CN104902420A - Method and system for realizing wide dynamic range compression based on curve compression - Google Patents

Abstract

The invention discloses a method for realizing wide dynamic range compression based on curve compression, comprising the steps as follows: S1, performing FFT to an input audio time domain signal to obtain a frequency domain sequence; S2, performing multichannel wide dynamic range compression to the frequency domain sequence; S3, calculating an average sound pressure value of a frequency domain signal of each sub-frequency band; S4, calculating an output sound pressure level according to an input sound pressure level; S5, calculating an amplitude gain value according to the input sound pressure level and the output sound pressure level; S6, performing IFFT to the amplitude and outputting a time-domain signal. The invention further provides a webpage instant messaging system with an e-mailed response. By using the manner of curve compression, the wide dynamic range realized on basis of curve compression adds a configurable parameter curve compression factor to reach sound pressure level compression to realize a function of compensating sound with common wide dynamic range compression algorithm. The method and the system of the invention well solve the problem that a straight line compression mode is not flexible and enables a processing result to be more idea and a user to have better experience.

Description

The wide dynamic range compression method realized based on curve compression and system

Technical field

The present invention relates to digital signal technique field, particularly relate to a kind of wide dynamic range compression method based on curve compression realization and system.

Background technology

In all kinds of algorithms of hearing aids, hearing compensation algorithm is the most basic, most important a kind of algorithm.The object of hearing compensation algorithm is that carrying out compression to sound amplifies, and by the hearing dynamic range of normal person by after algorithm process, is mapped in the hearing dynamic range of hearing impaired subject, and keeps identification and the low distortion of sound as far as possible.

As shown in Figure 1, patient and normal person's threshold of audibility and threshold of pain audiogram, reflect a certain patient and normal person respectively in the threshold of audibility and the threshold of pain level of each Frequency point.Hearing loss the chances are 50dB, wherein, TH _hifor the threshold of audibility of patient, UL _hifor the threshold of pain of patient; TH _nfor the threshold of audibility of normal person, UL _nfor the threshold of pain of normal person.

Listen threshold level and the threshold of pain level of normal person are certain, because sex reason may difference to some extent; But listen threshold level and the threshold of pain level of hearing loss patient are but that difference is totally different because individuality is different.

The threshold of audibility refers to, in a certain frequency range, a people can just hear the sound pressure level size of sound; The threshold of pain refers to, a receptible maximum hearing sound pressure level size of people.In general, the threshold of audibility of hearing impaired subject listens threshold value large than normal person, dynamic range on whole sub-audible sound frequency domain all reduces, and it directly embodies is exactly in a certain frequency range, and the higher or lower sound hearing impaired listener that normal person can hear cannot hear.

For compensating the hearing of this partial loss, make hearing impaired listener can hear the sound of normal person, this concept that DSD designer proposes " compression ".As shown in Figure 2, be I/O figure after a typical dynamic range compression acoustic pressure compensates.Wherein, be respectively linear amplification region along input/output sequence, straight line compressional zone, export-restriction district.

General, wide dynamic range compression algorithm adopts the mode of straight line compression in the middle of the threshold of audibility to the threshold of pain, although normal person's hearing dynamic range also can be mapped within the hearing dynamic range of hearing impaired subject by this processing method, the feedback of hearing impaired subject to the sound after compensation can not be responded well.Some patient may think that sound exceeds the expection of oneself enlargedly, and some patient may think that sound does not reach the expection of oneself enlargedly again.In brief, the mode underaction of straight line compression, does not reach optimal user and experiences.

Summary of the invention

In view of this, the invention provides a kind of wide dynamic range compression method based on curve compression realization and system, solve the shortcoming of straight line compress mode underaction, make result better.

The embodiment of the present invention provide a kind of based on curve compression realize wide dynamic range compression method, comprising: S1, to input audio time domain signal carry out fast fourier transform, obtain frequency domain sequence; S2, multi-channel wide dynamic range compression is carried out to described frequency domain sequence, be divided into multiple sub-frequency band; S3, calculate the average sound pressure value of the frequency-region signal of described how sub-frequency band, determine to input sound pressure level; S4, according to input sound pressure level calculate output sound pressure level; S5, calculate amplitude gain value according to described input sound pressure level and output sound pressure level; S6, inverse Fourier is carried out to amplitude become, output time-domain signal.

The embodiment of the present invention also provides a kind of wide dynamic range systems realized based on curve compression, comprising: Fast Fourier Transform Block, for carrying out fast fourier transform to the audio time domain signal of input, obtains frequency domain sequence; Multi-channel wide dynamic range compression module, for carrying out multi-channel wide dynamic range compression to described frequency domain sequence, is divided into multiple sub-frequency band; First computing module, for calculating the average sound pressure value of the frequency-region signal of described multiple sub-frequency band, determines to input sound pressure level; Second computing module, for calculating output sound pressure level according to input sound pressure level; 3rd computing module, for calculating amplitude gain value according to described input sound pressure level and output sound pressure level; Inverse Fourier transform module, for carrying out inverse Fourier transform to amplitude, output time-domain signal.

The wide dynamic range realized based on curve compression of the present invention adopts the mode of curve compression, increase this configurable parameter of the curve compression factor and reach sound pressure level compression, achieve the function of general wide dynamic range compression algorithm compensating sound, solve the shortcoming of straight line compress mode underaction equally well, make result better, make user have better experience.

Accompanying drawing explanation

In order to be illustrated more clearly in the embodiment of the present invention or technical scheme of the prior art, be briefly described to the accompanying drawing used required in embodiment or description of the prior art below, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.

Fig. 1 is existing patient and normal person's threshold of audibility and threshold of pain audiogram;

Fig. 2 is the input and output figure after existing dynamic range compression acoustic pressure compensates.

Fig. 3 is the input and output figure after dynamic range compression acoustic pressure compensation according to an embodiment of the invention;

Fig. 4 is the flow chart of wide dynamic range method according to an embodiment of the invention;

Fig. 5 is the structural representation of wide dynamic range systems according to an embodiment of the invention.

Specific embodiment

Below in conjunction with the drawings and specific embodiments, more detailed description is further done to technical scheme of the present invention.Obviously, described embodiment is only a part of embodiment of the present invention, instead of whole embodiments.Based on the embodiment in the present invention, the every other embodiment that those of ordinary skill in the art obtain under the prerequisite of not making creative work, all should belong to the scope of protection of the invention.

The present invention proposes a kind of wide dynamic range compression method realized based on curve compression, adopts the mode of curve compression, increases this configurable parameter of the curve compression factor and reaches sound pressure level compression.As shown in Figure 3, wherein, b2 represents normal straight compress mode, and b1 represents that curve makes up user and feels the fraud place of undercompensation, and b3 represents that curve makes up user and feels to compensate the fraud place exceeding expection.

In this wide dynamic range compression method, input discrete time signal x [n], through framing, Hamming windowing, FFT (Fast Fourier Transformation, fast fourier transform) transform to signal in frequency domain, again by solving the gain in 7 passages, at wide dynamic range compression (Wide Dynamic Range Compression, WDRC) after process, through IFFT (InverseFast Fourier Transformation, inverse Fourier transform), frequency-region signal is converted to time-domain signal again, exports x ' (n).

As shown in Figure 4, the present embodiment provides a kind of wide dynamic range compression method realized based on curve compression, comprising:

Step S1, to input audio time domain signal carry out FFT, obtain frequency domain sequence;

Particularly, to audio time domain signal x (n) of input, n=0 ~ 127, carry out FFT, obtain frequency domain sequence X (N), N=0 ~ 127.

S2, multi-channel wide dynamic range compression is carried out to described frequency domain sequence, be divided into multiple sub-frequency band;

Preferably, multichannel is 8 passages, and the frequency-region signal gain of each channel frequence band is unified value.The embodiment of the present invention gets f0=62Hz, f1=125Hz, f2=250Hz, f3=500Hz, f4=1000Hz, f5=2000Hz, f6=4000Hz, f7=8000Hz.

S3, calculate the average sound pressure value of the frequency-region signal of described multiple sub-frequency band, determine to input sound pressure level;

Described average sound pressure value

$\mathrm{spl}\left(n\right)=10\lg \left(\frac{\underset{k}{\mathrm{\Σ}}{X}^2\left(k\right)}{m\×\mathrm{Pr}{\mathrm{ef}}^2\×N}\right)$

Wherein, X (k) is for dropping on the frequency-region signal in a frequency band, and m drops on the signal number in this band frequency band, and N is the sampling number of FFT, and Pref is reference acoustic pressure.

Average energy P (the n)=X calculating input signal was also comprised before average sound pressure value calculates ²(k); And calculate instantaneous sound pressure SP (n)=20lg (P (n)/P _ref), wherein, Healthy People ear auditory perception has extremely wide dynamic range, can reach 0 ~ 140dB.

As expressed sound with Pascal (Pa), must process little of 20uPa, greatly to 2, the numeral of 000,000,000Pa.Thus, expressing sound or noise with Pa can rather inconvenience, and better simply way is the loud degree of expressing sound by a logarithmic scale.Herein, using 10 as radix, for avoiding Pa to need to process large amount of complex to express sound, heavy handled numeral, therefore by decibel (dB) this scale, this scale is using threshold of hearing 20uPa as reference sound pressure level, and to define this sound pressure level be 0dB.) reference acoustic pressure is P _ref=20uPa is also reference sound pressure value.

S4, calculate output sound pressure level spl_out according to input sound pressure level spl_in;

The acoustic pressure scope of sound pressure level is judged according to input sound pressure level, then according to acoustic pressure range computation output sound pressure level.

(1)spl_out[i]＝spl[i]+(THhi[i]-THn[i])

$\left(2\right)\mathrm{spl}\_\mathrm{out}\left[i\right]=\mathrm{THhi}\left[i\right]+\frac{\mathrm{ULhi}\left[i\right]-\mathrm{THhi}\left[i\right]}{\mathrm{ULn}\left[i\right]-\mathrm{THn}\left[i\right]}\left(\mathrm{spl}\right[i]-\mathrm{THn}[i\left]\right)$

(3)spl_out[i]＝ULhi[i]

Wherein, acoustic pressure scope comprises below 1-normal person's threshold of audibility; Between 2-normal person's threshold of audibility to the threshold of pain; More than the 3-normal person threshold of pain.

Input sound pressure level be in be less than normal person's threshold of audibility, input sound pressure level is between normal person's threshold of audibility threshold of pain, input sound pressure level is on the threshold of pain, and its computing formula is respectively as follows:

$O=\frac{O1}{I1}*I$

$O={\left(\frac{I-I1}{I2-I1}\right)}^{\mathrm{\α}}\×(O2-O1)+O1$

O＝O2

Wherein, a is exactly the curve compression factor, and I is input sound pressure level, and O is output sound pressure level.

S5, calculate amplitude gain value gain according to described input sound pressure level and output sound pressure level;

$\mathrm{spl}=10\lg \frac{X^2}{\mathrm{Pref}}$

$\mathrm{spl}\_\mathrm{out}=10\lg \frac{X^{\′2}}{\mathrm{Pr\; ef}}$

$\mathrm{spl}\_\mathrm{out}-\mathrm{spl}=10\lg \frac{X^{\′2}}{\mathrm{Pr\; ef}}-10\lg \frac{X^2}{\mathrm{Pr\; ef}}=10\lg {\left(\frac{X^{\′}}{X}\right)}^2=20\lg \frac{X^{\′}}{X}$

$\⇒\mathrm{gain}=\frac{X^{\′}}{X}={10}^{\frac{\mathrm{spl}\_\mathrm{out}-\mathrm{spl}}{20}}$

After described calculating amplitude gain value, spectrum signal is multiplied by described amplitude gain value, the amplitude X ' (0) be compensated ~ X ' (127).

S6, IFFT is carried out to amplitude, output time-domain signal.

IFFT is carried out to X ' (0) ~ X ' (127), obtains time domain and export, export as a frame these 64 of retaining above.

As shown in Figure 5, the present embodiment provides a kind of wide dynamic range compressing system realized based on curve compression, realizes, specifically comprise for wide dynamic range compression method:

FFT module 1, for carrying out FFT to the audio time domain signal of input, obtains frequency domain sequence;

Multi-channel wide dynamic range compression module 2, for carrying out multi-channel wide dynamic range compression to described frequency domain sequence, is divided into multiple sub-frequency band; ;

First computing module 3, for calculating the average sound pressure value of the frequency-region signal of described multiple sub-frequency band, determines to input sound pressure level;

Second computing module 4, for calculating output sound pressure level according to input sound pressure level;

3rd computing module 5, for calculating amplitude gain value according to described input sound pressure level and output sound pressure level;

IFFT module 6, for carrying out IFFT to amplitude, output time-domain signal.

Preferably, described second computing module 4 also comprises judge module, for judging the acoustic pressure scope of sound pressure level, and calculates output sound pressure level according to acoustic pressure.

Described IFFT module 6 also comprises compensating unit, for being multiplied by spectrum signal to described amplitude gain value, the amplitude be compensated.

It should be noted that, by the description of above embodiment, those skilled in the art can be well understood to the mode that the present invention can add required hardware platform by software and realize, and can certainly all be implemented by hardware.Based on such understanding, what technical scheme of the present invention contributed to background technology can embody with the form of software product in whole or in part, described computer software product can be stored in storage medium, as ROM/RAM, magnetic disc, CD etc., comprising some instructions in order to make a computer equipment (can be personal computer, server, or the network equipment etc.) perform the method described in some part of each embodiment of the present invention or embodiment.

The above disclosed preferred embodiment be only in the embodiment of the present invention, certainly can not limit the interest field of the present invention, therefore according to the equivalent variations that the claims in the present invention are done, still belong to the scope that the present invention is contained with this.

Claims (10)

1., based on the wide dynamic range compression method that curve compression realizes, it is characterized in that, comprising:

S1, to input audio time domain signal carry out fast fourier transform, obtain frequency domain sequence;

S2, multi-channel wide dynamic range compression is carried out to described frequency domain sequence, be divided into multiple sub-frequency band;

S3, calculate the average sound pressure value of the frequency-region signal of described multiple sub-frequency band, determine to input sound pressure level;

S4, according to input sound pressure level calculate output sound pressure level;

S5, calculate amplitude gain value according to described input sound pressure level and output sound pressure level;

S6, inverse Fourier transform is carried out to amplitude, output time-domain signal.

2., as claimed in claim 1 based on the wide dynamic range compression method that curve compression realizes, it is characterized in that, in described S2, multichannel is 8 passages, and the frequency-region signal gain of each channel frequence band is unified value.

3., as claimed in claim 1 based on the wide dynamic range compression method that curve compression realizes, it is characterized in that, in described S3, described average sound pressure value is according to formula:

$\mathrm{spl}\left(n\right)=101g\left(\frac{\underset{k}{\mathrm{\Σ}}{X}^2\left(k\right)}{m\×\mathrm{Pr}{\mathrm{ef}}^2\×N}\right)$

Calculate, wherein, X (k) is for dropping on the frequency-region signal in a frequency band, and m drops on the signal number in this band frequency band, and N is that the employing of fast fourier transform is counted, and Pref is reference acoustic pressure.

4. as claimed in claim 1 based on the wide dynamic range compression method that curve compression realizes, it is characterized in that, in described S4, judge the acoustic pressure scope of sound pressure level according to input sound pressure level, then according to acoustic pressure range computation output sound pressure level.

5., as claimed in claim 1 based on the wide dynamic range compression method that curve compression realizes, it is characterized in that, in described S5, after described calculating amplitude gain value, spectrum signal is multiplied by described amplitude gain value, the amplitude be compensated.

6., based on the wide dynamic range compressing system that curve compression realizes, it is characterized in that, comprising:

Fast Fourier Transform Block, for carrying out fast fourier transform to the audio time domain signal of input, obtains frequency domain sequence;

Multi-channel wide dynamic range compression module, for carrying out multi-channel wide dynamic range compression to described frequency domain sequence, is divided into multiple sub-frequency band;

First computing module, for calculating the average sound pressure value of the frequency-region signal of described multiple sub-frequency band, determines to input sound pressure level;

Second computing module, for calculating output sound pressure level according to input sound pressure level;

3rd computing module, for calculating amplitude gain value according to described input sound pressure level and output sound pressure level;

Inverse Fourier transform module, for carrying out inverse Fourier transform to amplitude, output time-domain signal.

7., as claimed in claim 6 based on the wide dynamic range compressing system that curve compression realizes, it is characterized in that, described multichannel is 8 passages, and the frequency-region signal gain of each channel frequence band is unified value.

8., as claimed in claim 6 based on the wide dynamic range compressing system that curve compression realizes, it is characterized in that, described average sound pressure value is according to formula:

$\mathrm{spl}\left(n\right)=101g\left(\frac{\underset{k}{\mathrm{\Σ}}{X}^2\left(k\right)}{m\×\mathrm{Pr}{\mathrm{ef}}^2\×N}\right)$

Calculate, wherein, X (k) is for dropping on the frequency-region signal in a frequency band, and m drops on the signal number in this band frequency band, and N is that the employing of fast fourier transform is counted, and Pref is reference acoustic pressure.

9. as claimed in claim 6 based on the wide dynamic range compressing system that curve compression realizes, it is characterized in that, described second computing module also comprises judge module, for judging the acoustic pressure scope of sound pressure level according to input sound pressure level, then according to acoustic pressure range computation output sound pressure level.

10., as claimed in claim 6 based on the wide dynamic range compressing system that curve compression realizes, it is characterized in that, described inverse Fourier transform module also comprises compensating unit, for being multiplied by spectrum signal to described amplitude gain value, the amplitude be compensated.