CN104936088A - Mixed virtual bass enhancing method - Google Patents

Abstract

The invention discloses a mixed virtual bass enhancing method. The method includes a signal processing process, and aims at enhancing the bass effect of audio signals. The method comprises the following steps that the same pathway signal of original audio is divided into two paths, high-pass filtering at the initial frequency of Fc is carried out on one path of the signals to obtain part of signals which is higher than the loudspeaker cutoff frequency Fc, time delay processing is carried out on the path of signals to obtain high-frequency signals, and low-pass filtering at the cutoff frequency Fc is carried out on the other path of signals to obtain low-frequency signals; the low-frequency signals are separated to obtain transient and stable components, and virtual bass processing are carried out on the transient and stable components respectively to obtain virtual low-frequency signals, and the virtual low-frequency signals are overlapped with high-frequency signals to obtain virtual bass enhanced audio signals. The audio signals after virtual bass processing reserve original quality with enhanced bass effect, and the audio loudness, fullness, deepness and spaciousness are enhanced.

Description

The virtual bass boosting method of a kind of mixing

Technical field

The present invention relates to audio signal processing technique field, be specifically related to the virtual bass boosting method of a kind of mixing.

Background technology

Along with miniaturization, the lighting of multimedia equipment, the loud speaker embedded in such devices has very strict restriction dimensionally, because the volume of loud speaker limits, the low frequency reproduction poor ability of miniature loudspeaker, but the low-frequency component in audio frequency plays a part very important to audition impression, directly affects the loud and clear degree of sound, richness, simple and honest sense and spatial impression.The low frequency behavior how improving miniature loudspeaker is Audio Design field problem demanding prompt solution.

Traditional low frequency Enhancement Method adopts audio equalizer directly to strengthen low frequency energy, and this method can cause loudspeaker enclosure reduction, replay signal produces distortion, even may damage speaker system time serious.Comparatively speaking, virtual bass strengthens technology is the comparatively effective solution of one.Virtual bass technology is formed at psychologic acoustics theory, during people's perception supper bass, mainly do not rely on the fundamental frequency of bass, and be each harmonic relying on this fundamental frequency more, even if the fundamental frequency of bass signal is suppressed, as long as the relation of its each harmonic and these harmonic waves still exists, so for people's ear, bass effect still can be felt.Virtual bass strengthens technology and make use of this phenomenon, by the composition filtering lower than loud speaker cut-off frequency, and suitably increase the harmonic energy at its times of frequency place, utilize loud speaker in the stronger advantage of the play capability of harmonic wave frequency range, acoustically can fictionalize bass effect from subjectivity.

Existing virtual bass boost algorithms mainly based on nonlinear device and based on phase vocoder also have based on both combine the virtual bass boost algorithms of mixing, method processing speed based on nonlinear device is very fast, the bass effect produced for transient signal is better, such as tum etc., but can nonlinear distortion be produced for steady-state signal.Be obtain the Time-Frequency Information of audio signal by fourier transformation in short-term based on phase vocoder algorithm, the mode then by increasing signal low-frequency component phase change rate produces harmonic wave.This Measures compare is flexible, effectively control signal distortion, but processing speed is slower, and phase vocoder is in order to obtain higher low frequency resolution, the analysis window that one very large must be had in the time domain, certain distortion can be produced like this, so have good performance based on the bass boost algorithms process steady-state signal of phase vocoder to transient signal.The virtual bass of mixing strengthens the advantage that system combines nonlinear device and phase vocoder, by audio signal respectively by two kinds of harmonic generation devices, by the signal obtained through a harmonic energy control module with the ratio both distributing.But, because the audio signal being input to two kinds of harmonic generation devices is the primary signal comprising transient components and stationary component, so can with the distortion of steady-state signal based on the harmonic wave of nonlinear device generation, and can with the distortion of transient signal based on the harmonic wave of phase vocoder generation, therefore the last virtual bass signal produced contains a large amount of distortions, can not the tonequality of good rediscover.

Summary of the invention

The object of the invention is the shortcoming and defect existed to solve prior art, providing a kind of novel mixing virtual bass boosting method.This process employs Sound seperation algorithm, carry out corresponding virtual bass process more respectively after the transient state in original audio signal and steady-state signal first being separated, overcome the deficiency that the conventional method distortion factor is high, significantly improve the effect that virtual bass strengthens.

In order to achieve the above object, present invention employs following technical proposals:

The virtual bass boosting method of a kind of mixing, first the same path signal of original audio is divided into two-way by this signal processing method, wherein a road obtains, higher than the part of loud speaker cut-off frequency Fc, after delay process, obtaining high-frequency signal after the high-pass filtering process that initial frequency is Fc; The low-pass filtering treatment that another road signal is Fc through cut-off frequency obtains low frequency signal; By low frequency signal through signal separation process, obtain transient components and stationary component; Transient components and stationary component are obtained virtual low frequency signal respectively through after corresponding virtual bass process, and superposes with high-frequency signal, obtain the audio signal after the enhancing of virtual bass.Its concrete steps are as follows:

(1) the same path signal of original audio is divided into two-way, wherein a road obtains higher than loud speaker cut-off frequency Fc through high-pass filtering process part, then obtains high-frequency signal after delay process; Another road obtains low frequency signal after the low-pass filtering treatment that cut-off frequency is Fc.

(2) by low frequency signal through transient state/stable state separation module, obtain transient components and the stationary component of low frequency signal.

(3) stationary component of low frequency signal is passed through the virtual bass generation module based on phase vocoder, obtain the virtual bass of stationary component; By the transient components of low frequency signal through the virtual bass generation module based on nonlinear device, obtain the virtual bass of transient components.

(4) stationary component produced step (3) according to contour of equal loudness and the virtual bass of transient components carry out harmonic amplitude control, each harmonic wave loudness is made to meet contour of equal loudness, and be collectively referred to as a road, the virtual bass signal after the amplitude that obtains controls.

(5) the virtual bass signal after the amplitude that high-frequency signal step (1) obtained and step (4) obtain controls synthesizes a road, obtains the audio signal after virtual bass process.

The virtual bass boosting method of one mixing of the present invention compared with prior art, has following advantage:

Contemplated by the invention in audio signal and generally there is transient components and stationary component, first transient state/steady-state signal separation algorithm is utilized the transient components in original audio and stationary component to be separated, recycle traditional virtual bass generation module based on nonlinear device and the virtual bass generation module based on phase vocoder carries out virtual bass process, greatly reduce stable state distortion and the transient distortion of virtual bass process, the audio signal after virtual bass process is made to enhance bass effect while reservation original quality, enhance the loud and clear degree of audio frequency, richness, simple and honest sense and spatial impression.

Accompanying drawing explanation

Fig. 1 is the main-process stream block schematic illustration that the present invention mixes virtual bass boosting method.

Fig. 2 is the idiographic flow schematic diagram of the virtual bass processing module in Fig. 1.

Embodiment

Below in conjunction with embodiment and accompanying drawing, the present invention is described in further detail, but embodiments of the present invention are not limited thereto.

The virtual bass boosting method of mixing of the present invention, transient components and stationary component is contain in audio signal, transient state/steady-state signal separation algorithm is utilized the transient components in original audio signal and stationary component to be separated, then virtual bass process is carried out respectively by the virtual bass generation module based on nonlinear device and the virtual bass generation module based on phase vocoder, then virtual for the two-way obtained bass signal is controlled make it meet contour of equal loudness and synthesize a road through harmonic amplitude, finally the high-frequency signal that the virtual bass signal obtained and original audio signal obtain after high-pass filtering and delay process is superposed, obtain mixing virtual bass and strengthen the audio signal after processing.As shown in Figure 1 and Figure 2, concrete implementation step is as follows:

(1) by original audio signal x _orit () is divided into two-way, initial frequency of wherein leading up to is the high pass filter of Fc, obtains the signal higher than loud speaker cut-off frequency Fc, then obtains the high-frequency signal x of primary signal through delay process module _ht (), cut-off frequency of separately leading up to is the low pass filter of Fc, obtains the low frequency signal x lower than loud speaker cut-off frequency Fc _l(t);

(2) by low frequency signal x _lt () is input to virtual bass processing module, obtain virtual bass signal x ' _l(t), concrete implementation step is as follows:

(2-1) by low frequency signal x _lt transient signal and steady-state signal can be separated by () after HPSS Sound seperation algorithm process from primary signal, obtain steady-state signal x _{l_H}(t), and transient signal x _{l_P}(t).Concrete separation algorithm is as follows:

A. by low frequency signal x _lt () carries out Short Time Fourier Transform (STFT), obtain the time-frequency spectrum matrix of low frequency signal, be designated as be the complex matrix of a N × K, wherein N represents time-frequency spectrum time number of samples altogether, and K represents the frequency number of samples that each time sampling point has.

B. to primary signal time-frequency spectrum carry out modulo operation, obtain its amplitude spectrum Y, Y is the real number matrix of a N × K, Y _n,krepresent the some time frequency point in time-frequency spectrum amplitude spectrum, n and k is integer, represents time index and the frequency indices of this time frequency point;

C. the initial value arranging steady-state signal amplitude spectrum H and instantaneous signal amplitude spectrum P is all wherein γ is decay factor, in the present embodiment γ=0.5;

D. following formula is utilized to carry out iterative processing:

$H_{n,k}^{\mathrm{\γ}}\←\frac{H_{n,k}^{(n-\mathrm{mean})}}{\sqrt{{\left(H_{n,k}^{(n-\mathrm{mean})}\right)}^2+{\left(P_{n,k}^{(n-\mathrm{mean})}\right)}^2}}{Y}_{n,k}^{\mathrm{\γ}}---\left(1\right)$

$P_{n,k}^{\mathrm{\γ}}\←\frac{P_{n,k}^{(n-\mathrm{mean})}}{\sqrt{{\left(H_{n,k}^{(n-\mathrm{mean})}\right)}^2+{\left(P_{n,k}^{(n-\mathrm{mean})}\right)}^2}}{Y}_{n,k}^{\mathrm{\γ}}---\left(2\right)$

Wherein,

$H_{n,k}^{(n-\mathrm{mean})}=\frac{1}{2N^{\′}}\underset{n^{\′}=1}{\overset{N^{\′}}{\mathrm{\Σ}}}(H_{n+n^{\′},k}^{\mathrm{\γ}}+H_{n-n^{\′},k}^{\mathrm{\γ}})---\left(3\right)$

$P_{n,k}^{(n-\mathrm{mean})}=\frac{1}{2K^{\′}}\underset{k^{\′}=1}{\overset{K^{\′}}{\mathrm{\Σ}}}(P_{n,k+k^{\′}}^{\mathrm{\γ}}+P_{n,k-k^{\′}}^{\mathrm{\γ}})---\left(4\right)$

E. the steady-state signal amplitude spectrum H obtained and the following formula of instantaneous signal amplitude spectrum P are obtained Weiner filter;

$W_H=\frac{H}{\sqrt{H^2+P^2}}---\left(5\right)$

$W_P=\frac{P}{\sqrt{H^2+P^2}}---\left(6\right)$

F. by primary signal time-frequency spectrum be multiplied by Weiner filter, obtain steady-state signal time-frequency spectrum with transient signal time-frequency spectrum computing formula is as follows;

$\hat{H}=\hat{Y}\×W_H---\left(7\right)$

$\hat{P}=\hat{Y}\×W_P---\left(8\right)$

G. to the steady-state signal time-frequency spectrum obtained with transient signal time-frequency spectrum carry out inverse Short Time Fourier Transform (iSTFT), obtain the steady-state signal x of time domain _{l_H}the transient signal x of (t) and time domain _{l_P}(t).

(2-2) by above-mentioned transient signal x _{l_P}t () is input to the virtual bass generation module based on nonlinear device, obtain the harmonic signal x ' of transient signal _{l_P}(t).Virtual bass generation module based on nonlinear device is the high order harmonic component utilizing Multiplier elements to produce low-frequency component, if multiplier two ends input signal is all frequency is f ₀tonal signal:

$X={\mathrm{Ae}}^{-j2\mathrm{\π}{f}_{0}t}---\left(9\right)$

Then the output signal of multiplier is frequency is 2f ₀tonal signal:

$\begin{array}{c}{X}_{\mathrm{out}}=X\×X\\ ={\mathrm{Ae}}^{-j2\mathrm{\π}{f}_{0}t}\×{\mathrm{Ae}}^{-j2\mathrm{\π}{f}_{0}t}\\ =A^2{e}^{-2\mathrm{\π}\left(2f_0\right)t}\end{array}---\left(10\right)$

In like manner, if input signal is frequency be respectively f ₀with 2f ₀pure tone, then the frequency outputed signal is 3f ₀;

(2-3) by above-mentioned steady-state signal x _{l_H}t () is input to the virtual bass generation module based on phase vocoder, obtain the harmonic signal x ' of steady-state signal _{l_H}(t).Virtual bass generation module based on phase vocoder utilizes phase vocoder to control harmonic wave, phase vocoder algorithm utilizes phase information to realize the method for the tone changing of signal based on Short Time Fourier Transform, can be applied to time-stretching and the compression of audio frequency.Based on phase vocoder virtual bass generation module first to input steady-state signal x _{l_H}t () carries out Short Time Fourier Transform, Short Time Fourier Transform can by windowing process by signal subsection, carry out FFT to every segment signal, the signal obtained can be combined by one group of sinusoidal signal, changes the phase change rate of sinusoidal signal, its frequency will change, thus generation high order harmonic component, and then inverse FFT operation is carried out to every segment signal, be reduced into time-domain signal, again every segment signal is added, obtains the harmonic signal x ' of steady-state signal _{l_H}(t);

(2-4) by above-mentioned transient state harmonic signal x ' _{l_P}(t) and stable state harmonic signal x ' _{l_H}t () is carried out harmonic amplitude control and is synthesized a road, due to tonal signal and the complex tone signal with different frequency component of different frequency, even if sound pressure level is identical, loudness can not be identical, need to carry out harmonic amplitude control according to contour of equal loudness to the harmonic components generated, make it reach desirable bass effect, then synthesize a road, obtain virtual bass signal x ' _l(t).

(3) the virtual bass signal x ' step (2) obtained _lt high-frequency signal x that () and step (1) obtain _ht () is mixed into a road signal, can obtain the signal x after mixing virtual bass process _bass(t).

In sum, the invention provides the virtual bass boosting method of a kind of mixing, first the method utilizes Sound seperation algorithm, original audio signal is separated into transient state and stable state two parts, feature again for each part signal carries out corresponding virtual bass generating process to it, while guarantee original audio tonequality, effectively reduce virtual bass and strengthen the distortion after processing, enhance the bass effect of audio signal.

Claims (4)

1. mix a virtual bass boosting method, it is characterized in that, the method comprises following processing procedure:

1) by original audio signal x _orit () is divided into two-way, wherein a road signal is the high pass filter of Fc by initial frequency, obtains the part higher than loud speaker cut-off frequency Fc, and obtains high-frequency signal x through delay process _ht (), the low pass filter process that another road signal is Fc through cut-off frequency obtains low frequency signal x _l(t);

2) to described low frequency signal x _lt () carries out signal separation process, obtain the steady-state signal x of low-frequency audio signal _{l_H}(t) and transient signal x _{l_P}(t);

3) to the steady-state signal x of described low frequency signal _{l_H}(t) and transient signal x _{l_P}t () carries out the process of virtual bass harmonic generation, obtain stable state harmonic signal x ' _{l_H}(t) and transient state harmonic signal x ' _{l_P}(t);

4) by described stable state harmonic signal x ' _{l_H}(t) and transient state harmonic signal x ' _{l_P}t () carries out harmonic amplitude control treatment, then synthesize a road, obtains virtual bass signal x ' _l(t);

5) by described high-frequency signal x _h(t) and virtual bass signal x ' _lt () synthesizes, obtain the audio signal x after the enhancing of virtual bass _bass(t).

2. according to claim 1 one mixing virtual bass boosting method, it is characterized in that, described step 2) in described low frequency signal x _lt () is carried out signal separation process process and is comprised the following steps:

A) to described low frequency signal x _lt () carries out Short Time Fourier Transform, obtain the time-frequency spectrum of low frequency signal

B) to described low frequency signal time-frequency spectrum carry out modulo operation, obtain the amplitude spectrum Y of low frequency signal time-frequency spectrum;

C) initial value arranging steady-state signal amplitude spectrum H and instantaneous signal amplitude spectrum P is all wherein γ is decay factor;

D) following formula is utilized to carry out iterative processing:

$H_{n,k}^{\mathrm{\γ}}\←\frac{H_{n,k}^{(n-\mathrm{mean})}}{\sqrt{{\left(H_{n,k}^{(n-\mathrm{mean})}\right)}^2+{\left(P_{n,k}^{(n-\mathrm{mean})}\right)}^2}}{Y}_{n,k}^{\mathrm{\γ}}---\left(1\right)$

$P_{n,k}^{\mathrm{\γ}}\←\frac{P_{n,k}^{(n-\mathrm{mean})}}{\sqrt{{\left(H_{n,k}^{(n-\mathrm{mean})}\right)}^2+{\left(P_{n,k}^{(n-\mathrm{mean})}\right)}^2}}{Y}_{n,k}^{\mathrm{\γ}}---\left(2\right)$

Wherein,

$H_{n,k}^{(n-\mathrm{mean})}=\frac{1}{2N^{\′}}\underset{n^{\′}=1}{\overset{N^{\′}}{\mathrm{\Σ}}}(H_{n+n^{\′}}^{\mathrm{\γ}}+H_{n-n^{\′},k}^{\mathrm{\γ}})---\left(3\right)$

$P_{n,k}^{(n-\mathrm{mean})}=\frac{1}{2K^{\′}}\underset{k^{\′}=1}{\overset{K^{\′}}{\mathrm{\Σ}}}(P_{n,k+k^{\′}}^{\mathrm{\γ}}+P_{{n,k-k}^{\′}}^{\mathrm{\γ}})---\left(4\right)$

E) the steady-state signal amplitude spectrum H obtained and the following formula of instantaneous signal amplitude spectrum P are obtained Weiner filter;

$W_H=\frac{H}{\sqrt{H^2+P^2}}---\left(5\right)$

$W_P=\frac{P}{\sqrt{H^2+P^2}}---\left(6\right)$

F) by described primary signal time-frequency spectrum be multiplied by Weiner filter, obtain steady-state signal time-frequency spectrum with transient signal time-frequency spectrum computing formula is as follows;

$\hat{H}=\hat{Y}\×W_H---\left(7\right)$

$\hat{P}=\hat{Y}\×W_P---\left(8\right)$

G) to the steady-state signal time-frequency spectrum obtained with transient signal time-frequency spectrum carry out inverse Short Time Fourier Transform, obtain the steady-state signal x of time domain _{l_H}the transient signal x of (t) and time domain _{l_P}(t).

3. according to claim 1 one mixing virtual bass boosting method, it is characterized in that, described step 3) in virtual bass harmonic generation processing procedure refer to described transient signal x _{l_P}t () is input to the harmonic generation device be made up of Multiplier elements and feedback loop, obtain the harmonic signal x ' of transient signal _{l_P}(t); By described steady-state signal x _{l_H}t () is input to the harmonic generation device be made up of phase vocoder, obtain the harmonic signal x ' of steady-state signal _{l_H}(t).

4. according to claim 1 one mixing virtual bass boosting method, it is characterized in that, described step 4) in harmonic amplitude control treatment process refer to, according to contour of equal loudness, to described transient state harmonic signal x ' _{l_P}(t) and stable state harmonic signal x ' _{l_H}t () carries out harmonic amplitude adjustment, make it reach desirable bass effect, then synthesize a road, obtains virtual bass signal x ' _l(t).