Loudspeaker stroke prediction system
By using low-pass filtering circuits and downsampling circuits in the speaker stroke prediction system, the sampling frequency of the audio signal and the order of the filter circuit are reduced, and the problem of deterioration of stroke prediction accuracy in the prior art is solved, thereby achieving high-accurate stroke prediction and saving computing resources.
Patent Information
- Application Number
- CN202311163913.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-09-11
- Publication Date
- 2025-03-11
AI Technical Summary
When existing speaker stroke prediction circuits reduce the impulse response order, they will cause significantly worse stroke prediction accuracy, making it difficult to balance between calculation and storage requirements and accuracy.
A low-pass filtering circuit and a downsampling circuit are used to generate a loudspeaker stroke prediction value with high accuracy by reducing the sampling frequency of the audio signal and the order of the filter circuit.
While reducing the filter circuit order of the impulse response generation unit, it is still possible to maintain a stroke prediction effect with high accuracy, saving memory space and DSP calculation amount.
Smart Images
Figure CN119603607A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a loudspeaker stroke estimation circuit, in particular to a loudspeaker stroke prediction system. Background Art
[0002] A speaker is generally composed of a magnet, a coil, and a diaphragm. When an electric current passes through the coil to generate an electromagnetic field, the coil and the diaphragm vibrate together, pushing the surrounding air to vibrate, and the speaker generates sound. The energy conversion process of the speaker is from electrical energy to magnetic energy, then from magnetic energy to mechanical energy, and then from mechanical energy to sound.
[0003] The excursion of a loudspeaker refers to the displacement of the diaphragm from a static position under the action of a driving force until it reaches the maximum displacement (hereinafter referred to as the stroke). The size of the stroke affects the performance of the loudspeaker, such as volume, power, and distortion, and is therefore crucial to the design and optimization of the loudspeaker. Because when the loudspeaker exceeds the maximum excursion limit, temporary or permanent mechanical damage may be caused to the loudspeaker, resulting in abnormal sound or failure to generate sound. Mechanical damage may be caused by collision between the coil, diaphragm, or voice coil frame and fixed parts. In order to avoid mechanical damage to the loudspeaker, the stroke of the diaphragm needs to be accurately estimated.
[0004] In terms of speaker stroke prediction, a common approach is based on acoustic modeling and simulation technology. By building a physical model of the speaker and the drive system, the vibration and acoustic characteristics of the speaker can be predicted and analyzed. In this case, the speaker is characterized by several basic parameters (for example, parameters include load characteristics, diaphragm elasticity and damping, etc.), and can also be estimated by measuring the speaker output current and voltage. Based on the aforementioned basic parameters of the speaker, a unit impulse response (Impulse Response) filter is established to process the output response of the speaker.
[0005] The unit impulse response of a speaker refers to the transient response of its output when the speaker is subjected to an impulse signal. It is used to evaluate the speaker's performance indicators such as frequency response and harmonic distortion. By performing Fourier series analysis on the frequency response of the pre-modeled impulse response of the speaker through the input audio signal, the speaker's resonant frequency and bandwidth can be obtained, and then the predicted stroke response under the audio input can be deduced. The speaker's stroke value is estimated through the infinite impulse response (IIR) or finite impulse response (FIR) stroke prediction circuit corresponding to the speaker, and the audio signal output to the speaker is limited according to the stroke value, thereby avoiding the problem of the speaker exceeding the maximum stroke limit (Xmax). Summary of the invention
[0006] However, in the above-mentioned stroke prediction circuit, factors such as frequency response and system stability usually need to be considered. In practical applications, the order of the impulse response is usually related to the required system accuracy. The stroke prediction circuit uses a higher-order impulse response to provide higher prediction accuracy, but it also leads to increased calculation and storage. If the filter is adjusted to a low order, it will be found that the accuracy of speaker stroke prediction is significantly deteriorated. For applications that estimate speaker stroke, it is necessary to balance the calculation and storage requirements with accuracy to determine the optimal impulse response order.
[0007] In the previous case, the 1024-order impulse response generated by the speaker under the same audio input can accurately simulate the predicted stroke response output. When the impulse response order is reduced to 128, the simulated predicted stroke response output will cause a large error offset. This phenomenon is a prediction error caused by insufficient resolution. The audio sampling rate refers to the number of times the sound signal is sampled over a period of time. It is usually expressed in units of samples per second, in Hertz (Hz). Common audio sampling rates include 44.1kHz, 48kHz, 96kHz, etc. The higher the sampling rate, the higher the sampling accuracy of the sound signal. Please refer to Figure 1A , which is a schematic diagram of the stroke response distribution of a commercially available full-range speaker. It can be found that most of the sound contribution is below the 3kHz bandwidth. Please refer to Figure 1B , is a schematic diagram of the stroke response distribution of another commercially available full-range speaker. It can also be found that most of the sound contribution is below the frequency bandwidth of less than 3kHz.
[0008] In view of the above-mentioned prior art, the present invention proposes a speaker stroke prediction system, which, in conjunction with the proposed system architecture, audio sampling rate and relative relationship between speaker stroke response distribution, can still maintain a high accuracy effect while reducing the impulse response order.
[0009] To achieve the above and other purposes, the present invention provides a speaker stroke prediction system, comprising: a low-pass filter circuit, a downsampling circuit and an impulse response generation unit. The low-pass filter circuit is configured to generate an impulse response according to a first sampling frequency F S1 The audio signal X(t) is generated and then outputted through the low-pass filter circuit. LPF (t). The down-sampling circuit is coupled to the low-pass filter circuit and is configured to convert the audio signal X output after passing through the low-pass filter circuit into LPF (t) The first sampling frequency F S1 Downsample to the second sampling frequency F S2 , to generate the downsampled audio signal X LPFDN (t). The impulse response generating unit is coupled to the down-sampling circuit and is configured to generate an impulse response according to the speaker stroke transfer function and the down-sampled audio signal X.LPFDN (t), generate the stroke prediction value Y(t).
[0010] In some embodiments, the first sampling frequency F S1 is 48kHz, the second sampling frequency F S2 To be adjustable.
[0011] In some embodiments, the second sampling frequency F S2 is the first sampling frequency F S1 1 / 8 times of.
[0012] In some embodiments, the second sampling frequency F S2 The component that covers 95% of the stroke response within the 0.5 times frequency point.
[0013] In some embodiments, the impulse response generation unit is an impulse response filter.
[0014] In some embodiments, when the second sampling frequency F S2 When the frequency is 6kHz, the order of the designed impulse response filter is 128.
[0015] In some embodiments, the cut-off frequency of the low-pass filter circuit is less than the second sampling frequency F S2 half.
[0016] In some embodiments, the cut-off frequency of the low-pass filter circuit is greater than the second sampling frequency F S2 The attenuation at 0.5 times the frequency point is less than -60dB.
[0017] In some embodiments, a stroke conversion circuit is further included, coupled to the impulse response generation unit and configured to generate the speaker stroke transfer function.
[0018] In some embodiments, the system further includes: a stroke conversion circuit coupled to the impulse response generation unit and configured to generate a speaker stroke transfer function; a protection circuit coupled to the impulse response generation unit and configured to generate a speaker stroke protection value according to the stroke prediction value; a gain controller coupled to the protection circuit and configured to generate an audio signal with a maximum stroke limit according to the delayed audio signal and the speaker stroke protection value; and a delay circuit coupled to the gain controller and configured to generate a delayed audio signal according to the audio signal.
[0019] Thus, the speaker stroke prediction system of the present invention improves the accuracy of the impulse response generation unit at low frequencies through a downsampling circuit and can effectively reduce the order of the filter circuit of the impulse response generation unit, thereby using less memory storage space and lower DSP calculation amount. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] Figure 1A It is a schematic diagram of the stroke response distribution of a commercially available full-range loudspeaker.
[0021] Figure 1B This is another diagram of the stroke response distribution of a commercially available full-range loudspeaker.
[0022] Figure 2 is a system block diagram of an embodiment of the present invention.
[0023] Figure 3 FIG. 4 is a waveform diagram of an input audio signal X(t) according to an embodiment of the present invention.
[0024] Figure 4 is the first sample F of the embodiment of the present invention S1 Frequency and second sampling F S2 Schematic diagram of frequency.
[0025] Figure 5 It is a schematic diagram of an impulse response of an embodiment of the present invention using a finite impulse response filter with an order of 128 for calculation.
[0026] Fig. 6A It is a stroke prediction output simulation waveform diagram of an embodiment of the present invention.
[0027] Figure 6B It is a waveform diagram of actual stroke output measurement of an embodiment of the present invention.
[0028] Figure 7 is a system block diagram of another embodiment of the present invention.
[0029] Figure 8 It is a system block diagram of another embodiment of the present invention.
[0030] The following are the descriptions of the reference numerals:
[0031] 101,201,301 Low-pass filter circuit
[0032] 102,202,302 Downsampling Circuit
[0033] 103,203,303 Impulse response generation unit
[0034] 204,304 stroke conversion circuit
[0035] 305 Protection Circuit
[0036] 306 Delay Circuit
[0037] 307 Gain Controller
[0038] F S1 First sampling frequency
[0039] FS2 Second sampling frequency DETAILED DESCRIPTION
[0040] In order to fully understand the purpose, features and effects of the present invention, the present invention is now described in detail through the following specific embodiments and in conjunction with the accompanying drawings, as follows:
[0041] Please refer to Figure 2 , is a system block diagram of an embodiment of the present invention. Figure 2 As shown, the speaker stroke prediction system includes: a low-pass filter circuit 101 , a down-sampling circuit 201 and an impulse response generation unit 103 .
[0042] The low-pass filter circuit 101 is configured to filter the signal according to a first sampling frequency F S1 The audio signal X(t) is generated by the low-pass filter circuit 101 and then outputted. LPF (t). Please refer to Figure 3 , is a waveform diagram of an input audio signal X(t) according to an embodiment of the present invention, wherein the first sampling frequency F of the input audio signal X(t) S1 =48kHz. Experiments have shown that the high-frequency audio signal X(t) is less helpful or even useless for stroke prediction. Therefore, the low-frequency band of the audio signal X(t) is extracted through the low-pass filter circuit 101 to generate the audio signal X output after passing through the low-pass filter circuit 101. LPF (t).
[0043] The down-sampling circuit 102 is coupled to the low-pass filter circuit 101. The down-sampling circuit 102 is configured to convert the audio signal X outputted after passing through the low-pass filter circuit 101 into LPF The first sample F of (t) S1 The frequency is downsampled to the second sampling frequency F S2 , to generate the downsampled audio signal X LPFDN (t). Please refer to Figure 4 , is the first sample F of the embodiment of the present invention S1 Frequency and second sampling F S2 It can be found that the audio signal X outputted after passing through the low-pass filter circuit 101 and the down-sampling circuit 102 is LPF (t) The second sampling frequency F S2 In addition, the downsampling circuit 102 is helpful for the order design of the filter circuit of the impulse response generation unit 103. More specifically, the downsampling circuit 101 allows the impulse response generation unit 103 to use a lower order filter. In some embodiments, the cutoff frequency of the low-pass filter circuit 101 is less than the second sampling frequency F S2 Preferably, the second sampling frequency FS2 Half of this is less than -60dB.
[0044] In some embodiments, the second sampling frequency F S2 The second sampling frequency F is adjustable. S2 is the first sampling frequency F S1 1 / N times, N is a positive integer. For example, the second sampling frequency F S2 is the first sampling frequency F S 1 / 8 times of 1. For example, when the first sampling frequency is 48 kHz, the second sampling frequency is 6 kHz. When the first sampling frequency is 96 kHz, the second sampling frequency is 12 kHz or lower. In some embodiments, the second sampling frequency F S2 The 0.5 times frequency point covers 95% of the components of the stroke response.
[0045] The impulse response generating unit 103 is coupled to the down-sampling circuit 102. The impulse response generating unit 103 is configured to generate an impulse response signal according to the speaker stroke transfer function and the down-sampled audio signal X. LPFDN More specifically, the impulse response generation unit 103 combines the speaker stroke transfer function H(t) with the downsampled audio signal X LPFDN (t) performs a time domain convolution operation to generate a stroke prediction value Y(t) corresponding to the displacement distance. Figure 5 , which is a schematic diagram of an impulse response of an embodiment of the present invention using a finite impulse response filter of order 128 for operation. The impulse response generating unit 103 may be an impulse response filter, such as a finite impulse response (FIR) filter or an infinite impulse response (IIR) filter. In some embodiments, when the second sampling frequency F S2 When the second sampling frequency F is 6kHz, the order of the finite impulse response filter is 128. S2 When it is 6kHz, it can be applied to speakers with a frequency response range of 3kHz.
[0046] Please refer to Fig. 6A and Figure 6B ,like Fig. 6A and Figure 6B As shown in the figure, the vertical axis is the stroke value, the unit is millimeter (mm), and the horizontal axis is the time, the unit is second (s). Fig. 6A A simulated waveform of the excursion prediction output generated for a loudspeaker excursion prediction system. Figure 6BThe actual output waveform of the speaker stroke measured by the laser rangefinder is shown in FIG. 1. It can be found that the two waveforms are very similar, that is, the speaker stroke prediction result of the embodiment of the present invention has high accuracy.
[0047] Please refer to Figure 7 , is a system block diagram of another embodiment of the present invention. Figure 3 As shown, the speaker stroke prediction system includes: a low-pass filter circuit 201, a down-sampling circuit 202, an impulse response generation unit 203 and a stroke conversion circuit 204. The low-pass filter circuit 201, the down-sampling circuit 202 and the impulse response generation unit 203 disclosed in this embodiment are respectively the same as Figure 2 The disclosed low-pass filter circuit 101, down-sampling circuit 102 and impulse response generation unit 103 are not described in detail below.
[0048] The stroke conversion circuit 204 is coupled to the impulse response generation unit 203. The stroke conversion circuit 204 is configured to generate a speaker stroke transfer function H(s) according to the speaker parameters. Since the speaker parameters belong to frequency domain signals, after the reverse Laplace operation, the speaker stroke transfer function H(t) of the time domain signal can be generated to be used in conjunction with the downsampled audio signal X which is also a time domain signal. LPFDN (t) Perform relevant operations.
[0049] Please refer to Figure 8 , is a system block diagram of another embodiment of the present invention. Figure 8 As shown, the speaker stroke prediction system includes: a low-pass filter circuit 301, a down-sampling circuit 302, an impulse response generation unit 303, a stroke conversion circuit 304, a protection circuit 305, a delay circuit 306 and a gain controller 307. The low-pass filter circuit 301, the down-sampling circuit 302, the impulse response generation unit 303 and the stroke conversion circuit 304 disclosed in this embodiment are respectively the same as Figure 7 The disclosed low-pass filter circuit 201, down-sampling circuit 202, impulse response generation unit 203 and stroke conversion circuit 204 are not described in detail below.
[0050] The protection circuit 305 is coupled to the impulse response generation unit 303. The protection circuit 305 is configured to generate a speaker stroke protection value according to the stroke prediction value Y(t) and send it to the gain controller 307. The delay circuit 306 is configured to generate a delayed audio signal according to the audio signal X(t) and send it to the gain controller 307.
[0051] The gain controller 307 is coupled to the protection circuit 305 and the delay circuit 306. The gain controller 307 is configured to generate a maximum excursion-limited audio signal Y(t) according to the delayed audio signal and the speaker excursion protection value to prevent the speaker from exceeding the maximum excursion limit (Xmax).
[0052] Next, the embodiment of the present invention compares the difference between the prior art and the embodiment of the present invention through experimental results, please refer to Table 1, as follows:
[0053] Table 1
[0054]
[0055]
[0056] As shown in Table 1 above, the stroke prediction circuit of the prior art uses a filter circuit with an order of 1024, while the embodiment of the present invention only needs to use a filter circuit with an order of 128. In comparison, the present invention can save about 1 / 8 of the memory storage space.
[0057] The prior art uses a sampling frequency of 48kHz and a filter circuit with an order of 1024. In terms of DSP computing power, a 48Mega pulse cycle is required to perform calculations to obtain the stroke prediction value. The embodiment of the present invention uses a sampling frequency of 6kHz, a filter circuit with an order of 128, and a low-pass filter circuit. According to estimates, the low-pass filter circuit will roughly take about 1440K (10+1440K) of DSP pulse cycles to perform calculations. In terms of DSP computing power, only 2208K pulse cycles are required to perform calculations to obtain the stroke prediction value Y(t). Overall, the embodiment of the present invention can save about 1 / 24 times the DSP computing power. In other words, while achieving the same high accuracy (>98%) of stroke prediction values, the embodiment of the present invention has less memory usage space and lower DSP computing power.
[0058] In summary, the speaker stroke prediction system of the present invention improves the accuracy of the impulse response generation unit at low frequencies through a downsampling circuit and can effectively reduce the order of the filter circuit of the impulse response generation unit, thereby using less memory storage space and lower DSP calculation amount.
[0059] The present invention has been disclosed above with preferred embodiments, but those skilled in the art should understand that the embodiments are only used to describe the present invention and should not be interpreted as limiting the scope of the present invention. It should be noted that all changes and substitutions equivalent to the embodiments should be considered to be within the scope of the present invention. Therefore, the scope of protection of the present invention shall be subject to the content defined in the claims.
Claims
1. A speaker stroke prediction system, characterized in that: include: A low-pass filter circuit is configured to, according to an audio signal X(t) having a first sampling frequency, Generate an audio signal X through the low-pass filter circuit LPF (t); A down sampling circuit is coupled to the low pass filter circuit and is configured to downsample the audio signal X output after passing through the low pass filter circuit. LPF (t) downsamples the first sampling frequency to a second sampling frequency to generate a downsampled audio signal X LPFDN (t); and The impulse response generating unit is coupled to the down-sampling circuit and is configured to generate an impulse response according to the speaker stroke transfer function and the down-sampled audio signal X. LPFDN (t), generate the stroke prediction value Y(t).
2. The speaker stroke prediction system according to claim 1, characterized in that The first sampling frequency is 48 kHz, and the second sampling frequency is adjustable.
3. The speaker stroke prediction system according to claim 1, characterized in that The second sampling frequency is 1 / 8 times the first sampling frequency.
4. The speaker stroke prediction system according to claim 1, characterized in that The second sampling frequency is 0.5 times the frequency point covering 95% of the components of the stroke response.
5. The speaker stroke prediction system according to claim 1, characterized in that: The impulse response generating unit is an impulse response filter.
6. The speaker stroke prediction system according to claim 5, characterized in that: When the second sampling frequency is 6 kHz, the order of the impulse response filter is designed to be 128.
7. The speaker stroke prediction system according to claim 1, characterized in that: The cut-off frequency of the low-pass filter circuit is less than half of the second sampling frequency.
8. The speaker stroke prediction system according to claim 1, characterized in that: Half of the second sampling frequency is less than -60dB.
9. The speaker stroke prediction system according to claim 1, characterized in that: The invention also includes a stroke conversion circuit, which is coupled to the impulse response generation unit and configured to generate the speaker stroke transfer function.
10. The speaker stroke prediction system according to claim 9, characterized in that: Also included are: A protection circuit, coupled to the impulse response generation unit, configured to generate a speaker stroke protection value according to the stroke prediction value Y(t); A gain controller, coupled to the protection circuit, configured to generate an audio signal with a maximum stroke limit according to the delayed audio signal and the speaker stroke protection value; and The delay circuit is coupled to the gain controller and configured to generate the delayed audio signal according to the audio signal.