurality of first target audios and the non-target audios.
  • the terminal device can enhance the first target audio effect in the first audio, suppress the non-target audio in the first audio, and thereby reduce the noise in the second audio, and improve the playback effect of the second audio.
  • FIG7 is a schematic diagram of a process of an audio processing method provided by an embodiment of the present disclosure. Please refer to FIG7, which includes: a first audio A, a first audio B, a first audio C, a target direction, a first model, and a second model.
  • the first audio A, the first audio B, and the first audio C are subjected to Fourier transform processing to obtain a spectrum A associated with the first audio A, a spectrum B associated with the first audio B, and a spectrum C associated with the first audio C.
  • the angle feature is determined by spectrum A, spectrum B, spectrum C and target direction, and the angle feature, spectrum A, spectrum B and spectrum C are input to the first model.
  • the first model can output the first weight A, the first weight B, the first weight C, the second weight A, the second weight B and the second weight C.
  • the first weight A and the second weight A are weights associated with the first audio A
  • the first weight B and the second weight B are weights associated with the first audio B
  • the first weight C and the second weight C are weights associated with the first audio C.
  • the product of the first audio A and the first weight A is determined as the first target audio A
  • the product of the first audio A and the second weight A is determined as the non-target audio A
  • the product of the first audio B and the first weight B is determined as the first target audio B
  • the product of the first audio B and the second weight B is determined as the non-target audio B
  • the product of the first audio C and the first weight C is determined as the first target audio C
  • the product of the first audio C and the second weight C is determined as the non-target audio C.
  • the first target audio A, the first target audio B, the first target audio C, the non-target audio A, the non-target audio B, and the non-target audio C are input into the second model, and the second model can output the target weight A, the target weight B, and the target weight C.
  • the product of the first audio A and the target weight A is determined as the sub-audio A
  • the product of the first audio B and the target weight B is determined as the sub-audio B
  • the product of the first audio C and the target weight C is determined as the sub-audio C
  • the sub-audio A, the sub-audio B, and the sub-audio C are fused to obtain the second audio, and the second audio is played.
  • the terminal device can enhance the first target audio effect in the first audio, suppress non-target audio in the first audio, and thereby reduce the noise in the second audio and improve the playback effect of the second audio.
  • FIG8 is a schematic diagram of the structure of an audio processing device provided by an embodiment of the present disclosure.
  • the audio processing device 800 includes an acquisition module 801, a first determination module 802, a second determination module 803 and a playback module 804, wherein:
  • the acquisition module 801 is used to acquire multiple first audios collected by multiple audio collection devices
  • the first determination module 802 is used to determine an angle feature based on the multiple first audios and the target direction, where the angle feature is used to indicate the proportion of the sound source in the target direction in each first audio;
  • the second determination module 803 is used to determine the second audio associated with the target direction based on the multiple first audios and the angle feature;
  • the playing module 804 is used for playing the second audio.
  • the second determining module 803 is specifically configured to:
  • the second audio is determined based on the plurality of first audios, the plurality of first target audios, and the non-target audio.
  • the second determining module 803 is specifically configured to:
  • the plurality of first target audios and the non-target audios are determined based on the plurality of first weights, the plurality of second weights, and the plurality of first audios.
  • the second determining module 803 is specifically configured to:
  • the non-target audio is determined based on the plurality of second weights and the plurality of first audios.
  • the second determining module 803 is specifically configured to:
  • the first model is obtained by training multiple groups of first samples, and the multiple groups of first samples include multiple sample audios, angle features associated with the multiple sample audios, and sample first weights and sample second weights associated with each sample audio.
  • the second determining module 803 is specifically configured to:
  • Target weight is the proportion of the audio associated with the target direction in each frequency point of the first audio
  • the second model is obtained by training multiple groups of second samples, and the multiple groups of second samples include multiple sample first target audios, multiple sample non-target audios associated with the multiple sample first target audios, and multiple sample target weights.
  • the first determining module 802 is specifically configured to:
  • the angular feature is determined based on the second phase difference and the plurality of first phase differences.
  • the first determining module 802 is specifically configured to:
  • the multiple cosine similarities are fused to obtain the angle feature.
  • the audio processing device provided in the embodiment of the present disclosure can be used to implement the technical solution of the above method embodiment, which realizes the original The principles and technical effects are similar and will not be described in detail in this embodiment.
  • the embodiments of the present disclosure further provide a computer-readable storage medium, in which computer-executable instructions are stored.
  • a processor executes the computer-executable instructions
  • the processor executes the methods described in the above-mentioned method embodiments.
  • the embodiments of the present disclosure also provide a computer program, which, when executed by a processor, implements the methods described in the above-mentioned various method embodiments.
  • the embodiments of the present disclosure further provide a computer program product, including a computer program, which implements the methods described in the above-mentioned various method embodiments when executed by a processor.
  • the present disclosure provides an audio processing method, apparatus, and terminal device, wherein the terminal device obtains multiple first audios collected by multiple audio collection devices, determines an angle feature based on the multiple first audios and target directions, wherein the angle feature is used to indicate the proportion of the sound source in the target direction in each first audio, determines a second audio associated with the target direction based on the multiple first audios and the angle feature, and plays the second audio.
  • the terminal device can determine the audio in the target direction that needs to be retained in the first audio, and the audio in other directions that needs to be suppressed in the first audio, thereby reducing the noise in the second audio and improving the playback effect of the second audio.
  • FIG9 is a schematic diagram of the structure of a terminal device provided by an embodiment of the present disclosure. Please refer to FIG9, which shows a schematic diagram of the structure of a terminal device 900 suitable for implementing an embodiment of the present disclosure.
  • the terminal device may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, personal digital assistants (PDAs), tablet computers (PADs), portable multimedia players (PMPs), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc.
  • PDAs personal digital assistants
  • PADs tablet computers
  • PMPs portable multimedia players
  • vehicle-mounted terminals such as vehicle-mounted navigation terminals
  • fixed terminals such as digital TVs, desktop computers, etc.
  • the terminal device shown in FIG9 is only an example and should not bring any limitations to the functions and scope of use of the embodiments of the present disclosure.
  • the terminal device 900 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 901, which may perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 902 or a program loaded from a storage device 908 to a random access memory (RAM) 903.
  • a processing device e.g., a central processing unit, a graphics processing unit, etc.
  • RAM random access memory
  • Various programs and data required for the operation of the terminal device 900 are also stored in the RAM 903.
  • the processing device 901, the ROM 902, and the RAM 903 are connected to each other via a bus 904.
  • An input/output (I/O) interface 905 is also connected to the bus 904.
  • the following devices may be connected to the I/O interface 905: input devices 906 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; output devices 907 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 908 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 909.
  • the communication device 909 may allow the terminal device 900 to communicate with other devices wirelessly or by wire to exchange data.
  • FIG. 9 shows a terminal device 900 with various devices, it should be understood that it is not required to implement or have all the devices shown. More or fewer devices may be implemented or have alternatively.
  • an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart.
  • the computer program can be downloaded and installed from a network through a communication device 909, or installed from a storage device 908, or installed from a ROM 902.
  • the processing device 901 the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.
  • the computer-readable medium mentioned above in the present disclosure may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two.
  • the computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above.
  • Computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory for short), an optical fiber, a portable compact disk read-only memory (CD-ROM for short), an optical storage device, a magnetic storage device, or any suitable combination of the above.
  • a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device.
  • a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code.
  • This propagated data signal can take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above.
  • Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate or transmit programs for use by or in conjunction with instruction execution systems, devices or devices.
  • the program code contained on the computer-readable medium can be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.
  • the computer-readable medium may be included in the terminal device, or may exist independently without being installed in the terminal device.
  • the computer-readable medium carries one or more programs.
  • the terminal device executes the method shown in the above embodiment.
  • Computer program code for performing the operations of the present disclosure may be written in one or more programming languages or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages.
  • the program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server.
  • the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).
  • LAN Local Area Network
  • WAN Wide Area Network
  • each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function.
  • the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved.
  • each square box in the block diagram and/or flow chart, and the combination of the square boxes in the block diagram and/or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.
  • the units involved in the embodiments described in the present disclosure may be implemented by software or hardware.
  • the name of a unit does not limit the unit itself in some cases.
  • the first acquisition unit may also be described as a "unit for acquiring at least two Internet Protocol addresses".
  • exemplary types of hardware logic components include: Field Programmable Gate Array (FPGA), Application Specific Integrated Circuit (ASIC), Application Specific Standard Product (ASSP), System On Chip (SOC), Complex Programmable Logic Device (CPLD), etc.
  • FPGA Field Programmable Gate Array
  • ASIC Application Specific Integrated Circuit
  • ASSP Application Specific Standard Product
  • SOC System On Chip
  • CPLD Complex Programmable Logic Device
  • a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment.
  • a machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium.
  • a machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing.
  • a more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
  • RAM random access memory
  • ROM read-only memory
  • EPROM or flash memory erasable programmable read-only memory
  • CD-ROM portable compact disk read-only memory
  • CD-ROM compact disk read-only memory
  • magnetic storage device or any suitable combination of the foregoing.
  • a prompt message is sent to the user to clearly prompt the user that the operation requested to be performed will require obtaining and using the user's personal information.
  • the user can autonomously choose whether to provide personal information to software or hardware such as a terminal device, application, server, or storage medium that performs the operation of the technical solution of the present disclosure according to the prompt message.
  • the prompt information in response to receiving an active request from the user, may be sent to the user in the form of a pop-up window, in which the prompt information may be presented in text form.
  • the pop-up window may also carry a selection control for the user to choose "agree” or “disagree” to provide personal information to the terminal device.
  • the data involved in this technical solution shall comply with the requirements of the relevant laws and regulations.
  • the data may include information, parameters and messages, such as flow switching indication information.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Computational Linguistics (AREA)
  • Signal Processing (AREA)
  • Health & Medical Sciences (AREA)
  • Human Computer Interaction (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Quality & Reliability (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • Circuit For Audible Band Transducer (AREA)

Abstract

本公开提供一种音频处理方法、装置及终端设备，该方法包括：获取多个音频采集设备采集的多个第一音频；基于所述多个第一音频和目标方向，确定角度特征，所述角度特征用于指示所述目标方向的音源在每个第一音频中所占的比重；基于所述多个第一音频和所述角度特征，确定所述目标方向相关联的第二音频，并播放所述第二音频。

Description

音频处理方法、装置及终端设备
相关申请交叉引用
本申请要求于2022年10月18日提交中国专利局、申请号为202211275143.4、发明名称为“音频处理方法、装置及终端设备”的中国专利申请的优先权，其全部内容通过引用并入本文。
技术领域
本公开实施例涉及音频处理技术领域，尤其涉及一种音频处理方法、装置及终端设备。
背景技术
终端设备可以采集空间中的音频，并播放该音频。例如，在会议场景中，终端设备中的麦克风可以采集用户的语音，并向其它终端设备发送该语音。
目前，终端设备可以通过麦克风采集空间中的音频，并向其它终端设备发送采集的音频，以使其它终端设备播放该音频。例如，会议室中的麦克风可以采集用户的语音，终端设备可以从麦克风中获取该语音，并向其它终端设备发送该语音，以使其它终端设备播放该语音，进而实现远程会议通话的效果。然而，在空间中既存在目标音频，也存在非目标音频(例如，与非目标对象关联的音频或者干扰音频)。空间中非目标音频的存在，导致音频播放的效果较差或者所播放的音频并非用户所关注的音频。
发明内容
本公开提供一种音频处理方法、装置及终端设备，用于解决现有技术中音频播放的效果较差的技术问题。
第一方面，本公开提供一种音频处理方法，该方法包括：
获取多个音频采集设备采集的多个第一音频；
基于所述多个第一音频和目标方向，确定角度特征，所述角度特征用于指示所述目标方向的音源在每个第一音频中所占的比重；
基于所述多个第一音频和所述角度特征，确定所述目标方向相关联的第二音频，并播放所述第二音频。
第二方面，本公开提供一种音频处理装置，该音频处理装置包括获取模块、第一确定模块、第二确定模块和播放模块，其中：
所述获取模块用于，获取多个音频采集设备采集的多个第一音频；
所述第一确定模块用于，基于所述多个第一音频和目标方向，确定角度特征，所述角度特征用于指示所述目标方向的音源在每个第一音频中所占的比重；
所述第二确定模块用于，基于所述多个第一音频和所述角度特征，确定所述目标方向相关联的第二音频；
所述播放模块用于，播放所述第二音频。
第三方面，本公开实施例提供一种终端设备，包括：处理器和存储器；
所述存储器存储计算机执行指令；
所述处理器执行所述存储器存储的计算机执行指令，使得所述至少一个处理器执行如上第一方面以及第一方面各种可能涉及的所述音频处理方法。
第四方面，本公开实施例提供一种计算机可读存储介质，所述计算机可读存储介质中存储有计算机执行指令，当处理器执行所述计算机执行指令时，实现如上第一方面以及第一方面各种可能涉及的所述音频处理方法。
第五方面，本公开实施例提供一种计算机程序产品，包括计算机程序，所述计算机程序被处理器执行时实现如上第一方面以及第一方面各种可能涉及的所述音频处理方法。
第六方面，本公开实施例提供一种计算机程序，该计算机程序被处理器执行时实现如上述第一方面所述的音频处理方法。
附图说明
为了更清楚地说明本公开实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作一简单地介绍，显而易见地，下面描述中的附图是本公开的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。
图1为本公开实施例提供的一种应用场景示意图；
图2为本公开实施例提供的一种音频处理方法的流程示意图；
图3为本公开实施例提供的一种确定多个第一相位差的过程示意图；
图4为本公开实施例提供的一种确定第二音频的方法示意图；
图5为本公开实施例提供的一种确定第一权重和第二权重的过程示意图；
图6为本公开实施例提供的一种确定第二音频的过程示意图；
图7为本公开实施例提供的一种音频处理方法的过程示意图；
图8为本公开实施例提供的一种音频处理装置的结构示意图；以及，
图9为本公开实施例提供的一种终端设备的结构示意图。
具体实施方式
这里将详细地对示例性实施例进行说明，其示例表示在附图中。下面的描述涉及附图时，除非另有表示，不同附图中的相同数字表示相同或相似的要素。以下示例性实施例中所描述的实施方式并不代表与本公开相一致的所有实施方式。相反，它们仅是与如所附权利要求书中所详述的、本公开的一些方面相一致的装置和方法的例子。
为了便于理解，下面，对本公开实施例涉及的概念进行说明。
终端设备：是一种具有无线收发功能的设备。终端设备可以部署在陆地上，包括室内或室外、手持、穿戴或车载；也可以部署在水面上(如轮船等)。所述终端设备可以是手机(mobile phone)、平板电脑(Portable Android Device，PAD)、带无线收发功能的电脑、虚拟现实(virtual reality，VR)终端设备、增强现实(augmented reality，AR)终端设备、工业控制(industrial control)中的无线终端、车载终端设备、无人驾驶(self driving)中的无线终端、远程医疗(remote medical)中的无线终端设备、智能电网(smart grid)中的无线终端设备、运输安全(transportation safety) 中的无线终端设备、智慧城市(smart city)中的无线终端设备、智慧家庭(smart home)中的无线终端设备、可穿戴终端设备等。本公开实施例所涉及的终端设备还可以称为终端、用户设备(user equipment，UE)、接入终端设备、车载终端、工业控制终端、UE单元、UE站、移动站、移动台、远方站、远程终端设备、移动设备、UE终端设备、无线通信设备、UE代理或UE装置等。终端设备也可以是固定的或者移动的。
在相关技术中，终端设备可以通过麦克风采集空间中的音频，并向其它终端设备发送采集的音频，以使其它终端设备播放该音频。例如，在远程会议的场景中，会议室中的麦克风可以采集参会的用户的语音，会议室中的终端设备可以从麦克风中获取参会的用户的语音，并向远程连接的终端设备发送该语音，以使远程连接的终端设备播放参会的用户的语音。但是，在空间中的非目标音频较多时，例如，在会议室内的环境噪音较多时，麦克风采集的音频的噪声较大，进而导致音频播放的效果较差。
为了解决相关技术中的技术问题，本公开实施例提供一种音频处理方法，终端设备获取多个音频采集设备采集的多个第一音频，基于多个第一音频和目标方向确定角度特征，其中，角度特征用于指示目标方向的音源在每个第一音频中所占的比重，基于多个第一音频和角度特征，确定每个第一音频在目标方向相关联的第一目标音频，以及确定每个第一音频在目标方向相关联的非目标音频，基于多个第一音频、多个第一目标音频和非目标音频，确定第二音频，并播放第二音频，在上述方法中，由于角度特征可以指示目标方向的音源在每个第一音频中所占的比重，因此，终端设备可以增强第一音频中的第一目标音频效果，抑制第一音频中的非目标音频，进而降低第二音频中的噪音，提高第二音频的播放效果。
下面，结合图1，对本公开实施例的应用场景进行说明。
图1为本公开实施例提供的一种应用场景示意图。请参见图1，包括：用户、音频采集设备和终端设备。其中，用户向音频采集设备发送语音A，音频采集设备还可以采集到环境中的噪音A和噪音B。音频采集设备向终端设备发送语音A、噪音A和噪音B的混合音频，终端设备可以确定该混合音频的角度特征，并根据混合音频和角度特征，确定用户的语音A，并播放语音A。这样，由于角度特征可以指示语音A在混合语音中所占的比重，因此，终端设备可以放大语音A的信号，抑制噪音A和噪音B的信号，进而提高音频的播放效果。
需要说明的是，图1只是示例性的示意本公开实施例的应用场景，并非对本公开实施例的应用场景的限定。
下面以具体地实施例对本公开的技术方案以及本公开的技术方案如何解决上述技术问题进行详细说明。下面这几个具体的实施例可以相互结合，对于相同或相似的概念或过程可能在某些实施例中不再赘述。下面将结合附图，对本公开的实施例进行描述。
图2为本公开实施例提供的一种音频处理方法的流程示意图。请参见图2，该方法可以包括：
S201、获取多个音频采集设备采集的多个第一音频。
本公开实施例的执行主体为终端设备，也可以为设置在终端设备中的音频处理装置。其中，音频处理装置可以通过软件实现，音频处理装置也可以通过软件和硬件的结合实现，本公开实施例对此不作限定。
可选的，音频采集设备用于采集空间中的音频。例如，音频采集设备可以为麦克风。例如，音频采集设备可以为会议室中的麦克风，音频采集设备也可以为终端设备中的麦克风， 本公开实施例对此不作限定。
可选的，第一音频为音频采集设备采集的音频。例如，在会议场景中，会议室中可以设置多个麦克风，第一音频可以为多个麦克风采集的音频。例如，在会议场景中，在参会的用户发出语音时，会议室中的麦克风可以采集用户的语音，并将用户的语音确定为第一音频。
可选的，终端设备可以通过音频采集设备获取多个第一音频。例如，终端设备可以和多个音频采集设备连接，多个音频采集设备获取到多个第一音频时，可以向终端设备发送多个第一音频。
可选的，终端设备可以通过其它终端设备获取多个第一音频。例如，终端设备可以和其它终端设备通信连接，在其它终端设备中的麦克风获取多个第一音频时，其它终端设备可以向终端设备发送多个第一音频。
需要说明的是，一个音频采集设备可以采集一个第一音频。例如，在空间中包括一个音源时，若空间中包括1个音频采集设备，则1个音频采集设备针对于该音源可以采集1个第一音频，若空间中包括6个音频采集设备，则6个音频采集设备针对于该音源可以采集6个第一音频。
S202、基于多个第一音频和目标方向，确定角度特征。
可选的，角度特征用于指示目标方向的音源在每个第一音频中所占的比重。例如，若终端设备获取的第一音频包括第一音频A和第一音频B，则角度特征用于指示目标方向相关联的音频在第一音频A中的占比和目标方向相关联的音频在第一音频B中的占比。
可选的，终端设备可以基于如下可行的实现方式，确定角度特征：确定多个第一音频之间的相位差，得到多个第一相位差，确定目标方向相关联的第二相位差，基于第二相位差和多个第一相位差，确定角度特征。
可选的，确定多个第一音频之间的相位差，具体为：终端设备在多个第一音频中确定初始音频。可选的，终端设备可以将多个第一音频中的任意一个第一音频确定为初始音频。例如，若终端设备获取的多个第一音频包括第一音频A、第一音频B和第一音频C，则终端设备可以将第一音频A确定为初始音频，终端设备可以将第一音频B确定为初始音频，终端设备也可以将第一音频C确定为初始音频。
可选的，获取初始音频与每个第一音频之间的相位差，得到多个第一相位差。例如，若终端设备采集的多个第一音频包括第一音频A、第一音频B和第一音频C，终端设备确定初始音频为第一音频A，则终端设备可以获取第一音频A和第一音频B之间的相位差A，终端设备可以获取第一音频A和第一音频C之间的相位差B，进而将相位差A和相位差B确定为第一相位差。
需要说明的是，本公开实施例中终端设备通过多个第一音频和目标方向，确定角度特征时，终端设备可以先对多个第一音频进行傅里叶变换，进而得到多个第一音频相关联的多个频谱图，进而通过多个频谱图，确定多个第一相位差，这样可以提高第一相位差确定的准确度，进而得到准确的角度特征，提高音频处理的效果。
下面，结合图3，对确定多个第一相位差的过程进行说明。
图3为本公开实施例提供的一种确定多个第一相位差的过程示意图。请参见图3，包括第一音频A、第一音频B和第一音频C。对第一音频A、第一音频B和第一音频C进行傅里叶变换处理，得到第一音频A相关联的频谱A、第一音频B相关联的频谱B和第一音频C相关 联的频谱C。将第一音频A确定为初始音频，将频谱A和频谱B之间的相位差，确定为第一相位差A，将频谱A和频谱C之间的相位差，确定为第一相位差B。
可选的，终端设备可以将目标方向的音源与音频采集设备之间的相位差，确定为目标方向相关联的第二相位差。例如，目标方向相关联的第二相位差是由目标方向上的音源到达麦克风之间的路程差引起的相位差，该相位差为目标相位差。
可选的，基于第二相位差和多个第一相位差，确定角度特征，具体为：确定第二相位差与每个第一相位差之间的余弦相似度，得到多个余弦相似度。例如，若目标方向相关联的第二相位差为相位差A，第一相位差包括相位差B和相位差C，则多个余弦相似度包括相位差A与相位差B之间的余弦相似度和相位差A与相位差C之间的余弦相似度。
对多个余弦相似度进行融合处理，得到角度特征。例如，终端设备可以将多个余弦相似度进行拼接处理，进而得到角度特征，终端设备也可以将多个余弦相似度进行叠加，得到角度特征，本公开实施例对此不作限定。例如，终端设备可以将余弦相似度A和余弦相似度B进行拼接处理，进而得到角度特征。
可选的，终端设备可以响应于用户的触控操作，确定目标方向。例如，终端设备可以基于用户对视频会议显示屏幕中的任何位置的触控操作，将用户触控的位置对应的空间中的方向确定为目标方向。
可选的，终端设备可以响应于用户的语音操作，确定目标方向。例如，在用户A发出“请用户B发言”的语音时，终端设备可以将用户B所在的位置，确定为目标方向。需要说明的是，终端设备也可以根据其它方式确定目标方向，本公开实施例对此不作限定。
S203、基于多个第一音频和角度特征，确定目标方向相关联的第二音频。
可选的，第二音频为终端设备在目标方向获取的音频。例如，空间中的音源位于麦克风的正北方向，若目标方向为正东方向，则第二音频为该音源在正东方向的音频。
可选的，终端设备可以基于如下可行的实现方式，确定目标方向相关联的第二音频：基于多个第一音频和角度特征，确定多个第一目标音频和非目标音频，基于多个第一音频、多个第一目标音频和非目标音频，确定第二音频。可选的，第一目标音频为第一音频在目标方向相关联的音频，非目标音频为第一音频在其它方向相关联的音频。例如，由于角度特征可以指示目标方向的音源在每个第一音频中的占比，因此，通过角度特征和多个第一音频，可以得到每个第一音频在目标方向上的第一目标音频和每个第一音频在目标方向上的非目标音频(其它方向的音频)。
S204、播放第二音频。
可选的，终端设备对多个子音频进行融合处理，得到第二音频之后，可以播放第二音频。
可选的，终端设备得到第二音频之后，也可以向其它终端设备或服务器发送该第二音频，本公开实施例对此不作限定。
本公开实施例提供一种音频处理方法，终端设备获取多个音频采集设备采集的多个第一音频，基于多个第一音频和目标方向确定角度特征，基于多个第一音频和角度特征，确定每个第一音频在目标方向相关联的第一目标音频，以及确定每个第一音频在目标方向相关联的非目标音频，基于第二模型对多个第一目标音频和非目标音频进行处理，得到多个目标权重，基于多个目标权重和多个第一音频，确定多个子音频，对多个子音频进行融合处理，得到第二音频，并播放第二音频。在上述方法中，由于角度特征可以指示目标方向的音源在每个第 一音频中所占的比重，因此，终端设备可以增强第一音频中的第一目标音频效果，抑制第一音频中的非目标音频，进而降低第二音频中的噪音，提高第二音频的播放效果。
在图2所示的实施例的基础上，下面，结合图4，对上述音频处理方法中，基于多个第一音频和角度特征，确定目标方向相关联的第二音频的过程进行详细的说明。
图4为本公开实施例提供的一种确定第二音频的方法示意图。请参见图4，该方法流程包括：
S401、基于多个第一音频和角度特征，确定多个第一目标音频和非目标音频。
可选的，第一目标音频为第一音频在目标方向相关联的音频，非目标音频为第一音频在其它方向相关联的音频。例如，每个第一音频都包括一个第一目标音频和一个非目标音频。例如，终端设备通过音频采集设备获取6个第一音频，则终端设备可以得到6个第一目标音频和6个非目标音频。
可选的，终端设备可以基于如下可行的实现方式，确定多个第一目标音频和非目标音频：基于多个第一音频和角度特征，确定多个第一音频相关联的多个第一权重和多个第二权重，基于多个第一权重、多个第二权重和多个第一音频，确定多个第一目标音频和非目标音频。可选的，第一权重可以为目标方向的音频在第一音频中所占的权重，第二权重可以为其它方向的音频在第一音频中所占的权重。
可选的，基于多个第一音频和角度特征，确定多个第一音频相关联的多个第一权重和多个第二权重，具体为：基于第一模型、多个第一音频和角度特征，确定第一权重和第二权重。可选的，第一模型为对多组第一样本训练得到的，多组第一样本包括多个样本音频、多个样本音频相关联的角度特征和每个样本音频相关联的样本第一权重和样本第二权重。例如，获取样本音频1、样本音频1相关联的角度特征1和每个样本音频相关联的样本第一权重1、样本第二权重1，其中，该组第一样本中包括样本音频1、样本角度特征1、样本第一权重1和样本第二权重1，采用该种方法可以得到多组第一样本。
可选的，第一模型的结构可以包括编码器、双路递归神经网络(Dual Path Recurrent Neural Network，简称DPRNN)和解码器。例如，DPRNN可以有效地建模音频的频率之间的相关性和帧之间的相关性，进而可以有效地对音频进行分解，编码器用于编码角度特征和多个频谱图的特征，DPRNN可以由堆叠的SA-DPRNN块和自适应的特征融合块(Adaptive Feature Fusion Block，简称AFFB)组成，SA-DPRNN块包括块内递归神经网络(Recurrent Neural Network，简称RNN)和块间RNN，AFFB根据不同的中间特征，对训练任务的贡献分配不同的权重，进而可以自适应地聚合不同的中间特征，在DPRNN中可以引入注意力机制，进而可以提高模型的训练效果。
可选的，在通过第一模型对多个第一音频和角度特征进行处理时，可以对多个第一音频进行傅里叶变换处理，得到多个第一音频对应的多个频谱图，通过第一模型提取多个频谱图的融合频谱特征，并将该融合频谱特征与角度特征进行拼接，进而对拼接之后的特征进行处理，得到多个第一音频相关联的多个第一权重和多个第二权重。例如，终端设备向第一模型输入6个第一音频对应的6个频谱图和角度特征，第一模型可以在6个频谱图中提取融合频谱特征，并将融合频谱特征和角度特征进行拼接，得到拼接特征，并对拼接特征进行卷积处理，得到6个第一权重和6个第二权重。
下面，结合图5，对确定多个第一权重和多个第二权重的过程进行说明。
图5为本公开实施例提供的一种确定第一权重和第二权重的过程示意图。请参见图5，包括：第一音频A、第一音频B、第一音频C、目标方向和第一模型。对第一音频A、第一音频B和第一音频C进行傅里叶变换处理，得到第一音频A相关联的频谱A、第一音频B相关联的频谱B和第一音频C相关联的频谱C。
请参见图5，通过频谱A、频谱B、频谱C和目标方向，确定角度特征，并向第一模型输入角度特征、频谱A、频谱B和频谱C，第一模型可以输出第一权重A、第一权重B、第一权重C、第二权重A、第二权重B和第二权重C。其中，第一权重A和第二权重A为第一音频A相关联的权重，第一权重B和第二权重B为第一音频B相关联的权重，第一权重C和第二权重C为第一音频C相关联的权重。
可选的，基于多个第一权重、多个第二权重和多个第一音频，确定多个第一目标音频和非目标音频，具体为：基于多个第一权重和多个第一音频，确定多个第一目标音频。例如，将每个第一音频与对应的第一权重相乘，得到该第一音频对应的第一目标音频，采用该种方法可以得到多个第一音频相关联的多个第一目标音频。例如，若第一音频A对应的第一权重为权重A，第一音频B对应的第一权重为权重B，则将第一音频A和权重A的乘积确定为第一音频A相关联的第一目标音频，将第一音频B和权重B的乘积确定为第一音频B相关联的第一目标音频。
可选的，基于多个第二权重和多个第一音频，确定非目标音频。例如，将每个第一音频与对应的第二权重相乘，得到该第一音频对应的非目标音频，采用该种方法可以得到多个第一音频相关联的非目标音频。例如，若第一音频1对应的第二权重为权重1，第一音频2对应的第二权重为权重2，则将第一音频1和权重1的乘积确定为第一音频1相关联的非目标音频，将第一音频2和权重2的乘积确定为第一音频2相关联的非目标音频。
S402、基于多个第一音频、多个第一目标音频和非目标音频，确定第二音频。
可选的，终端设备可以基于如下可行的实现方式，确定第二音频：基于第二模型对多个第一目标音频和非目标音频进行处理，得到多个目标权重。可选的，目标权重为目标方向相关联的音频在第一音频的每个频点中所占的比重。例如，目标权重可以为权重向量，若第一音频中包括100个频点，则权重向量中包括100个子向量，若第一音频中包括1000个频点，则权重向量中包括1000个子向量。
可选的，第二模型为对多组第二样本训练得到的，其中，多组第二样本包括多个样本第一目标音频、多个样本第一目标音频相关联的多个样本非目标音频和多个样本目标权重。例如，获取样本第一目标音频1和样本第一目标音频1相关联的样本非目标音频1和样本权重1，得到一组第二样本，其中，该组第二样本中包括样本第一目标音频1、样本非目标音频1和样本权重1。
基于多个目标权重和多个第一音频，确定多个子音频，并对多个子音频进行融合处理，得到第二音频。例如，终端设备通过第二模型对6个第一目标音频和6个非目标音频进行处理之后，得到6个目标权重，每个第一音频都对应1个目标权重，通过6个第一音频与相关联的6目标权重进行相乘，得到6个子音频，对6个子音频进行合并，得到第二音频。
下面，结合图6，对确定第二音频的过程进行说明。
图6为本公开实施例提供的一种确定第二音频的过程示意图。请参见图6，包括第一目标音频A、第一目标音频B、第一目标音频C、非目标音频A、非目标音频B、非目标音频C和 第二模型。向第二模型中输入第一目标音频A、第一目标音频B、第一目标音频C、非目标音频A、非目标音频B和非目标音频C，第二模型可以输出目标权重A、目标权重B和目标权重C。
请参见图6，将第一音频A和目标权重A的乘积确定为子音频A，将第一音频B和目标权重B的乘积确定为子音频B，将第一音频C和目标权重C的乘积确定为子音频C，并对子音频A、子音频B和子音频C进行融合处理，得到第二音频。
本公开实施例提供一种确定第二音频的方法，基于多个第一音频和角度特征，确定多个第一音频相关联的多个第一权重和多个第二权重，基于多个第一权重、多个第二权重和多个第一音频，确定多个第一目标音频和非目标音频，基于多个第一音频、多个第一目标音频和非目标音频，确定第二音频。这样，由于第一权重可以指示目标方向的音频在第一音频中所占的比重，第二权重可以指示其它方向的音频在第一音频中所占的比重，因此，终端设备可以增强第一音频中的第一目标音频效果，抑制第一音频中的非目标音频，进而降低第二音频中的噪音，提高第二音频的播放效果。
在上述任意一个实施例的基础上，下面，结合图7，对上述音频处理方法的过程进行说明。
图7为本公开实施例提供的一种音频处理方法的过程示意图。请参见图7，包括：第一音频A、第一音频B、第一音频C、目标方向、第一模型和第二模型。对第一音频A、第一音频B和第一音频C进行傅里叶变换处理，得到第一音频A相关联的频谱A、第一音频B相关联的频谱B和第一音频C相关联的频谱C。
请参见图7，通过频谱A、频谱B、频谱C和目标方向，确定角度特征，并向第一模型输入角度特征、频谱A、频谱B和频谱C，第一模型可以输出第一权重A、第一权重B、第一权重C、第二权重A、第二权重B和第二权重C。其中，第一权重A和第二权重A为第一音频A相关联的权重，第一权重B和第二权重B为第一音频B相关联的权重，第一权重C和第二权重C为第一音频C相关联的权重。
请参见图7，将第一音频A和第一权重A的乘积确定为第一目标音频A，将第一音频A和第二权重A的乘积确定为非目标音频A，将第一音频B和第一权重B的乘积确定为第一目标音频B，将第一音频B和第二权重B的乘积确定为非目标音频B，将第一音频C和第一权重C的乘积确定为第一目标音频C，将第一音频C和第二权重C的乘积确定为非目标音频C。
请参见图7，向第二模型中输入第一目标音频A、第一目标音频B、第一目标音频C、非目标音频A、非目标音频B和非目标音频C，第二模型可以输出目标权重A、目标权重B和目标权重C。将第一音频A和目标权重A的乘积确定为子音频A，将第一音频B和目标权重B的乘积确定为子音频B，将第一音频C和目标权重C的乘积确定为子音频C，并对子音频A、子音频B和子音频C进行融合处理，得到第二音频，并播放第二音频。
这样，由于第一权重可以指示目标方向的音频在第一音频中所占的比重，第二权重可以指示其它方向的音频在第一音频中所占的比重，因此，终端设备可以增强第一音频中的第一目标音频效果，抑制第一音频中的非目标音频，进而降低第二音频中的噪音，提高第二音频的播放效果。
图8为本公开实施例提供的一种音频处理装置的结构示意图。请参见图8，该音频处理装置800包括获取模块801、第一确定模块802、第二确定模块803和播放模块804，其中：
所述获取模块801用于，获取多个音频采集设备采集的多个第一音频；
所述第一确定模块802用于，基于所述多个第一音频和目标方向，确定角度特征，所述角度特征用于指示所述目标方向的音源在每个第一音频中所占的比重；
所述第二确定模块803用于，基于所述多个第一音频和所述角度特征，确定所述目标方向相关联的第二音频；
所述播放模块804用于，播放所述第二音频。
根据本公开一个或多个实施例，所述第二确定模块803具体用于：
基于所述多个第一音频和所述角度特征，确定多个第一目标音频和非目标音频，所述第一目标音频为所述第一音频在目标方向相关联的音频，所述非目标音频为所述第一音频在其它方向相关联的音频；
基于所述多个第一音频、所述多个第一目标音频和所述非目标音频，确定所述第二音频。
根据本公开一个或多个实施例，所述第二确定模块803具体用于：
基于所述多个第一音频和所述角度特征，确定所述多个第一音频相关联的多个第一权重和多个第二权重；
基于所述多个第一权重、所述多个第二权重和所述多个第一音频，确定所述多个第一目标音频和所述非目标音频。
根据本公开一个或多个实施例，所述第二确定模块803具体用于：
基于所述多个第一权重和所述多个第一音频，确定所述多个第一目标音频；
基于所述多个第二权重和所述多个第一音频，确定所述非目标音频。
根据本公开一个或多个实施例，所述第二确定模块803具体用于：
基于第一模型、所述多个第一音频和所述角度特征，确定所述多个第一权重和所述多个第二权重；
其中，所述第一模型为对多组第一样本训练得到的，所述多组第一样本包括多个样本音频、所述多个样本音频相关联的角度特征和所述每个样本音频相关联的样本第一权重和样本第二权重。
根据本公开一个或多个实施例，所述第二确定模块803具体用于：
基于第二模型对所述多个第一目标音频和所述非目标音频进行处理，得到多个目标权重，所述目标权重为所述目标方向相关联的音频在所述第一音频的每个频点中所占的比重；
基于所述多个目标权重和所述多个第一音频，确定多个子音频，并对所述多个子音频进行融合处理，得到所述第二音频；
其中，所述第二模型为对多组第二样本训练得到的，所述多组第二样本包括多个样本第一目标音频、多个样本第一目标音频相关联的多个样本非目标音频和多个样本目标权重。
根据本公开一个或多个实施例，所述第一确定模块802具体用于：
确定所述多个第一音频之间的相位差，得到多个第一相位差；
确定所述目标方向相关联的第二相位差；
基于所述第二相位差和所述多个第一相位差，确定所述角度特征。
根据本公开一个或多个实施例，所述第一确定模块802具体用于：
确定所述第二相位差与每个第一相位差之间的余弦相似度，得到多个余弦相似度；
对所述多个余弦相似度进行融合处理，得到所述角度特征。
本公开实施例提供的音频处理装置，可用于执行上述方法实施例的技术方案，其实现原 理和技术效果类似，本实施例此处不再赘述。
本公开实施例还提供一种计算机可读存储介质，该计算机可读存储介质中存储有计算机执行指令，当处理器执行该计算机执行指令时，使得该处理器执行如上述各个方法实施例所述方法。
本公开实施例还提供一种计算机程序，该计算机程序被处理器执行时实现如上述各个方法实施例所述方法。
本公开实施例还提供一种计算机程序产品，包括计算机程序，该计算机程序被处理器执行时实现如上述各个方法实施例所述方法。
本公开提供一种音频处理方法、装置及终端设备，终端设备获取多个音频采集设备采集的多个第一音频，基于多个第一音频和目标方向，确定角度特征，其中，角度特征用于指示目标方向的音源在每个第一音频中所占的比重，基于多个第一音频和角度特征，确定目标方向相关联的第二音频，并播放第二音频。在上述方法中，由于角度特征可以指示目标方向的音源在每个第一音频中所占的比重，因此，终端设备可以确定第一音频中的需要保留的目标方向的音频，以及第一音频中需要抑制的其它方向的音频，进而可以降低第二音频中的噪音，提高第二音频的播放效果。
图9为本公开实施例提供的一种终端设备的结构示意图。请参见图9，其示出了适于用来实现本公开实施例的终端设备900的结构示意图。其中，终端设备可以包括但不限于诸如移动电话、笔记本电脑、数字广播接收器、个人数字助理(Personal Digital Assistant，简称PDA)、平板电脑(PAD)、便携式多媒体播放器(Portable Media Player，简称PMP)、车载终端(例如车载导航终端)等等的移动终端以及诸如数字TV、台式计算机等等的固定终端。图9示出的终端设备仅仅是一个示例，不应对本公开实施例的功能和使用范围带来任何限制。
如图9所示，终端设备900可以包括处理装置(例如中央处理器、图形处理器等)901，其可以根据存储在只读存储器(Read Only Memory，简称ROM)902中的程序或者从存储装置908加载到随机访问存储器(Random Access Memory，简称RAM)903中的程序而执行各种适当的动作和处理。在RAM 903中，还存储有终端设备900操作所需的各种程序和数据。处理装置901、ROM 902以及RAM 903通过总线904彼此相连。输入/输出(Input/Output，简称I/O)接口905也连接至总线904。
通常，以下装置可以连接至I/O接口905：包括例如触摸屏、触摸板、键盘、鼠标、摄像头、麦克风、加速度计、陀螺仪等的输入装置906；包括例如液晶显示器(Liquid Crystal Display，简称LCD)、扬声器、振动器等的输出装置907；包括例如磁带、硬盘等的存储装置908；以及通信装置909。通信装置909可以允许终端设备900与其他设备进行无线或有线通信以交换数据。虽然图9示出了具有各种装置的终端设备900，但是应理解的是，并不要求实施或具备所有示出的装置。可以替代地实施或具备更多或更少的装置。
特别地，根据本公开的实施例，上文参考流程图描述的过程可以被实现为计算机软件程序。例如，本公开的实施例包括一种计算机程序产品，其包括承载在计算机可读介质上的计算机程序，该计算机程序包含用于执行流程图所示的方法的程序代码。在这样的实施例中，该计算机程序可以通过通信装置909从网络上被下载和安装，或者从存储装置908被安装，或者从ROM 902被安装。在该计算机程序被处理装置901执行时，执行本公开实施例的方法中限定的上述功能。
需要说明的是，本公开上述的计算机可读介质可以是计算机可读信号介质或者计算机可读存储介质或者是上述两者的任意组合。计算机可读存储介质例如可以是——但不限于——电、磁、光、电磁、红外线、或半导体的系统、装置或器件，或者任意以上的组合。计算机可读存储介质的更具体的例子可以包括但不限于：具有一个或多个导线的电连接、便携式计算机磁盘、硬盘、随机访问存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(Erasable Programmable Read-Only Memory，简称EPROM或闪存)、光纤、便携式紧凑磁盘只读存储器(Compact Disk Read-Only Memory，简称CD-ROM)、光存储器件、磁存储器件、或者上述的任意合适的组合。在本公开中，计算机可读存储介质可以是任何包含或存储程序的有形介质，该程序可以被指令执行系统、装置或者器件使用或者与其结合使用。而在本公开中，计算机可读信号介质可以包括在基带中或者作为载波一部分传播的数据信号，其中承载了计算机可读的程序代码。这种传播的数据信号可以采用多种形式，包括但不限于电磁信号、光信号或上述的任意合适的组合。计算机可读信号介质还可以是计算机可读存储介质以外的任何计算机可读介质，该计算机可读信号介质可以发送、传播或者传输用于由指令执行系统、装置或者器件使用或者与其结合使用的程序。计算机可读介质上包含的程序代码可以用任何适当的介质传输，包括但不限于：电线、光缆、RF(Radio Frequency，射频)等等，或者上述的任意合适的组合。
上述计算机可读介质可以是上述终端设备中所包含的；也可以是单独存在，而未装配入该终端设备中。
上述计算机可读介质承载有一个或者多个程序，当上述一个或者多个程序被该终端设备执行时，使得该终端设备执行上述实施例所示的方法。
可以以一种或多种程序设计语言或其组合来编写用于执行本公开的操作的计算机程序代码，上述程序设计语言包括面向对象的程序设计语言—诸如Java、Smalltalk、C++，还包括常规的过程式程序设计语言—诸如“C”语言或类似的程序设计语言。程序代码可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中，远程计算机可以通过任意种类的网络——包括局域网(Local Area Network，简称LAN)或广域网(Wide Area Network，简称WAN)—连接到用户计算机，或者，可以连接到外部计算机(例如利用因特网服务提供商来通过因特网连接)。
附图中的流程图和框图，图示了按照本公开各种实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段、或代码的一部分，该模块、程序段、或代码的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。也应当注意，在有些作为替换的实现中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个接连地表示的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或操作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。
描述于本公开实施例中所涉及到的单元可以通过软件的方式实现，也可以通过硬件的方式来实现。其中，单元的名称在某种情况下并不构成对该单元本身的限定，例如，第一获取单元还可以被描述为“获取至少两个网际协议地址的单元”。
本文中以上描述的功能可以至少部分地由一个或多个硬件逻辑部件来执行。例如，非限制性地，可以使用的示范类型的硬件逻辑部件包括：现场可编程门阵列(Field Programmable Gate Array，简称FPGA)、专用集成电路(Application Specific Integrated Circuit，简称ASIC)、专用标准产品(Application Specific Standard Product，简称ASSP)、片上系统(System On Chip，简称SOC)、复杂可编程逻辑设备(Complex Programmable Logic Device，简称CPLD)等等。
在本公开的上下文中，机器可读介质可以是有形的介质，其可以包含或存储以供指令执行系统、装置或设备使用或与指令执行系统、装置或设备结合地使用的程序。机器可读介质可以是机器可读信号介质或机器可读储存介质。机器可读介质可以包括但不限于电子的、磁性的、光学的、电磁的、红外的、或半导体系统、装置或设备，或者上述内容的任何合适组合。机器可读存储介质的更具体示例会包括基于一个或多个线的电气连接、便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦除可编程只读存储器(EPROM或快闪存储器)、光纤、便捷式紧凑盘只读存储器(CD-ROM)、光学储存设备、磁储存设备、或上述内容的任何合适组合。
需要注意，本公开中提及的“一个”、“多个”的修饰是示意性而非限制性的，本领域技术人员应当理解，除非在上下文另有明确指出，否则应该理解为“一个或多个”。
本公开实施方式中的多个装置之间所交互的消息或者信息的名称仅用于说明性的目的，而并不是用于对这些消息或信息的范围进行限制。
可以理解的是，在使用本公开各实施例公开的技术方案之前，均应当依据相关法律法规通过恰当的方式对本公开所涉及个人信息的类型、使用范围、使用场景等告知用户并获得用户的授权。
例如，在响应于接收到用户的主动请求时，向用户发送提示信息，以明确地提示用户，其请求执行的操作将需要获取和使用到用户的个人信息。从而，使得用户可以根据提示信息来自主地选择是否向执行本公开技术方案的操作的终端设备、应用程序、服务器或存储介质等软件或硬件提供个人信息。
作为一种可选的但非限定性的实现方式，响应于接收到用户的主动请求，向用户发送提示信息的方式例如可以是弹窗的方式，弹窗中可以以文字的方式呈现提示信息。此外，弹窗中还可以承载供用户选择“同意”或者“不同意”向终端设备提供个人信息的选择控件。
可以理解的是，上述通知和获取用户授权过程仅是示意性的，不对本公开的实现方式构成限定，其它满足相关法律法规的方式也可应用于本公开的实现方式中。
可以理解的是，本技术方案所涉及的数据(包括但不限于数据本身、数据的获取或使用)应当遵循相应法律法规及相关规定的要求。数据可以包括信息、参数和消息等，如切流指示信息。
以上描述仅为本公开的较佳实施例以及对所运用技术原理的说明。本领域技术人员应当理解，本公开中所涉及的公开范围，并不限于上述技术特征的特定组合而成的技术方案，同时也应涵盖在不脱离上述公开构思的情况下，由上述技术特征或其等同特征进行任意组合而形成的其它技术方案。例如上述特征与本公开中公开的(但不限于)具有类似功能的技术特征进行互相替换而形成的技术方案。
此外，虽然采用特定次序描绘了各操作，但是这不应当理解为要求这些操作以所示出的特定次序或以顺序次序执行来执行。在一定环境下，多任务和并行处理可能是有利的。同样 地，虽然在上面论述中包含了若干具体实现细节，但是这些不应当被解释为对本公开的范围的限制。在单独的实施例的上下文中描述的某些特征还可以组合地实现在单个实施例中。相反地，在单个实施例的上下文中描述的各种特征也可以单独地或以任何合适的子组合的方式实现在多个实施例中。
尽管已经采用特定于结构特征和/或方法逻辑动作的语言描述了本主题，但是应当理解所附权利要求书中所限定的主题未必局限于上面描述的特定特征或动作。相反，上面所描述的特定特征和动作仅仅是实现权利要求书的示例形式。

Claims (13)

  1. 一种音频处理方法，包括：
    获取多个音频采集设备采集的多个第一音频；
    基于所述多个第一音频和目标方向，确定角度特征，所述角度特征用于指示所述目标方向的音源在每个第一音频中所占的比重；
    基于所述多个第一音频和所述角度特征，确定所述目标方向相关联的第二音频，并播放所述第二音频。
  2. 根据权利要求1所述的方法，其中，所述基于所述多个第一音频和所述角度特征，确定所述目标方向相关联的第二音频，包括：
    基于所述多个第一音频和所述角度特征，确定多个第一目标音频和非目标音频，所述第一目标音频为所述第一音频在目标方向相关联的音频，所述非目标音频为所述第一音频在其它方向相关联的音频；
    基于所述多个第一音频、所述多个第一目标音频和所述非目标音频，确定所述第二音频。
  3. 根据权利要求2所述的方法，其中，所述基于所述多个第一音频和所述角度特征，确定多个第一目标音频和非目标音频，包括：
    基于所述多个第一音频和所述角度特征，确定所述多个第一音频相关联的多个第一权重和多个第二权重；
    基于所述多个第一权重、所述多个第二权重和所述多个第一音频，确定所述多个第一目标音频和所述非目标音频。
  4. 根据权利要求3所述的方法，其中，所述基于所述多个第一权重、所述多个第二权重和所述多个第一音频，确定所述多个第一目标音频和所述非目标音频，包括：
    基于所述多个第一权重和所述多个第一音频，确定所述多个第一目标音频；
    基于所述多个第二权重和所述多个第一音频，确定所述非目标音频。
  5. 根据权利要求3所述的方法，其中，基于所述多个第一音频和所述角度特征，确定每个第一音频相关联的第一权重和第二权重，包括：
    基于第一模型、所述多个第一音频和所述角度特征，确定所述多个第一权重和所述多个第二权重；
    其中，所述第一模型为对多组第一样本训练得到的，所述多组第一样本包括多个样本音频、所述多个样本音频相关联的角度特征和所述每个样本音频相关联的样本第一权重和样本第二权重。
  6. 根据权利要求2-5任一项所述的方法，其中，所述基于所述多个第一音频、所述多个第一目标音频和所述非目标音频，确定所述第二音频，包括：
    基于第二模型对所述多个第一目标音频和所述非目标音频进行处理，得到多个目标权重，所述目标权重为所述目标方向相关联的音频在所述第一音频的每个频点中所占的比重；
    基于所述多个目标权重和所述多个第一音频，确定多个子音频，并对所述多个子音频进行融合处理，得到所述第二音频；
    其中，所述第二模型为对多组第二样本训练得到的，所述多组第二样本包括多个样本第一目标音频、多个样本第一目标音频相关联的多个样本非目标音频和多个样本目标权重。
  7. 根据权利要求1-5任一项所述的方法，其中，所述基于所述多个第一音频和目标方向， 确定角度特征，包括：
    确定所述多个第一音频之间的相位差，得到多个第一相位差；
    确定所述目标方向相关联的第二相位差；
    基于所述第二相位差和所述多个第一相位差，确定所述角度特征。
  8. 根据权利要求7所述的方法，其中，所述基于所述第二相位差和所述多个第一相位差，确定所述角度特征，包括：
    确定所述第二相位差与每个第一相位差之间的余弦相似度，得到多个余弦相似度；
    对所述多个余弦相似度进行融合处理，得到所述角度特征。
  9. 一种音频处理装置，包括获取模块、第一确定模块、第二确定模块和播放模块，其中：
    所述获取模块用于，获取多个音频采集设备采集的多个第一音频；
    所述第一确定模块用于，基于所述多个第一音频和目标方向，确定角度特征，所述角度特征用于指示所述目标方向的音源在每个第一音频中所占的比重；
    所述第二确定模块用于，基于所述多个第一音频和所述角度特征，确定所述目标方向相关联的第二音频；
    所述播放模块用于，播放所述第二音频。
  10. 一种终端设备，包括：处理器和存储器；
    所述存储器存储计算机执行指令；
    所述处理器执行所述存储器存储的计算机执行指令，使得所述处理器执行如权利要求1-8任一项所述的音频处理方法。
  11. 一种计算机可读存储介质，所述计算机可读存储介质中存储有计算机执行指令，当处理器执行所述计算机执行指令时，实现如权利要求1-8任一项所述的音频处理方法。
  12. 一种计算机程序产品，包括计算机程序，所述计算机程序被处理器执行时实现如权利要求1-8任一项所述的音频处理方法。
  13. 一种计算机程序，所述计算机程序被处理器执行时实现权利要求1-8任一项所述的音频处理方法。
PCT/CN2023/113597 2022-10-18 2023-08-17 音频处理方法、装置及终端设备 WO2024082800A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211275143.4A CN117953908A (zh) 2022-10-18 2022-10-18 音频处理方法、装置及终端设备
CN202211275143.4 2022-10-18

Publications (1)

Publication Number Publication Date
WO2024082800A1 true WO2024082800A1 (zh) 2024-04-25

Family

ID=90736940

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/113597 WO2024082800A1 (zh) 2022-10-18 2023-08-17 音频处理方法、装置及终端设备

Country Status (2)

Country Link
CN (1) CN117953908A (zh)
WO (1) WO2024082800A1 (zh)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140044279A1 (en) * 2012-08-07 2014-02-13 Microsoft Corporation Multi-microphone audio source separation based on combined statistical angle distributions
CN110488225A (zh) * 2019-10-17 2019-11-22 南京雷鲨信息科技有限公司 声音方位的指示方法、装置、可读存储介质及移动终端
CN112382306A (zh) * 2020-12-02 2021-02-19 苏州思必驰信息科技有限公司 分离说话人音频的方法及装置
CN113014983A (zh) * 2021-03-08 2021-06-22 Oppo广东移动通信有限公司 视频播放方法、装置、存储介质及电子设备
CN113763982A (zh) * 2020-06-05 2021-12-07 阿里巴巴集团控股有限公司 音频处理方法、装置、电子设备及可读存储介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140044279A1 (en) * 2012-08-07 2014-02-13 Microsoft Corporation Multi-microphone audio source separation based on combined statistical angle distributions
CN110488225A (zh) * 2019-10-17 2019-11-22 南京雷鲨信息科技有限公司 声音方位的指示方法、装置、可读存储介质及移动终端
CN113763982A (zh) * 2020-06-05 2021-12-07 阿里巴巴集团控股有限公司 音频处理方法、装置、电子设备及可读存储介质
CN112382306A (zh) * 2020-12-02 2021-02-19 苏州思必驰信息科技有限公司 分离说话人音频的方法及装置
CN113014983A (zh) * 2021-03-08 2021-06-22 Oppo广东移动通信有限公司 视频播放方法、装置、存储介质及电子设备

Also Published As

Publication number Publication date
CN117953908A (zh) 2024-04-30

Similar Documents

Publication Publication Date Title
US11240050B2 (en) Online document sharing method and apparatus, electronic device, and storage medium
US20220386061A1 (en) Audio processing method and apparatus, readable medium, and electronic device
CN109947387B (zh) 音频采集方法、音频播放方法、系统、设备及存储介质
US11822854B2 (en) Automatic volume adjustment method and apparatus, medium, and device
WO2024109706A1 (zh) 媒体内容显示方法、装置、设备、可读存储介质及产品
US11886484B2 (en) Music playing method and apparatus based on user interaction, and device and storage medium
WO2024094130A1 (zh) 内容分享方法、装置、设备、计算机可读存储介质及产品
US11783865B2 (en) Method and apparatus for displaying video playback page, and electronic device and medium
US20240129598A1 (en) Mehtod, system and device for playing effect in live room
US20240007591A1 (en) Method and apparatus of interaction video connection, electronic device, and storage medium
WO2024139843A1 (zh) 跨语言的视频处理方法、装置、设备、介质及产品
CN113889140A (zh) 音频信号播放方法、装置和电子设备
WO2024082800A1 (zh) 音频处理方法、装置及终端设备
WO2020224294A1 (zh) 用于处理信息的系统、方法和装置
US20220215482A1 (en) User information processing method, apparatus, electronic device, and computer storage medium
WO2023284411A1 (zh) 应用于直播的音频的输入输出的切换方法、直播设备
US20240040191A1 (en) Livestreaming audio processing method and device
CN111460211A (zh) 音频信息播放方法、装置和电子设备
CN112261236B (zh) 一种在多人语音中用于静音处理的方法与设备
WO2024131576A1 (zh) 视频处理方法、装置及电子设备
WO2024120027A1 (zh) 语音处理方法、装置及电子设备
US20240119920A1 (en) Audio processing method, audio processing apparatus and device
CN117953907A (zh) 音频处理方法、装置及终端设备
WO2024104235A1 (zh) 音频处理方法、装置及电子设备
CN112019919B (zh) 视频贴纸的添加方法、装置及电子设备

Legal Events

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

Ref document number: 23878795

Country of ref document: EP

Kind code of ref document: A1