CN111785298A - Acoustic performance testing method and device, electronic equipment, computer readable medium - Google Patents

Abstract

The disclosure provides an acoustic performance testing method of voice equipment, belongs to the field of voice recognition, and can be used for performance testing of an artificial intelligent sound box. The method comprises the following steps: acquiring a first audio signal recorded by the voice equipment, wherein the first audio signal is an audio signal played by the voice equipment or an environmental noise signal generated by an external noise source; acquiring a second audio signal generated after the voice equipment performs signal processing on the first audio signal; generating first acoustic performance data of the speech device from the first audio signal and the second audio signal; and generating a first test result according to the first acoustic performance data and preset first standard data. The disclosure also provides an acoustic performance testing device, an electronic apparatus, and a computer-readable medium.

Description

Acoustic performance testing method and device, electronic equipment, computer readable medium

technical field

The embodiments of the present disclosure relate to the technical field of speech recognition, and in particular, to a method and apparatus for testing acoustic performance of speech equipment, electronic equipment, and computer-readable media.

Background technique

Intelligent interactive devices, especially voice interactive devices, have been widely used in people's daily life, work, and even production processes, for example, in-vehicle voice recognition systems in the automotive field.

Existing solutions for evaluating voice interaction devices usually evaluate the device's speech recognition rate and effect, but lack a solution for performance testing of the underlying audio hardware of the voice interaction device.

SUMMARY OF THE INVENTION

Embodiments of the present disclosure provide a method and apparatus for testing acoustic performance of a voice device, an electronic device, and a computer-readable medium.

In a first aspect, an embodiment of the present disclosure provides an acoustic performance testing method, the acoustic performance testing method comprising:

acquiring a first audio signal recorded by the voice device, where the first audio signal is an audio signal played by the voice device itself or an environmental noise signal generated by an external noise source;

acquiring a second audio signal generated after the voice device performs signal processing on the first audio signal;

generating first acoustic performance data of the speech device from the first audio signal and the second audio signal;

A first test result is generated according to the first acoustic performance data and the preset first standard data.

In a second aspect, an embodiment of the present disclosure provides an acoustic performance testing device for a voice device, the acoustic performance testing device comprising:

a first acquisition module, configured to acquire a first audio signal recorded by the voice device, where the first audio signal is an audio signal played by the voice device itself or an environmental noise signal generated by an external noise source;

a second acquisition module, configured to acquire a second audio signal generated after the voice device performs signal processing on the first audio signal;

a performance parameter generation module, configured to generate first acoustic performance data of the voice device according to the first audio signal and the second audio signal;

An evaluation module, configured to generate a first test result according to the first acoustic performance data and the preset first standard data.

In a third aspect, an embodiment of the present disclosure provides an electronic device, which includes:

one or more processors;

A memory having one or more programs stored thereon that, when executed by the one or more processors, cause the one or more processors to implement the acoustics provided by any of the above embodiments performance testing methods;

One or more I/O interfaces, connected between the processor and the memory, are configured to realize the information interaction between the processor and the memory.

In a fourth aspect, an embodiment of the present disclosure provides a computer-readable medium on which a computer program is stored, wherein when the computer program is executed, the acoustic performance testing method provided by any of the foregoing embodiments is implemented.

The method and device, electronic device, and computer-readable medium for testing the acoustic performance of a voice device provided by the embodiments of the present disclosure can obtain test results under different test scenarios by evaluating the performance of the underlying audio hardware of the voice device in different test scenarios. , the test results are used as the quality evaluation standard of the underlying audio hardware of the voice device, which can provide hardware guarantee for the back-end speech recognition algorithm, and the quality risk evaluation of the underlying audio hardware can be pre-positioned, which can save the research and development cost of the product.

Description of drawings

The accompanying drawings are used to provide a further understanding of the embodiments of the present disclosure, and constitute a part of the specification, and together with the embodiments of the present disclosure, they are used to explain the present disclosure, and do not constitute a limitation to the present disclosure. The above and other features and advantages will become more apparent to those skilled in the art by describing detailed example embodiments with reference to the accompanying drawings, in which:

1 is a flowchart of a method for testing acoustic performance of a voice device according to an embodiment of the present disclosure;

Fig. 2 is a flow chart of a specific embodiment of step 13 in Fig. 1;

Fig. 3 is a flow chart of a specific implementation manner of step 14 in Fig. 1;

4 is a flowchart of another acoustic performance testing method provided by an embodiment of the present disclosure;

Fig. 5 is a flow chart of a specific implementation of step 23 in Fig. 4;

Fig. 6 is a flow chart of another specific implementation manner of step 23 in Fig. 4;

Fig. 7 is a flow chart of a specific implementation of step 24 in Fig. 4;

Fig. 8 is a flow chart of another specific implementation manner of step 24 in Fig. 4;

FIG. 9 is a block diagram of a composition of an apparatus for testing acoustic performance of a voice device provided by an embodiment of the present disclosure;

FIG. 10 is a block diagram of an electronic device according to an embodiment of the present disclosure.

Detailed ways

In order for those skilled in the art to better understand the technical solutions of the present disclosure, the method and apparatus, electronic device, and computer-readable medium for testing the acoustic performance of a speech device provided by the present disclosure are described in detail below with reference to the accompanying drawings.

Example embodiments are described more fully hereinafter with reference to the accompanying drawings, but which may be embodied in different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

Various embodiments of the present disclosure and various features of the embodiments may be combined with each other without conflict.

As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items.

The terminology used herein is used to describe particular embodiments only and is not intended to limit the present disclosure. As used herein, the singular forms "a" and "the" are intended to include the plural forms as well, unless the context clearly dictates otherwise. It will also be understood that when the terms "comprising" and/or "made of" are used in this specification, the stated features, integers, steps, operations, elements and/or components are specified to be present, but not precluded or Add one or more other features, integers, steps, operations, elements, components, and/or groups thereof.

Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art. It will also be understood that terms such as those defined in common dictionaries should be construed as having meanings consistent with their meanings in the context of the related art and the present disclosure, and will not be construed as having idealized or over-formal meanings, unless expressly so limited herein.

FIG. 1 is a flowchart of an acoustic performance testing method of a voice device provided by an embodiment of the present disclosure. As shown in FIG. 1 , the method may be performed by an acoustic performance testing apparatus, and the apparatus may be implemented by software and/or hardware. Implementation, the device can be integrated in electronic equipment such as a server. The acoustic performance testing method includes step 11 and step 14 .

Step 11: Acquire a first audio signal recorded by the voice device, where the first audio signal is an audio signal played by the voice device itself or an environmental noise signal generated by an external noise source.

In the embodiments of the present disclosure, the voice device is an intelligent terminal, device, or system that can provide users with intelligent voice interactive services. For example, the voice device may be a vehicle-mounted voice recognition system, a smart audio system, a smart video speaker, a smart story machine, or an intelligent interactive platform. Wait. Wherein, the underlying audio hardware of the voice device includes but is not limited to: a sound playback component, a sound pickup component, and an audio digital signal processing (Digital Signal Process, DSP) chip, wherein the sound playback component may include a speaker, and the sound pickup component may include a microphone ( MIC) array.

In a test scenario, when the test environment is kept quiet, the audio signal used for testing is played by the sound playback component of the voice device itself, and the audio signal used for the test is recorded by the sound pickup component of the voice device itself, and the recorded audio signal is used as For the first audio signal, in step 11, the first audio signal recorded by the sound pickup component of the voice device is obtained.

In another test scenario, in the test environment, the voice device itself is kept in a quiet state (that is, in a state where no audio signal is emitted), and an external noise source is used to generate environmental noise signals (such as white noise, air conditioning noise, driving wind noise, etc.) , and use the sound pickup component of the voice device to perform recording, and the recorded audio signal is used as the first audio signal. In step 11, the first audio signal recorded by the sound pickup component of the voice device is obtained.

Step 12: Acquire a second audio signal generated after the voice device performs signal processing on the first audio signal.

In the embodiment of the present disclosure, after the voice device records the first audio signal, the recorded first audio signal will be transmitted to an audio digital signal processing chip (DSP), and the audio digital signal processing chip will perform signal processing on the first audio signal For example, signal processing such as noise reduction, acoustic echo cancellation (AEC), and electrical noise (noise generated by the circuit) is performed to perform hardware noise cancellation, thereby obtaining a second audio signal generated after the signal processing. In step 12, a second audio signal generated by performing signal processing on the first audio signal is obtained from the audio digital signal processing chip of the voice device.

Step 13: Generate first acoustic performance data of the voice device according to the first audio signal and the second audio signal.

In the embodiment of the present disclosure, the audio digital signal processing chip performs signal processing on the first audio signal, which theoretically is to eliminate the audio signal played by the voice device itself or to eliminate the environmental noise signal generated by the external noise source, that is, theoretically signal processing The second audio signal generated later does not contain the first audio signal. In the embodiment of the present disclosure, by comparing and analyzing the first audio signal recorded by the voice device and the second audio signal generated after signal processing, the signal processing capability of the underlying hardware of the voice device in different test scenarios can be analyzed.

Among them, in the wake-up scene of voice interruption, that is, the test scene in which the voice device itself plays audio signals, when the voice device is woken up, the audio signal (such as music) played by the voice device itself will interfere with the wake-up of the voice device. Therefore, in order to test The noise reduction ability of the voice device to the noise (first audio signal) generated by itself, by acquiring the first audio signal played by the voice device itself and the second audio signal generated after processing the first audio signal, and by comparative analysis, In this way, the signal processing capability of the underlying hardware of the voice device in the wake-up scene from voice interruption can be analyzed and obtained.

After the voice device wakes up, it will no longer play the audio signal. At this time, the noise generated by the external noise source is the main factor affecting the word accuracy and sentence accuracy of the voice recognition. Therefore, in the quasi-sentence recognition scene, that is, the test scene in which the environmental noise signal is generated by the external noise source, the environmental noise signal generated by the external noise source will interfere with the speech recognition of the voice device. The noise reduction capability of the noise (first audio signal), by obtaining the first audio signal generated by the external noise source and the second audio signal generated by the voice device after processing the first audio signal, and through comparative analysis, so as to be able to analyze The signal processing capability of the underlying hardware of the speech device in the scene of word quasi-sentence quasi-recognition is obtained.

FIG. 2 is a flowchart of a specific implementation manner of step 13 in FIG. 1 . As shown in FIG. 2 , in some embodiments, step 13 includes step 131 and step 132 .

Step 131: Perform fast Fourier transform calculation processing on the first audio signal and the second audio signal respectively, and calculate the energy value corresponding to the first audio signal and the energy value corresponding to the second audio signal.

Specifically, a fast Fourier transform (FFT) calculation process is performed on the first audio signal to obtain a first frequency spectrum corresponding to the first audio signal. The value of each frequency point in the first frequency spectrum is a complex number, and the first frequency spectrum is calculated. The modulo value of the complex number corresponding to each frequency point in the first frequency spectrum, and statistical average processing is performed on the modulo value corresponding to each frequency point in the first frequency spectrum to obtain a first statistical average value, and the first statistical average value corresponds to the first audio signal. energy value.

Similarly, perform fast Fourier transform (FFT) calculation processing on the second audio signal to obtain a second frequency spectrum corresponding to the second audio signal. The value of each frequency point in the second frequency spectrum is a complex number, and the second frequency spectrum is calculated. The modulo value of the complex number corresponding to each frequency point in the second frequency spectrum, and statistical average processing is performed on the modulo value corresponding to each frequency point in the second frequency spectrum to obtain a second statistical average value, and the second statistical average value corresponds to the second audio signal. energy value.

Step 132 , using the energy value corresponding to the first audio signal and the energy value corresponding to the second audio signal to calculate the noise cancellation amount, and the first acoustic performance data includes the noise cancellation amount.

The noise removal amount Y is the difference between the energy value S1 corresponding to the first audio signal and the energy value S2 corresponding to the second audio signal, that is, Y=S1-S2.

For example, in the case where the first audio signal is an audio signal played by the voice device itself, in step 132, the energy value corresponding to the first audio signal and the energy value corresponding to the second audio signal are used to calculate the first noise removal amount Y1, the first acoustic performance data includes the first noise cancellation amount Y1.

For example, in the case where the first audio signal is an external noise source to generate an environmental noise signal, in step 132, the second noise elimination amount is calculated by using the energy value corresponding to the first audio signal and the energy value corresponding to the second audio signal Y2, the first acoustic performance data includes the second noise cancellation amount Y2.

Step 14: Generate a first test result according to the first acoustic performance data and the preset first standard data.

Specifically, the first test result is obtained by comparing the first acoustic performance data with the preset first standard data. Wherein, the first acoustic performance data includes a noise cancellation amount, and the first standard data includes a first evaluation standard value.

FIG. 3 is a flowchart of a specific implementation manner of step 14 in FIG. 1 . As shown in FIG. 3 , in some embodiments, step 14 includes step 141 and step 142 .

Step 141: Compare the noise elimination amount with the first evaluation standard value to obtain a comparison result.

Step 142: Generate a first test result according to the comparison result between the noise elimination amount and the first evaluation standard value.

It should be noted that, in different test scenarios, the value of the first evaluation criterion value may be different. For example, in the case where the first audio signal is the audio signal played by the voice device itself, the first evaluation criterion value can be set to A1. , and compare the first noise elimination amount Y1 with the first evaluation standard value A1 to obtain a comparison result. In step 142, when the comparison result is that the first noise elimination amount Y1 is greater than or equal to the first evaluation standard value A1, then it is determined that the test result of the signal processing capability (hardware noise reduction capability) of the voice device in this situation is qualified, The first test result includes information indicating that the test is qualified, otherwise the test result is judged to be unqualified, and the first test result includes information indicating that the test is unqualified.

In the case where the first audio signal is an environmental noise signal generated by an external noise source, the first criterion value can be set to A2. After obtaining the second noise elimination amount Y2 in this situation through step 132, in step 141, The second noise elimination amount Y2 is compared with the first evaluation standard value A2 to obtain a comparison result. In step 142, when the comparison result is that the second noise elimination amount Y2 is greater than or equal to the first evaluation standard value A2, then it is determined that the signal processing capability (hardware noise reduction capability) test result of the voice device in this situation is qualified, The first test result includes information indicating that the test is qualified, otherwise the test result is judged to be unqualified, and the first test result includes information indicating that the test is unqualified.

It can be understood that the above steps 11 to 14 are the process of testing the signal processing (ie noise reduction, echo cancellation, electrical noise cancellation, etc.) capabilities of the underlying audio hardware of the voice device in some test scenarios, as shown in Figure 1. The process of testing the signal processing capability of the underlying audio hardware of the voice device in some test scenarios is presented.

In some embodiments, the acoustic performance testing method further includes a process of testing the ability of the underlying audio hardware of the speech device to eliminate (ie, noise reduction, blind source separation, etc.) ineffective audio parts. FIG. 4 is a flowchart of another acoustic performance testing method provided by an embodiment of the present disclosure. As shown in FIG. 4 , the interactive system evaluation method further includes:

Step 21: Obtain a fourth audio signal generated after the voice device performs signal processing on the recorded third audio signal, where the third audio signal is an audio signal played by an external audio playback device.

In some embodiments, the external audio playback device may be an artificial mouth, and the external audio playback device may be placed in each sound area of the voice device to simulate actual usage scenarios of the voice device. In a test scenario, in the test environment, the voice device itself is kept in a quiet state (that is, the state where no audio signal is emitted), the audio signal is played by an external audio playback device, and the sound pickup component of the voice device itself is used for recording, so The recorded audio signal is the third audio signal.

After the voice device records the third audio signal, the recorded third audio signal will be transmitted to the audio digital signal processing chip (DSP), and the audio digital signal processing chip will perform signal processing on the third audio signal, such as noise reduction, blindness Signal processing such as source separation is performed to obtain a fourth audio signal generated after the signal processing. In step 21, a fourth audio signal generated by performing signal processing on the third audio signal is obtained from the audio digital signal processing chip of the voice device.

Step 22, obtain the sixth audio signal generated after the voice device performs signal processing on the recorded fifth audio signal, and the fifth audio signal includes the audio signal played by the voice device itself and the audio signal played by an external audio playback device, or the fifth audio signal. Audio signals include ambient noise signals generated by external noise sources and audio signals played by external audio playback devices.

In a test scenario, the audio signal is played by the sound playback component of the voice device itself, the audio signal (which is the same audio signal as the aforementioned third audio signal) is played by the external audio playback device, and the audio signal is played by the sound pickup component of the voice device itself. In recording, the recorded audio signal is the fifth audio signal. At this time, the fifth audio signal includes the audio signal played by the voice device itself and the audio signal played by the external audio playback device.

In this test scenario, after the voice device records the fifth audio signal, the recorded fifth audio signal will be transmitted to the audio digital signal processing chip (DSP), and the audio digital signal processing chip will perform signal processing on the fifth audio signal, For example, signal processing such as noise reduction and blind source separation is performed to obtain the sixth audio signal generated after the signal processing. In step 22, a sixth audio signal generated by performing signal processing on the fifth audio signal is obtained from the audio digital signal processing chip of the voice device.

In another test scenario, an external audio playback device is used to play an audio signal (the same audio signal as the aforementioned third audio signal), an external noise source is used to generate an environmental noise signal, and the sound pickup component of the voice device is used for recording, The recorded audio signal is a fifth audio signal, and at this time, the fifth audio signal includes an audio signal played by an external audio playback device and an environmental noise signal generated by an external noise source.

In this test scenario, after the voice device records the fifth audio signal, the recorded fifth audio signal will be transmitted to the audio digital signal processing chip (DSP), and the audio digital signal processing chip will perform signal processing on the fifth audio signal, For example, signal processing such as noise reduction and blind source separation is performed to obtain the sixth audio signal generated after the signal processing. In step 22, a sixth audio signal generated by performing signal processing on the fifth audio signal is obtained from the audio digital signal processing chip of the voice device.

Step 23: Generate second acoustic performance data of the voice device according to the fourth audio signal and the sixth audio signal.

In the embodiment of the present disclosure, the audio signal (that is, the third audio signal) played by the external audio playback device (the artificial mouth) is the effective audio part, and other noise signals (the audio signal played by the voice device itself, the environmental noise generated by the external noise source) signal, etc.) are inactive audio parts.

Therefore, the audio digital signal processing chip performs signal processing on the third audio signal, which theoretically eliminates the non-effective audio part, that is, theoretically, the fourth audio signal generated after the signal processing only contains the third audio signal and does not contain other audio signals. noise signal. In the same way, the audio digital signal processing chip performs signal processing on the fifth audio signal, which theoretically eliminates the non-effective audio part, that is, theoretically, the sixth audio signal generated after the signal processing only contains the third audio signal, but does not contain the third audio signal. other noise signals.

In the embodiment of the present disclosure, the fourth audio signal generated by performing signal processing on the third audio signal and the sixth audio signal generated by performing signal processing on the fifth audio signal can be analyzed and obtained by comparative analysis. The ability of the underlying hardware to eliminate inactive audio parts.

In a test scenario (such as a voice interrupt wake-up scenario), in the case that the voice device itself does not play any audio signal, the external audio playback device (the artificial mouth) is made to play the audio signal to obtain the fourth audio signal; then in the In the case that the voice device itself plays the audio signal, the external audio playing device (the artificial mouth) is made to play the audio signal, so as to obtain the sixth audio signal. By comparing and analyzing the fourth audio signal and the sixth audio signal obtained by the voice device in this test scenario, the ability of the underlying hardware of the voice device to eliminate ineffective audio parts in this test scenario can be analyzed.

In another test scenario (such as the quasi-sentence recognition scene), under the situation that the voice device itself does not play any audio signal, the external audio playing device (the artificial mouth) is made to play the audio signal to obtain the fourth audio signal; Then, in the case that the external noise source generates the ambient noise signal, the external audio playing device (the artificial mouth) is caused to play the audio signal to obtain the sixth audio signal. By comparing and analyzing the fourth audio signal and the sixth audio signal recorded by the voice device in this test scenario, it is possible to analyze the ability of the underlying hardware of the voice device to eliminate ineffective audio parts in this test scenario.

FIG. 5 is a flowchart of a specific implementation of step 23 in FIG. 4 . As shown in FIG. 5 , in some embodiments, the fifth audio signal includes an audio signal played by the voice device itself and an audio signal played by an external audio playback device. In the case of an audio signal, step 23 includes step 231a and step 232a.

Step 231a: Perform fast Fourier transform processing on the fourth audio signal and the sixth audio signal, respectively, to obtain an energy value corresponding to the fourth audio signal and an energy value corresponding to the sixth audio signal.

Specifically, a fast Fourier transform (FFT) calculation process is performed on the fourth audio signal to obtain a fourth frequency spectrum corresponding to the fourth audio signal. The value of each frequency point in the fourth frequency spectrum is a complex number, and the fourth frequency spectrum is calculated. The modulo value of the complex number corresponding to each frequency point in the fourth frequency spectrum, and statistical average processing is performed on the modulo value corresponding to each frequency point in the fourth frequency spectrum to obtain a fourth statistical average value, and the fourth statistical average value corresponds to the fourth audio signal. energy value.

Similarly, perform fast Fourier transform (FFT) calculation processing on the sixth audio signal to obtain a sixth frequency spectrum corresponding to the sixth audio signal. The value of each frequency point in the sixth frequency spectrum is a complex number, and the sixth frequency spectrum is calculated. The modulo value of the complex number corresponding to each frequency point in the sixth spectrum, and the modulo value corresponding to each frequency point in the sixth frequency spectrum is statistically averaged to obtain a sixth statistical average value, which is the sixth audio signal corresponding to the energy value.

Step 232a, using the energy value corresponding to the fourth audio signal and the energy value corresponding to the sixth audio signal to calculate the energy attenuation amount, and the second acoustic performance data includes the energy attenuation amount.

The energy attenuation Y3 is the difference between the energy value S3 corresponding to the fourth audio signal and the energy value S4 corresponding to the sixth audio signal, that is, Y3=S3-S4.

FIG. 6 is a flowchart of another specific implementation of step 23 in FIG. 4 . As shown in FIG. 7 , in some embodiments, the fifth audio signal includes an environmental noise signal generated by an external noise source and an external audio playback device. In the case of a played audio signal, step 23 includes step 231b and step 232b.

Step 231b: Perform fast Fourier transform processing on the fourth audio signal and the sixth audio signal, respectively, to obtain an energy value corresponding to the fourth audio signal and an energy value corresponding to the sixth audio signal.

For the calculation process of the energy value corresponding to the fourth audio signal and the energy value corresponding to the sixth audio signal, refer to the above description of step 231a, and details are not repeated here.

Step 232b, using the energy value corresponding to the fourth audio signal and the energy value corresponding to the sixth audio signal to calculate the amount of speech distortion, and the second acoustic performance data includes the amount of speech distortion.

Specifically, using the energy value S3 corresponding to the fourth audio signal, the energy value S4 corresponding to the sixth audio signal, and the preset speech distortion amount SDR formula: SDR=10log ₁₀ [(S3)/(S4-S3)], calculate Get the speech distortion amount SDR.

Step 24: Generate a second test result according to the second acoustic performance data and the preset second standard data.

Specifically, the second test result is obtained by comparing the second acoustic performance data with the preset second standard data.

Wherein, in the case where the fifth audio signal includes the audio signal played by the voice device itself and the audio signal played by the external audio playback device, the second acoustic performance data includes the energy attenuation Y3, and the second standard data includes the second evaluation standard value B .

In the case where the fifth audio signal includes an ambient noise signal generated by an external noise source and an audio signal played by an external audio playback device, the second acoustic performance data includes the amount of speech distortion SDR, and the second standard data includes a third evaluation standard value C.

FIG. 7 is a flowchart of a specific implementation of step 24 in FIG. 4. As shown in FIG. 8, in some embodiments, the fifth audio signal includes an audio signal played by the voice device itself and an audio signal played by an external audio playback device. In the case of an audio signal, the second acoustic performance data includes the energy attenuation amount Y3, the second standard data includes the second evaluation standard value B, and step 24 includes steps 241a and 242a.

Step 241a, compare the energy attenuation with the second evaluation standard value to obtain a comparison result.

Step 242a: Generate a second test result according to the comparison result between the energy attenuation and the second evaluation standard value.

Specifically, when the comparison result is that the energy attenuation Y3 is less than the second evaluation standard value B, then the test result of the voice device's ability to eliminate ineffective audio in this situation is judged to be qualified, and the second test result includes an indication test Qualified information, otherwise the test result is judged to be unqualified, and the second test result includes information indicating that the test is unqualified.

FIG. 8 is a flowchart of another specific implementation of step 24 in FIG. 4 . As shown in FIG. 9 , in some embodiments, the fifth audio signal includes an environmental noise signal generated by an external noise source and an external audio playback device. In the case of the played audio signal, the second acoustic performance data includes the amount of speech distortion SDR, the second standard data includes the third evaluation standard value C, and step 24 includes steps 241b and 242b.

Step 241b, compare the voice distortion amount with the third evaluation standard value to obtain a comparison result.

Step 242b: Generate a second test result according to the comparison result between the speech distortion amount and the third evaluation standard value.

Specifically, when the comparison result is that the voice distortion amount SDR is greater than or equal to the third evaluation standard value C, then the test result of the voice device's ability to eliminate ineffective audio in this situation is judged to be qualified, and the second test result includes Information indicating that the test is qualified, otherwise the test result is judged to be unqualified, and the second test result includes information indicating that the test is unqualified.

In the acoustic performance testing method provided by the embodiments of the present disclosure, by evaluating the performance of the underlying audio hardware of the voice device in different test scenarios, test results under different test scenarios are obtained, and the test results are used as the quality of the underlying audio hardware of the voice device. The evaluation standard can provide hardware guarantee for the back-end speech recognition algorithm, and pre-position the quality risk evaluation of the underlying audio hardware, which can save the research and development cost of the product.

FIG. 9 is a block diagram of an acoustic performance testing apparatus for voice equipment provided by an embodiment of the present disclosure. As shown in FIG. 9 , the apparatus is used to implement the above-mentioned acoustic performance testing method. The acoustic performance testing apparatus includes: a first An acquisition module 301 , a second acquisition module 302 , a performance parameter generation module 303 and an evaluation module 304 .

The first acquisition module 301 is configured to acquire a first audio signal recorded by the voice device, where the first audio signal is an audio signal played by the voice device itself or an environmental noise signal generated by an external noise source.

The second obtaining module 302 is configured to obtain a second audio signal generated after the voice device performs signal processing on the first audio signal.

The performance parameter generating module 303 is configured to generate first acoustic performance data of the voice device according to the first audio signal and the second audio signal.

The evaluation module 304 is configured to generate a first test result according to the first acoustic performance data and the preset first standard data.

In some embodiments, the performance parameter generation module 303 is specifically configured to perform fast Fourier transform processing on the first audio signal and the second audio signal, respectively, to obtain the energy value corresponding to the first audio signal and the energy corresponding to the second audio signal value; using the energy value corresponding to the first audio signal and the energy value corresponding to the second audio signal, the noise elimination amount is calculated, and the first acoustic performance data includes the noise elimination amount.

In some embodiments, the first standard data includes a first evaluation standard value, and the evaluation module 304 is specifically configured to compare the noise elimination amount with the first evaluation standard value; according to the comparison result between the noise elimination amount and the first evaluation standard value, A first test result is generated.

In some embodiments, the first acquisition module 301 is further configured to acquire a fourth audio signal generated after the voice device performs signal processing on the recorded third audio signal, where the third audio signal is an audio signal played by an external audio playback device. The second obtaining module 302 is further configured to obtain a sixth audio signal generated after the voice device performs signal processing on the recorded fifth audio signal, where the fifth audio signal includes an audio signal played by the voice device itself and an audio signal played by an external audio playing device , or, the fifth audio signal includes an ambient noise signal generated by an external noise source and an audio signal played by an external audio playback device. The performance parameter generating module 303 is further configured to generate second acoustic performance data of the speech device according to the fourth audio signal and the sixth audio signal. The evaluation module 304 is further configured to generate a second test result according to the second acoustic performance data and the preset second standard data.

In some embodiments, when the fifth audio signal includes an audio signal played by the voice device itself and an audio signal played by an external audio playback device, the performance parameter generation module 303 is specifically configured to: The audio signal is processed by fast Fourier transform to obtain the energy value corresponding to the fourth audio signal and the energy value corresponding to the sixth audio signal; using the energy value corresponding to the fourth audio signal and the energy value corresponding to the sixth audio signal, the calculation is obtained. The amount of energy attenuation, the second acoustic performance data includes the amount of energy attenuation.

In some embodiments, the second standard data includes a second evaluation standard value; the evaluation module 304 is specifically configured to compare the energy attenuation amount with the second evaluation standard value; according to the comparison result between the energy attenuation amount and the second evaluation standard value, A second test result is generated.

In some embodiments, when the fifth audio signal includes an ambient noise signal generated by an external noise source and an audio signal played by an external audio playback device, the performance parameter generation module 303 is specifically configured to: respectively perform a The six audio signals are subjected to fast Fourier transform processing to obtain the energy value corresponding to the fourth audio signal and the energy value corresponding to the sixth audio signal; using the energy value corresponding to the fourth audio signal and the energy value corresponding to the fourth and sixth audio signals, The amount of speech distortion is obtained by calculation, and the second acoustic performance data includes the amount of speech distortion.

In some embodiments, the second standard data includes a third evaluation standard value; the evaluation module 304 is specifically configured to compare the speech distortion amount with the third evaluation standard value; according to the comparison result between the speech distortion amount and the third evaluation standard value, A second test result is generated.

In addition, the acoustic performance testing device provided by the embodiment of the present disclosure is specifically used to implement the aforementioned acoustic performance testing method, and for details, reference may be made to the description of the aforementioned acoustic performance testing method, which will not be repeated here.

FIG. 10 is a block diagram of an electronic device provided by an embodiment of the present disclosure. As shown in FIG. 10, the electronic device includes: one or more processors 501; One or more programs are executed by one or more processors 501, so that one or more processors 501 implement the above-mentioned acoustic performance testing method; one or more I/O interfaces 503 are connected between the processor 501 and the memory 502; is configured to realize the information interaction between the processor 501 and the memory 502 .

Embodiments of the present disclosure also provide a computer-readable storage medium on which a computer program is stored, wherein, when the computer program is executed, the aforementioned acoustic performance testing method is implemented.

Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, functional modules/units in the systems, and devices can be implemented as software, firmware, hardware, and appropriate combinations thereof. In a hardware implementation, the division between functional modules/units mentioned in the above description does not necessarily correspond to the division of physical components; for example, one physical component may have multiple functions, or one function or step may be composed of several physical components Components execute cooperatively. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor or microprocessor, or as hardware, or as an integrated circuit, such as an application specific integrated circuit . Such software may be distributed on computer-readable media, which may include computer storage media (or non-transitory media) and communication media (or transitory media). As known to those of ordinary skill in the art, the term computer storage media includes both volatile and nonvolatile implemented in any method or technology for storage of information, such as computer readable instructions, data structures, program modules or other data flexible, removable and non-removable media. Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cartridges, magnetic tape, magnetic disk storage or other magnetic storage devices, or may Any other medium used to store desired information and which can be accessed by a computer. In addition, communication media typically embodies computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and can include any information delivery media, as is well known to those of ordinary skill in the art .

Example embodiments have been disclosed herein, and although specific terms are employed, they are used and should only be construed in a general descriptive sense and not for purposes of limitation. In some instances, it will be apparent to those skilled in the art that features, characteristics and/or elements described in connection with a particular embodiment may be used alone or in combination with other embodiments, unless expressly stated otherwise. Features and/or elements are used in combination. Accordingly, it will be understood by those of ordinary skill in the art that various changes in form and details may be made therein without departing from the scope of the present disclosure as set forth in the appended claims.

Claims (18)

1. A method for testing acoustic performance of a voice device comprises the following steps:

acquiring a first audio signal recorded by the voice equipment, wherein the first audio signal is an audio signal played by the voice equipment or an environmental noise signal generated by an external noise source;

acquiring a second audio signal generated after the voice equipment performs signal processing on the first audio signal;

generating first acoustic performance data of the speech device from the first audio signal and the second audio signal;

and generating a first test result according to the first acoustic performance data and preset first standard data.

2. The acoustic performance testing method of claim 1, wherein the generating first acoustic performance data of the speech device from the first audio signal and the second audio signal comprises:

respectively carrying out fast Fourier transform processing on the first audio signal and the second audio signal to obtain an energy value corresponding to the first audio signal and an energy value corresponding to the second audio signal;

and calculating to obtain a noise elimination amount by using an energy value corresponding to the first audio signal and an energy value corresponding to the second audio signal, wherein the first acoustic performance data comprises the noise elimination amount.

3. The acoustic performance testing method according to claim 2, wherein the first standard data includes a first criterion value, and the generating a first test result according to the first acoustic performance data and a preset first standard data includes:

comparing the noise removal amount with the first evaluation criterion value;

and generating the first test result according to the comparison result of the noise elimination amount and the first judgment standard value.

4. The acoustic performance testing method of claim 1, wherein the method further comprises:

acquiring a fourth audio signal generated after the voice device performs signal processing on a recorded third audio signal, wherein the third audio signal is an audio signal played by an external audio playing device;

acquiring a sixth audio signal generated after the audio device performs signal processing on a recorded fifth audio signal, where the fifth audio signal includes an audio signal played by the audio device itself and an audio signal played by the external audio playing device, or the fifth audio signal includes an environmental noise signal generated by an external noise source and an audio signal played by the external audio playing device;

generating second acoustic performance data of the speech device from the fourth audio signal and the sixth audio signal;

and generating a second test result according to the second acoustic performance data and preset second standard data.

5. The acoustic performance testing method of claim 4, wherein in case that the fifth audio signal comprises an audio signal played by the voice device itself and an audio signal played by the external audio playing device,

generating second acoustic performance data of the speech device from the fourth audio signal and the sixth audio signal, comprising:

performing fast fourier transform processing on the fourth audio signal and the sixth audio signal respectively to obtain an energy value corresponding to the fourth audio signal and an energy value corresponding to the sixth audio signal;

and calculating to obtain an energy attenuation amount by using an energy value corresponding to the fourth audio signal and an energy value corresponding to the sixth audio signal, wherein the second acoustic performance data comprises the energy attenuation amount.

6. The acoustic performance testing method according to claim 5, wherein the second standard data includes a second criterion value, and the generating a second test result according to the second acoustic performance data and a preset second standard data includes:

comparing the energy attenuation amount with the second judgment standard value;

and generating the second test result according to the comparison result of the energy attenuation and the second judgment standard value.

7. The acoustic performance testing method of claim 4, wherein in a case where the fifth audio signal comprises an ambient noise signal generated by an external noise source and an audio signal played by the external audio playing device,

generating second acoustic performance data of the speech device from the fourth audio signal and the sixth audio signal, comprising:

performing fast fourier transform processing on the fourth audio signal and the sixth audio signal respectively to obtain an energy value corresponding to the fourth audio signal and an energy value corresponding to the sixth audio signal;

and calculating to obtain the voice distortion quantity by using the energy value corresponding to the fourth audio signal and the energy value corresponding to the sixth audio signal, wherein the second acoustic performance data comprises the voice distortion quantity.

8. The acoustic performance testing method according to claim 7, wherein the second standard data includes a third criterion value, and the generating a second test result according to the second acoustic performance data and a preset second standard data includes:

comparing the voice distortion amount with the third criterion value;

and generating the second test result according to the comparison result of the voice distortion quantity and the third judgment standard value.

9. An acoustic performance testing apparatus for a speech device, comprising:

the first acquisition module is used for acquiring a first audio signal recorded by the voice equipment, wherein the first audio signal is an audio signal played by the voice equipment or an environmental noise signal generated by an external noise source;

the second obtaining module is used for obtaining a second audio signal generated after the voice equipment performs signal processing on the first audio signal;

a performance parameter generation module, configured to generate first acoustic performance data of the speech device according to the first audio signal and the second audio signal;

and the evaluation module is used for generating a first test result according to the first acoustic performance data and preset first standard data.

10. The acoustic performance testing apparatus according to claim 9, wherein the performance parameter generating module is specifically configured to perform fast fourier transform processing on the first audio signal and the second audio signal, respectively, to obtain an energy value corresponding to the first audio signal and an energy value corresponding to the second audio signal; and calculating to obtain a noise elimination amount by using an energy value corresponding to the first audio signal and an energy value corresponding to the second audio signal, wherein the first acoustic performance data comprises the noise elimination amount.

11. The acoustic performance testing apparatus of claim 10, wherein the first criterion data comprises a first criterion value, the evaluation module being configured to compare the amount of noise cancellation to the first criterion value; and generating the first test result according to the comparison result of the noise elimination amount and the first judgment standard value.

12. The acoustic performance testing apparatus of claim 9,

the first obtaining module is further configured to obtain a fourth audio signal generated by the voice device after performing signal processing on a recorded third audio signal, where the third audio signal is an audio signal played by an external audio playing device;

the second obtaining module is further configured to obtain a sixth audio signal generated by the voice device after performing signal processing on a recorded fifth audio signal, where the fifth audio signal includes an audio signal played by the voice device itself and an audio signal played by the external audio playing device, or the fifth audio signal includes an environmental noise signal generated by an external noise source and an audio signal played by the external audio playing device;

the performance parameter generating module is further configured to generate second acoustic performance data of the speech device according to the fourth audio signal and the sixth audio signal;

the evaluation module is further used for generating a second test result according to the second acoustic performance data and preset second standard data.

13. The acoustic performance testing apparatus of claim 12, wherein in case that the fifth audio signal comprises an audio signal played by the voice device itself and an audio signal played by the external audio playing device,

the performance parameter generating module is specifically configured to perform fast fourier transform processing on the fourth audio signal and the sixth audio signal respectively to obtain an energy value corresponding to the fourth audio signal and an energy value corresponding to the sixth audio signal; and calculating to obtain an energy attenuation amount by using an energy value corresponding to the fourth audio signal and an energy value corresponding to the sixth audio signal, wherein the second acoustic performance data comprises the energy attenuation amount.

14. The acoustic performance testing apparatus of claim 13, wherein the second criterion data comprises a second criterion value;

the evaluation module is specifically used for comparing the energy attenuation with the second evaluation standard value; and generating the second test result according to the comparison result of the energy attenuation and the second judgment standard value.

15. The acoustic performance testing apparatus of claim 12, wherein in case the fifth audio signal comprises an ambient noise signal generated by an external noise source and an audio signal played by the external audio playing device,

the performance parameter generating module is specifically configured to perform fast fourier transform processing on the fourth audio signal and the sixth audio signal respectively to obtain an energy value corresponding to the fourth audio signal and an energy value corresponding to the sixth audio signal; and calculating to obtain the voice distortion quantity by using the energy value corresponding to the fourth audio signal and the energy value corresponding to the sixth audio signal, wherein the second acoustic performance data comprises the voice distortion quantity.

16. The acoustic performance testing apparatus of claim 15, wherein the second criterion data comprises a third criterion value;

the evaluating module is specifically used for comparing the voice distortion quantity with the third evaluating standard value; and generating the second test result according to the comparison result of the voice distortion quantity and the third judgment standard value.

17. An electronic device, comprising:

one or more processors;

memory having one or more programs stored thereon that, when executed by the one or more processors, cause the one or more processors to implement the acoustic performance testing method of any of claims 1-8;

one or more I/O interfaces connected between the processor and the memory and configured to enable information interaction between the processor and the memory.

18. A computer-readable medium, on which a computer program is stored, wherein the computer program, when executed, implements the acoustic performance testing method of any of claims 1-8.

Images