Intelligent sound box audio adjustment processing method

Abstract

The invention relates to the technical field of audio processing, in particular to an intelligent sound box audio adjustment processing method, which comprises the following steps of: S1, acquiring various noise data of different public areas, converting the noise data into audio data, recording decibels and waveforms of noise, establishing a noise library model, and training by using the noise data, arranging a noise detection module in the sound box, and inputting the noise library model into the noise detection module; by establishing the noise library and recording the sound waves of various noises, in the playing process of the sound box, if the noises exist, the processing method can immediately control the sound box to improve the volume of the audio, so that a user can hear the played music clearly in an environment with relatively loud noises, and if the sound of the sound box is too loud, the user can hear the music clearly. According to the method, the sound box can be immediately controlled to reduce the volume of the audio, so that the resident disturbing condition is avoided, a user does not need to adjust by keys, and the method is more convenient and intelligent.

Description

Technical Field

[0001] The present invention relates to the technical field of audio processing, and in particular to an audio adjustment processing method for an intelligent speaker. Background Art

[0002] Audio processing is an important work involving multiple fields and technologies. It has a wide range of applications in modern life. Audio processing covers a series of operations such as the collection, editing, enhancement, compression, and synthesis of audio signals. In the audio signal collection stage, professional microphones, recorders and other equipment are used to convert sound into digital signals, laying the foundation for subsequent processing. Bluetooth speakers refer to speaker devices with built-in Bluetooth chips that use Bluetooth connections instead of traditional wire connections. They are connected to Bluetooth playback devices such as mobile phones, tablets, and notebooks to achieve the purpose of convenience and speed, allowing users to avoid the entanglement of annoying wires and listen to music in various ways at ease. Bluetooth speakers are mainly portable speakers, and their appearance is generally small and portable. Bluetooth speaker technology has also gradually been valued and accepted by consumers for its convenience.

[0003] With the rise of square dancing, portable Bluetooth speakers are used for square dancing to entertain lovers physically and mentally, but nuisance incidents occur from time to time. The audio volume of existing Bluetooth speakers is controlled by buttons on the user's mobile phone. When the surrounding noise is loud, it is difficult to increase the volume and the music cannot be heard clearly. If the volume is too loud and difficult to turn down in time, it will disturb others. Summary of the invention

[0004] The purpose of the present invention is to provide an intelligent speaker audio adjustment processing method to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a method for adjusting audio of an intelligent speaker, the method comprising the following steps:

[0006] S1. Collect various noise data from different public areas, convert them into audio data, record the decibel number and waveform of the noise, establish a noise library model, use the noise data for training, set up a noise detection module in the speaker, and input the noise library model into the noise detection module;

[0007] S2, input the preset maximum decibel value A of the sound played by the speaker into the speaker max , and then use the speaker to play music, and the speaker will make a sound wave curve according to the played music, and input the sound wave curve into the filter module of the speaker;

[0008] S3, while playing music, the speaker receives surrounding sounds, including surrounding noise and the music played by the speaker itself. The received sounds are filtered out by the filtering module to remove the interference of the music, and only the noise is transmitted to the noise detection module. The noise detection module detects the decibel number of the noise, which is recorded as the decibel value A1;

[0009] S4, the speaker monitors the decibel value A2 of its own music at the same time. If A1>80-90% of A2, the control module in the speaker will control the speaker to increase the volume. max , the control module stops increasing or decreasing the volume.

[0010] Preferably, in step S1, the noise library model is established and the training steps are as follows:

[0011] S1.1. Input the sound wave, type, time and distance of the noise source of the noise data, and send it to the noise library model through the data transmission unit, and classify it according to the decibel number of the noise data;

[0012] S1.2, extracting part of the noise audio data to form a noise audio training set, performing time domain framing and feature extraction processing on the audio data in the noise audio training set, obtaining multiple frame feature vectors of the noise audio training set and multiple frame feature vectors of the reference channel, inputting them into the noise library model for training, and constructing a noise classification detection operation method;

[0013] S1.3. After the model training is completed, a parameter-free classification method is used to select several samples in the training set that are most similar to the noise audio data to form a test set. The test set is input into the model to obtain the result. If the result is inconsistent with the test set, the model is trained again until the training test is completed.

[0014] S1.4. Input the trained model into the noise detection module.

[0015] Preferably, in step S1.2, a differential sequence of the noise audio signal to be input is obtained through multiple noise audio frames, an autocorrelation function operation is performed according to the differential sequence, multiple correlation values ​​are obtained from the autocorrelation operation result, and the multiple correlation values ​​are combined into the correlation value sequence.

[0016] Preferably, in step S1, the places where noise is collected are various squares.

[0017] Preferably, in step S2, according to different city requirements, the value of Amax is divided into four categories, the value range of category 0 is 40-50dB, the value range of category 1 is 45-55dB, the value range of category 2 is 50-60dB, the value range of category 3 is 55-65dB, and the value range of category 4 is 55-70dB.

[0018] Preferably, in step S2, before using the speaker to play music, the noise level classification of the area is obtained.

[0019] Preferably, in step S3, a plurality of music songs are pre-stored inside the speaker, and the sound wave curve of the audio is drawn, and the music song audio signal is framed according to a preset frame rule to determine each original sub-audio signal in the music song audio signal.

[0020] Preferably, in step S3, the corresponding relationship between the music song audio signal and the simultaneous noise audio signal is obtained.

[0021] Preferably, in step S3, the speaker edits the collected noise audio data into an audio signal and sends it to the user end via a Bluetooth module.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] The present invention establishes a noise library to record the sound waves of various noises. During the playback of the speaker, if there is noise, the processing method can immediately control the speaker to increase the volume of the audio, so that the user can hear the music being played clearly in a noisy environment. If the sound of the speaker is too loud, the method can immediately control the speaker to lower the volume of the audio, thereby avoiding disturbing the public. Moreover, the user does not need to use buttons to adjust, which is more convenient and intelligent. DETAILED DESCRIPTION

[0024] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] Embodiment 1:

[0026] A method for adjusting audio of an intelligent speaker comprises the following steps:

[0027] S1. Collect various noise data from different public areas, convert them into audio data, record the decibel number and waveform of the noise, establish a noise library model, use the noise data for training, set up a noise detection module in the speaker, and input the noise library model into the noise detection module;

[0028] The steps to establish the noise library model and training are as follows:

[0029] S1.1. Input the sound wave, type, time and distance of the noise source of the noise data, and send it to the noise library model through the data transmission unit, and classify it according to the decibel number of the noise data;

[0030] S1.2, extracting part of the noise audio data to form a noise audio training set, performing time domain framing and feature extraction processing on the audio data in the noise audio training set, obtaining multiple frame feature vectors of the noise audio training set and multiple frame feature vectors of the reference channel, inputting them into the noise library model for training, and constructing a noise classification detection operation method;

[0031] S1.3. After the model training is completed, a parameter-free classification method is used to select several samples in the training set that are most similar to the noise audio data to form a test set. The test set is input into the model to obtain the result. If the result is inconsistent with the test set, the model is trained again until the training test is completed.

[0032] S1.4, input the trained model into the noise detection module;

[0033] S2, input the preset maximum decibel value A of the sound played by the speaker into the speaker max ,According to different city requirements, the Amax values ​​are divided into four categories, the value range of category 0 is 40-50dB, the value range of category 1 is 45-55dB, the value range of category 2 is 50-60dB, the value range of category 3 is 55-65dB, and the value range of category 4 is 55-70dB. Then, the speaker is used to play music, and the speaker makes a sound wave curve according to the played music, and the sound wave curve is input into the filter module of the speaker;

[0034] S3, while playing music, the speaker receives surrounding sounds, including surrounding noise and the music played by the speaker itself. The received sound is filtered out by the filter module to remove the interference of the music, and only the noise is transmitted to the noise detection module. The noise detection module detects the decibel number of the noise, which is recorded as the decibel value A 1， Pre-store multiple music songs in the speaker, draw the sound wave curve of the audio, perform frame processing on the music song audio signal according to the preset frame rule, and determine each original sub-audio signal in the music song audio signal;

[0035] S4, the speaker monitors the decibel value A2 of its own music at the same time. If A1>80-90% of A2, the control module in the speaker will control the speaker to increase the volume. max , the control module stops increasing or decreasing the volume.

[0036] Embodiment 2:

[0037] A method for adjusting audio of an intelligent speaker comprises the following steps:

[0038] S1. Collect various noise data from different public areas, convert them into audio data, record the decibel number and waveform of the noise, establish a noise library model, use the noise data for training, set up a noise detection module in the speaker, and input the noise library model into the noise detection module;

[0039] The steps to establish the noise library model and training are as follows:

[0040] S1.1. Input the sound wave, type, time and distance of the noise source of the noise data, and send it to the noise library model through the data transmission unit, and classify it according to the decibel number of the noise data;

[0041] S1.2. Extract part of the noise audio data to form a noise audio training set, perform time domain framing and feature extraction processing on the audio data in the noise audio training set, obtain multiple frame feature vectors of the noise audio training set and multiple frame feature vectors of the reference channel, input them into the noise library model for training, construct a noise classification detection operation method, obtain a differential sequence of the noise audio signal to be input through multiple noise audio frames, perform autocorrelation function operation according to the differential sequence, obtain multiple correlation values ​​from the autocorrelation operation result, and form the multiple correlation values ​​into the correlation value sequence;

[0042] S1.3. After the model training is completed, a parameter-free classification method is used to select several samples in the training set that are most similar to the noise audio data to form a test set. The test set is input into the model to obtain the result. If the result is inconsistent with the test set, the model is trained again until the training test is completed.

[0043] S1.4, input the trained model into the noise detection module;

[0044] S2, input the preset maximum decibel value A of the sound played by the speaker into the speaker max ,According to different city requirements, the value of Amax is divided into four categories, the value range of category 0 is 40-50dB, the value range of category 1 is 45-55dB, the value range of category 2 is 50-60dB, the value range of category 3 is 55-65dB, and the value range of category 4 is 55-70dB.,The corresponding relationship between the audio signal of the music song and the simultaneous noise audio signal is obtained, and then the music is played by the speaker. At the same time, the speaker makes a sound wave curve according to the played music, and the sound wave curve is input into the filter module of the speaker;

[0045] S3, while playing music, the speaker receives surrounding sounds, including surrounding noise and the music played by the speaker itself. The received sound is filtered out by the filter module to remove the interference of the music, and only the noise is transmitted to the noise detection module. The noise detection module detects the decibel number of the noise, which is recorded as the decibel value A 1， Pre-store multiple music songs in the speaker, draw the sound wave curve of the audio, perform frame processing on the music song audio signal according to the preset frame rule, and determine each original sub-audio signal in the music song audio signal;

[0046] S4, the speaker monitors the decibel value A2 of its own music at the same time. If A1>80-90% of A2, the control module in the speaker will control the speaker to increase the volume. max , the control module stops increasing or decreasing the volume.

[0047] Embodiment three:

[0048] A method for adjusting audio of an intelligent speaker comprises the following steps:

[0049] S1. Collect various noise data from different public areas, convert them into audio data, record the decibel number and waveform of the noise, establish a noise library model, use the noise data for training, set up a noise detection module in the speaker, and input the noise library model into the noise detection module. The places where noise is collected are various squares;

[0050] The steps to establish the noise library model and training are as follows:

[0051] S1.1. Input the sound wave, type, time and distance of the noise source of the noise data, and send it to the noise library model through the data transmission unit, and classify it according to the decibel number of the noise data;

[0052] S1.2. Extract part of the noise audio data to form a noise audio training set, perform time domain framing and feature extraction processing on the audio data in the noise audio training set, obtain multiple frame feature vectors of the noise audio training set and multiple frame feature vectors of the reference channel, input them into the noise library model for training, construct a noise classification detection operation method, obtain a differential sequence of the noise audio signal to be input through multiple noise audio frames, perform autocorrelation function operation according to the differential sequence, obtain multiple correlation values ​​from the autocorrelation operation result, and form the multiple correlation values ​​into the correlation value sequence;

[0053] S1.3. After the model training is completed, a parameter-free classification method is used to select several samples in the training set that are most similar to the noise audio data to form a test set. The test set is input into the model to obtain the result. If the result is inconsistent with the test set, the model is trained again until the training test is completed.

[0054] S1.4, input the trained model into the noise detection module;

[0055] S2. Before using the speaker to play music, obtain the noise level classification of the area and input the maximum decibel value A of the sound played by the speaker into the speaker in advance. max ,According to different city requirements, the value of Amax is divided into four categories, the value range of category 0 is 40-50dB, the value range of category 1 is 45-55dB, the value range of category 2 is 50-60dB, the value range of category 3 is 55-65dB, and the value range of category 4 is 55-70dB.,The corresponding relationship between the audio signal of the music song and the simultaneous noise audio signal is obtained, and then the music is played by the speaker. At the same time, the speaker makes a sound wave curve according to the played music, and the sound wave curve is input into the filter module of the speaker;

[0056] S3, while playing music, the speaker receives surrounding sounds, including surrounding noise and the music played by the speaker itself. The received sound is filtered out by the filter module to remove the interference of the music, and only the noise is transmitted to the noise detection module. The noise detection module detects the decibel number of the noise, which is recorded as the decibel value A 1， Pre-store multiple music songs in the speaker, draw the sound wave curve of the audio, perform frame processing on the music song audio signal according to the preset frame division rule, determine each original sub-audio signal in the music song audio signal, and the speaker edits the collected noise audio data into an audio signal and sends it to the user end through the Bluetooth module;

[0057] S4, the speaker monitors the decibel value A2 of its own music at the same time. If A1>80-90% of A2, the control module in the speaker will control the speaker to increase the volume. max , the control module stops increasing or decreasing the volume.

[0058] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for adjusting audio of an intelligent speaker, characterized in that: The method comprises the following steps: S1. Collect various noise data from different public areas, convert them into audio data, record the decibel number and waveform of the noise, establish a noise library model, use the noise data for training, set up a noise detection module in the speaker, and input the noise library model into the noise detection module; S2, input the preset maximum decibel value A of the sound played by the speaker into the speaker max , and then use the speaker to play music, and the speaker will make a sound wave curve according to the played music, and input the sound wave curve into the filter module of the speaker; S3, while playing music, the speaker receives surrounding sounds, including surrounding noise and the music played by the speaker itself. The received sounds are filtered out by the filtering module to remove the interference of the music, and only the noise is transmitted to the noise detection module. The noise detection module detects the decibel number of the noise, which is recorded as the decibel value A1; S4, the speaker monitors the decibel value A2 of its own music at the same time. If A1>80-90% of A2, the control module in the speaker will control the speaker to increase the volume. max , the control module stops increasing or decreasing the volume.

2. The method for adjusting audio of an intelligent speaker according to claim 1, characterized in that: In step S1, the noise library model is established and the training steps are as follows: S1.

1. Input the sound wave, type, time and distance of the noise source of the noise data, and send it to the noise library model through the data transmission unit, and classify it according to the decibel number of the noise data; S1.2, extracting part of the noise audio data to form a noise audio training set, performing time domain framing and feature extraction processing on the audio data in the noise audio training set, obtaining multiple frame feature vectors of the noise audio training set and multiple frame feature vectors of the reference channel, inputting them into the noise library model for training, and constructing a noise classification detection operation method; S1.

3. After the model training is completed, a parameter-free classification method is used to select several samples in the training set that are most similar to the noise audio data to form a test set. The test set is input into the model to obtain the result. If the result is inconsistent with the test set, the model is trained again until the training test is completed. S1.

4. Input the trained model into the noise detection module.

3. The method for adjusting audio frequency of an intelligent speaker according to claim 2, characterized in that: In the step S1.2, a differential sequence of the noise audio signal to be input is obtained through multiple noise audio frames, an autocorrelation function operation is performed according to the differential sequence, multiple correlation values ​​are obtained from the autocorrelation operation result, and the multiple correlation values ​​are combined into the correlation value sequence.

4. The method for adjusting audio of an intelligent speaker according to claim 1, characterized in that: In the step S1, the places where noise is collected are various squares.

5. The method for adjusting audio of an intelligent speaker according to claim 1, characterized in that: In step S2, according to different city requirements, the value of Amax is divided into four categories, the value range of category 0 is 40-50dB, the value range of category 1 is 45-55dB, the value range of category 2 is 50-60dB, the value range of category 3 is 55-65dB, and the value range of category 4 is 55-70dB.

6. The method for adjusting audio of an intelligent speaker according to claim 1, characterized in that: In step S2, before using the speaker to play music, the noise level classification of the area is obtained.

7. The method for adjusting audio of an intelligent speaker according to claim 1, characterized in that: In step S3, a plurality of music songs are stored in advance inside the speaker, and the sound wave curve of the audio is drawn, and the music song audio signal is framed according to a preset frame rule to determine each original sub-audio signal in the music song audio signal.

8. The method for adjusting audio of an intelligent speaker according to claim 7, characterized in that: In the step S3, the corresponding relationship between the music song audio signal and the simultaneous noise audio signal is obtained. In the step S3, the speaker edits the collected noise audio data into an audio signal and sends it to the user end through the Bluetooth module.