Low-Complexity Howling Suppression for Portable Karaoke

Abstract

Provided is a low-complexity howling suppression system and method for a portable Karaoke system. When performing howling suppression, at least one infinite impulse response (IIR) filter is introduced to estimate the acoustic feedback picked up by a microphone from a real environment, and thereby eliminate the acoustic feedback from the microphone input signal.

Description

Technical Field

[0001] The present invention generally relates to acoustic feedback cancellation technology. More specifically, the present invention relates to a low-complexity howling suppression method and system for a portable karaoke system. Background Art

[0002] Karaoke, as an interactive entertainment activity, has become increasingly popular in East Asia. Anyone can use a karaoke machine in a karaoke club to sing with their friends. Recently, people are not only satisfied with singing in clubs, but also use portable all-in-one karaoke machines to sing anywhere. Therefore, many portable karaoke products have entered the market. However, the sound quality of most of these products is not as good as expected. Many of these products suffer from "acoustic howling" or do not have sufficient sound level due to this howling problem.

[0003] In fact, howling is a common problem in karaoke or other sound reinforcement systems. A karaoke system typically includes at least one microphone and a speaker. When a sound signal is picked up by the microphone and then amplified and fed to the speaker, the speaker sound is usually picked up by the microphone through a direct sound path or some reflection paths. The acoustic coupling between the speaker and the microphone forms a closed signal loop, in which case acoustic feedback in the form of unwanted howling may occur. In such a situation, this can lead to a howling problem and even damage the speaker or limit the sound reinforcement performance, because the applicable amplification amount is limited if the karaoke system is required to be stable.

[0004] Therefore, different measures have been taken to prevent this problem. There are mainly four types of howling suppression methods. The first method is the frequency shift method, which can change the microphone input signal in each loop by a frequency shift of several Hz. It will be more helpful when the shift frequency is large (e.g., 20 Hz), but the side effect on the sound quality is too serious and thus considered unacceptable. Therefore, usually not much change can be made and a compromise has to be made in the suppression performance. The second method is notch filter-based feedback suppression (NFS), where the notch filter is used to suppress the problematic frequencies where howling has been detected. A notch filter is a stopband filter with a very narrow stopband, which can significantly reduce the gain of a specific frequency band and thus suppress these frequencies in the microphone input signal. However, the NFS method usually includes a detection stage and a suppression stage, which means that the howling sound needs to be detected first, so the howling sound is often heard before suppressing it. The third method is the beamforming method, which uses a microphone array or a speaker array to modify the directivity to reduce the direct sound propagation between the microphone and the speaker. But this method requires additional hardware and more calculations, and thus the number of microphones and speakers is often limited. The fourth method is the echo cancellation (AEC) method, which uses an adaptive filter to approximate the transfer function between the microphone and the speaker, and filters the output signal from the speaker to eliminate the approximate feedback signal picked up by the microphone. If the adaptive filter perfectly approximates the transfer function, no howling will occur. Usually, the AEC method can perform well due to the adaptive mechanism, but it consumes high computing power and power consumption as well as a long system processing time, which is not suitable for a portable karaoke system with specific playtime and delay requirements.

[0005] However, most of these measures focus on the usual large karaoke systems, where the distance between the speaker and the microphone is usually much larger than that in a portable all-in-one karaoke system. On the other hand, even with the small karaoke machine, users may still want to sing as loudly as possible. Therefore, if we want the karaoke product to exceed the user's expectations, better technologies need to be developed to suppress howling. Summary of the Invention

[0006] The present invention overcomes some of the above-mentioned drawbacks by providing a low-complexity howling suppression system for a portable karaoke system. The howling suppression system includes: at least one microphone for capturing an input signal, the input signal including a source signal and acoustic feedback propagating in the environment; an electroacoustic path for compressing and equalizing the input signal and then feeding it to an output signal after amplification; and a speaker for playing back the output signal. The howling suppression system further includes at least one infinite impulse response (IIR) filter for estimating the acoustic feedback and thereby eliminating the acoustic feedback from the input signal.

[0007] The present invention also provides a low-complexity howling suppression method for a portable karaoke system. The howling suppression method includes the steps of: capturing an input signal via at least one microphone, the input signal including a source signal and acoustic feedback; compressing and equalizing the input signal in an electroacoustic path and then feeding the output signal to a speaker after amplification; and playing back the output signal using the speaker, and the output signal propagates in the environment. The howling suppression method further includes the steps of: estimating the acoustic feedback via at least one infinite impulse response (IIR) filter and thereby eliminating the acoustic feedback from the input signal. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] The present invention can be better understood by reading the following description of non-limiting embodiments with reference to the accompanying drawings. In the figures, like reference numerals denote corresponding parts, and hereinafter:

[0009] Figure 1 An exemplary howling situation in a karaoke system having one microphone and one speaker is shown.

[0010] Figures 2A to 2B An exemplary system diagram for howling suppression in a karaoke system according to the present invention is shown.

[0011] Figure 3 An exemplary flowchart showing a howling suppression method according to the present invention is shown.

[0012] Figure 4 An exemplary hardware system of a portable integrated karaoke system according to the present invention is shown.

[0013] Figures 5A to 5E Some simulated comparison results related to the present invention are shown, wherein Figure 5A the sound spectrum of the speaker signal is shown; Figure 5B the sound spectrum of the speaker signal in which howling sounds are present at approximately 2 kHz and 9.3 kHz is shown; Figure 5C the sound spectrum obtained using a frequency shift method is shown;Figure 5D shows the sound spectrum obtained by the NFS method, and Figure 5E shows the sound spectrum obtained by the present invention. Detailed Description of the Invention

[0014] A detailed description of one or more embodiments of the present invention is disclosed below; however, it should be understood that the disclosed embodiments are only examples of the present invention, and the present invention may be embodied in various alternative forms. The figures are not necessarily to scale; some features may be enlarged or minimized to show details of particular components. Therefore, the specific structural and functional details disclosed herein should not be construed as limiting, but only as a representative basis for teaching those skilled in the art to adopt the present invention in various ways.

[0015] Figure 1 shows an exemplary howling situation in a karaoke system having a microphone and a speaker. In Figure 1 it can be seen that the system 100 includes a microphone 110 and a speaker 140, and the microphone 110 is connected to an electroacoustic path 120 and then connected to the speaker after amplification 130. When setting up the karaoke system, the microphone 110 captures the source signal S(z) from the user's singing voice; and processes the input signal X(z) in the electroacoustic path G(z), the processing including compression and equalization; and then amplifies it by a gain factor K. The output signal Y(z) is fed to the speaker 140 and propagates in the environment 150.

[0016] When the sound signal is picked up by the microphone 110 and then amplified and fed to the speaker 140, the audio sound played back by the speaker 140 may be picked up by the microphone 110 through a direct path or a reflection path. The acoustic coupling between the speaker 140 and the microphone 110 causes the output signal Y(z) propagating through the environmental transfer function F(z) to form acoustic feedback, and then is also picked up by the microphone 110, and thus Figure 1 the input signal X(z) shown in includes the source signal S(z) and the acoustic feedback Y(z)·F(z).

[0017] Therefore, representing all signals and transfer functions in the frequency domain, Figure 1 the process in can be expressed by the following equations:

[0018] Y(z) = X(z)·G,(z)·K (I)

[0019] X(z) = y(z).F(z) + S(z) (2)

[0020] Regarding the above equation (2), it can be envisaged that the input signal X(z) to the karaoke system may further include, for example, an audio stream (such as the accompaniment music of a song) input via a Bluetooth or AUX interface, which has been omitted herein.

[0021] And then the overall transfer function H(z) from the source signal S(z) to the output signal Y(z) can be calculated:

[0022]

[0023] Where the term F(z)·G(z)·K refers to the loop response of the system, and its magnitude and phase responses represent the loop gain and loop phase, respectively. Therefore, howling will occur when the system becomes unstable, which is summarized in the Nyquist stability criterion.

[0024] F(z)·G(z)·K≥1, ∠F(z)G(z) = n2π (4)

[0025] In the past few decades, many howling suppression methods have been discussed. Undoubtedly, the first preventive measure considered is to optimize the entire karaoke system, such as the directivity of the speakers and microphones, the distance between the speakers and the microphones, the overall gain of the system, and the amplitudes of some problematic frequencies. However, such optimization is usually limited, especially in portable integrated karaoke systems, because the form factor and sound performance of portable integrated karaoke systems usually have certain requirements - as small as possible in size, but as high as possible in sound level. For these restrictive cases, the process must be automated or other measures need to be taken to avoid howling feedback. Therefore, the present invention provides a portable integrated karaoke system using a low-complexity howling suppression method.

[0026] In order to better suppress the howling sound in a portable integrated karaoke machine and reduce processing calculations, power consumption, and system latency, the present invention provides a low-complexity howling suppression method for a portable karaoke system, in which a filter is used to eliminate unwanted components from the microphone signal.

[0027] Figure 2A An exemplary system diagram for howling suppression in a karaoke system according to the present invention is shown. A filter 260 is introduced into this system 200, and F est (z) is the estimated transfer function established by the filter 260, which is designed and adjusted to be similar to the actual environmental transfer function F(z) in the environment 250. Many algorithms have been proposed to implement this method, such as the least mean square and normalized least mean square algorithms. In Figure 2A the example of est (z) converges perfectly and thus F est(z) = F(z), X est (z) = X(z), then all feedback signals from the speaker 240 will be eliminated. Therefore, the input signal only consists of the source signal S(z) such as the user's singing voice, and howling will not occur, which can be expressed by the following equation:

[0028] X(z) - X est (z) = S(z) + F(z)Y(z) - F est (z)Y(z) ≈ S(z) (5)

[0029] Using an adaptive filter to estimate the environmental transfer function can achieve good howling suppression effect. However, in practice, there are still some problems in howling suppression by estimating the environmental transfer function through an adaptive filter. First, the delay cannot meet the requirements of this small karaoke machine. Since the adaptive algorithm may require a long processing time, and the speaker is very close to the microphone, the sound propagation time may be less than the processing time, and thus the algorithm is invalid. Second, the adaptive algorithm may also consume high power and the battery will be quickly depleted, which is an obvious defect of portable devices. Third, the adaptive algorithm sometimes does not converge smoothly, which leads to obvious differences in the adaptive filter, and this will always affect the sound quality of the user's singing. In addition, there is a high correlation between the speaker and the microphone in the karaoke system, which makes this structure perform poorly in this case.

[0030] Therefore, in an exemplary system, several second-order infinite impulse response (IIR) filters 260' are used to model the transfer function F(z), as Figure 2B shown. Since the form factor of the portable integrated karaoke system determines the fixed distance between the speaker 240 and the microphone 210, the transfer function F iir (z) can be measured and estimated offline and approximated by a multi-band IIR filter. In addition, in some situations, such as an integrated karaoke machine, the relative positions of the speaker and the microphone are relatively fixed. Therefore, even without an adaptive process, a howling suppression effect of 80 - 90% can still be achieved, and the IIR filter can have a fixed suppression effect. Using an IIR filter in this system not only saves power consumption but also saves chip computing resources.

[0031] In addition, decorrelation 215 can also be introduced into the exemplary system 200 to reduce the correlation between the speaker and the microphone signals. In the Figure 2B model shown, the speaker signal and the microphone signal are decorrelated by shifting the input signal in frequency before compression and equalization, which means that in decorrelation 215, the total input signal X(z) - X iir (z) is frequency-shifted, and Y(z) is made decorrelated from X(z) - X iir(z) decorrelates. Thus, the output signal is decorrelated from the input signal. For example, a frequency-shifted output signal x 移位 (t) can be obtained in the time domain as follows,

[0032]

[0033] where Δf is the shift frequency, and is the Hilbert transform form of the original signal x(t).

[0034] All the advantages of frequency shifting can be incorporated into the method of the present invention. In the proposed model for modeling the environmental transfer function through several IIR filters, adaptive processing is not required due to delay problems and high power consumption. In addition, due to the acoustic feedback problem caused by a part of the output signal of the speaker returning to the input microphone through acoustic coupling in the air, frequency shifting is used to decorrelate the reference signal from the error signal, which can be used to mitigate the biased filter estimation.

[0035] Figure 3 An exemplary flowchart is shown, which shows the howling suppression method according to the present invention.

[0036] In step 310, an input signal is provided to a portable karaoke system, the input signal including a source signal (such as a user's singing voice) and acoustic feedback. At least one microphone captures this part of the input signal; in addition, the input signal further includes an audio stream (such as the accompaniment music of a song), and this part of the input signal can be uploaded to the karaoke system in a wired or wireless manner (such as via Bluetooth or through an AUX interface).

[0037] In step 315, decorrelation is introduced in the provided system to decorrelate the input signal. In this step, the speaker signal of the speaker is decorrelated from the input signal by frequency shifting the input signal.

[0038] In step 320, the frequency-shifted input signal is then processed (including compression and equalization) in the electroacoustic path and then amplified by a gain factor K to obtain an output signal.

[0039] In step 340, the output signal is fed to a speaker for playback and propagation in the environment.

[0040] In this step, after propagation in the environment, the output signal playback from the speaker is picked up by the microphone as acoustic feedback through a direct path or some reflection paths, and this acoustic feedback enters the microphone when at least one microphone receives other parts of the input signal, as mentioned above in step 310.

[0041] Next, in step 360, in order to eliminate unwanted components from the microphone signal, several IIR filters are used to model the environmental transfer function. This step includes off-line measuring and estimating the environmental transfer function, and approximating the function by, for example, a multi-band IIR filter. The resulting estimated signal is approximately equal to the acoustic feedback portion of the input signal that enters the microphone in step 350. Therefore, by subtracting the estimated acoustic feedback from the input signal captured by the microphone, the acoustic feedback that can cause howling in the system can be eliminated.

[0042] Figure 4 FIG. shows an exemplary hardware system of a portable all-in-one Karaoke system, in which the low-complexity howling suppression method provided herein is implemented in the portable all-in-one Karaoke system. As Figure 4 shown in, beamforming techniques are additionally used in the howling suppression method, and the beamforming techniques use a microphone array or a speaker array to modify the directivity.

[0043] For example, two microphones with different directivities of the microphone array 410 are used to form Figure 4 the cardioid directivity pattern shown in. This can also be regarded as a special beamforming arrangement for suppressing more howling components. It should be understood that more microphones can be used to arrange the microphone array.

[0044] The beamforming output X 波束 (z) is written as:

[0045] X 波束 (z) = ∑ k W k (z)·X k (z) (7)

[0046] where W k (z) and X k (z) are the k-th beamforming filter and the k-th microphone input signal, respectively.

[0047] In addition, in an actual portable Karaoke product, the microphone 410 can be wrapped with sound-absorbing cotton 480 to further reduce the sound energy from the speaker 440. It can be seen that a passive radiator 470 is alternatively used in this exemplary machine, as Figure 4 schematically shown in.

[0048] Figures 5A to 5E Some simulated comparison results are shown. Figure 5A shows the sound spectrum of the speaker signal, and Figure 5B shows the howling sound at approximately 2 kHz and 9.3 kHz, which increases sharply and continues until the end. In Figure 5CThrough the frequency shift method, the howling frequency is reduced in each loop and the howling power is suppressed to a lower level, but the howling is still obvious. Figure 5D It is shown that the NFS method can successfully suppress howling, but it only starts to work when the howling can be heard and detected. The sound spectrum obtained by the low-complexity howling suppression method provided by the present invention ( Figure 5E shown in) indicates that there is no obvious increase in the sound energy at the problematic frequencies, which means that the provided method can successfully and significantly suppress howling feedback and performs best among these methods.

[0049] In the case of a portable karaoke machine, it is always desired to have a long playing time but still expect good sound quality, such as high volume and low howling problems. In the present invention, the provided low-complexity howling suppression method adopts an IIR filter structure to reduce power consumption and system delay. To further suppress howling, a non-linear algorithm (such as frequency shift) is also combined with a microphone beamforming method.

[0050] The low-complexity howling suppression method and system provided by the present invention are suitable for those applications in which the system contains both a speaker and a microphone with relatively fixed positions and the speaker plays the input signal of the microphone in real time. Exemplary applications include, but are not limited to, portable karaoke machines, integrated speakers, and conference systems, etc.

[0051] As used in this application, an element or step stated in the singular and preceded by the word "a" or "an" should be understood not to exclude a plurality of the said elements or steps, unless the exclusion is specified. In addition, the reference to "an embodiment" or "an example" of the present disclosure is not to be construed as excluding the existence of additional embodiments that also include the stated features. The terms "first", "second", and "third", etc. are used only as labels and are not intended to impose a numerical requirement or a particular positional order on their objects.

[0052] Although the exemplary embodiments are described above, these embodiments are not intended to describe all possible forms of the present invention. Rather, the words used in the specification are for description rather than limitation, and it should be understood that various changes can be made without departing from the spirit and scope of the present invention. Additionally, the features of various implemented embodiments can be combined to form other embodiments of the present invention.

Claims

1. A low-complexity howling suppression method for a portable karaoke system, the method comprising the following steps: Capturing an input signal through at least one microphone, the input signal including a source signal and acoustic feedback; Compressing and equalizing the input signal in an electroacoustic path and then feeding it to an output signal after amplification; And Playing back the output signal through a speaker, the output signal propagating in the environment; Wherein the method further comprises the steps of: estimating the acoustic feedback by at least one infinite impulse response (IIR) filter and thereby eliminating the acoustic feedback from the input signal, Wherein the method further comprises the steps of: Modeling an environmental transfer function by the at least one IIR filter, and Wherein the at least one IIR filter includes at least one multi-band IIR filter, and modeling the environmental transfer function includes offline measuring and estimating the environmental transfer function and approximating the environmental transfer function by the at least one multi-band IIR filter.

2. The method according to claim 1, wherein the acoustic feedback includes howling generated in a closed signal loop formed by acoustic coupling between the speaker and the at least one microphone.

3. The method according to claim 1, wherein the relative positions of the at least one microphone and the speaker are fixed.

4. The method according to claim 1, further comprising decorrelating the output signal from the input signal by frequency-shifting the input signal.

5. The method according to claim 1, wherein approximating the environmental transfer function can be performed by the at least one multi-band IIR filter.

6. The method according to claim 1, the method further comprising arranging the at least one microphone in different directivities as a microphone array for beamforming.

7. The method according to claim 6, wherein the microphone array forms a cardioid directivity pattern.

8. A portable karaoke system having a howling suppression function, the system comprising: At least one microphone for capturing an input signal, the input signal including a source signal and acoustic feedback propagating in the environment; An electroacoustic path for compressing and equalizing the input signal and then feeding it to an output signal after amplification; A speaker for playing back the output signal, Wherein the system further includes at least one infinite impulse response (IIR) filter for estimating the acoustic feedback and thereby eliminating the acoustic feedback from the input signal, the at least one infinite impulse response (IIR) filter further modeling an environmental transfer function, and Wherein the at least one IIR filter further includes at least one multi-band IIR filter for offline measuring and estimating the environmental transfer function and approximating the environmental transfer function.

9. The system according to claim 8, wherein the acoustic feedback includes a howling generated in a closed signal loop formed by acoustic coupling between the loudspeaker and the at least one microphone.

10. The system according to claim 8, wherein the relative positions of the at least one microphone and the loudspeaker are fixed.

11. The system according to claim 8, wherein the at least one IIR filter further includes decorrelating the output signal from the input signal by frequency-shifting the input signal.

12. The system according to claim 8, wherein the environmental transfer function can be approximated by the at least one multi-band IIR filter.

13. The system according to claim 9, wherein the at least one microphone is further arranged as a microphone array with different directivities for beamforming.

14. The system according to claim 13, wherein the microphone array forms a cardioid directivity pattern.

Images