CN1937854A - Apparatus and method of reproduction virtual sound of two channels - Google Patents

Abstract

A stereo sound generation apparatus of reproducing multi-channel sound input signals through two-channel speakers. The stereo sound generation apparatus includes: a preprocessing filter unit to reduce correlation between two-channel audio signals from among multi-channel audio signals and to generate a presence perception, a virtual speaker filter unit to convert the two-channel audio signals output from the preprocessing filter unit into a virtual sound source at a predetermined position, a signal correction filter unit to correct a signal characteristic between remaining multi-channel audio signals excluding the two-channel audio signals, and the two-channel audio signals output from the virtual speaker filter unit, and an addition unit to add signals to be output to a first channel from among the multi-channel audio signals output from the virtual speaker filter unit and the signal correction filter unit, and to add signals to be output to a second channel from among the multi-channel audio signals output from the virtual speaker filter unit and the signal correction filter unit.

Description

Be used to reproduce the apparatus and method of virtual sound of two channels

Technical field

The present invention relates to a kind of stereophonic sound system, relate to a kind of stereo generation device and method particularly, it is used for producing binaural audio signal in the output gain of adjusting residue channel audio input signal and time delay virtual sound source makes it possible to provide the stereo perception of nature.

Background technology

Usually, audio reproducing system provides surrounding sound effect by only using two loud speakers, such as 5.1 sound channel systems.

The traditional stereo generation system that is used for reproducing by 2 channel loudspeakers 5.1 channel audios has been described in WO 99/49574 (on January 6th, 1999 application, exercise question be the PCT/AU99/00002 of " acoustic signal processing method and device ").

Fig. 1 is the block diagram of the traditional stereo generation of expression system 1.Referring to Fig. 1, described traditional sound generating system comprises: the part relevant with the convolution of input signal and impulse response, and it produces 5.1 channel stereo sensation by head related transfer function (HRTF) is used as down hybrid technology by two-way speaker; And the part that is used for adding convolution signal to dual track.

Referring to Fig. 1, import 5.1 channel audio signal.Described 5.1 sound channels comprise left front sound channel 2, right front channels, the preceding sound channel of central authorities, left surround channel, right surround channel and low-frequency effect (LFE) sound channel.Therefore, be associated corresponding left front impulse response function 4 and left front signal 3 convolution with left front sound channel 2.Left front impulse response function 4 is the impulse response that will be received as the desirable spiking of exporting from the left front channel loudspeaker that is positioned at ideal position by listener's left ear, and uses HRTF.Output signal 7 is added to the left channel signals 10 that is used for earphone.Similarly, be used for the impulse response function 5 and left front signal 3 convolution corresponding to listener's auris dextra of right channel loudspeaker, so that produce the output signal 9 that will be added to right-channel signals 11.

Therefore, before left front sound channel 2, right front channels, the central authorities audio signal of sound channel, left surround channel, right surround channel and LFE sound channel respectively with corresponding impulse response convolution so that be each sound channel generation two signals, i.e. left signal and right signal.Then, with the left signal addition each other of described 6 sound channels, with the right signal addition each other of described 6 sound channels, so that obtain the dual track output signal at last.

If reproduce the dual track output signal, then the loud speaker by two reality produces stereo perception, and is promptly left front, right front, central just as virtual speaker, left around being disposed in around the listener with right circulating loudspeaker.

But, produce system 1 according to the conventional stereo acoustic sound shown in Fig. 1, if the relevance height between left surround channel and right surround channel then is difficult to produce audiovideo at listener's rear portion.

Here, high relevance represents that sound characteristic almost is identical, if explained later relevance height then why be difficult to produce the reason of audiovideo at listener's rear portion.

Use HRTF to form virtual sound source, described HRTF is the feature of the audible signal located at listener's ear (being people's ear), and it depends on the shape of listener's head and ear.By HRTF, the phenomenon perception three-dimensional audio that can produce by feature from pahtfinder hard, the feature of described pahtfinder hard for example is diffraction on the skin of listeners head and the reflection that causes by auricle, except the simple path difference such as time difference (ITD) between the grade differential between the sense of hearing (ILD) and the sense of hearing, three-dimensional audio is also with respect to the incident direction of sound and change.

But, though HRTF makes it possible to easily distinguish the left and right sides audiovideo on horizontal surface, because standard HRTF error makes to be difficult to difference front and back audiovideo.For the position of audiovideo before and after distinguishing, should measure actual user's precise frequency.Because use the artificial head (dummy head) of standard usually, therefore, owing to differing between artificial head and actual user's frequency characteristic causes taking place front/rearly obscuring.

When using surround channel, only when being positioned at left back of listener with right back, audiovideo can obtain the effect of surround channel.When the relevance of the audio input signal of left and right sides surround channel was high, described audiovideo was positioned at the central authorities at listener rear portion.And, owing to use the standard artificial head, therefore also cause taking place front/rearly obscuring, and be difficult to obtain the effect of surround channel.

Summary of the invention

The invention provides a kind of stereo generation device and method, thus, produce the stereo perception that provides by multi-channel speaker system by using the two-way speaker system.In addition, in multi-channel audio signal, generation is used for the virtual sound source of binaural audio signal, and adjusts output gain and the time delay that is used to remain channel audio signal (promptly except described binaural audio signal), makes it possible to provide the stereo perception of nature.

Other aspects of the present invention will partly provide in the explanation of back, and partly apparent from described explanation, perhaps can come cognitive by practice of the present invention.

Above-mentioned and/or other aspects of the present invention can realize by a kind of stereo generation device is provided, described stereo generation device is used for the reproduction of multichannel audio input signal is exported as dual track, described device comprises: the pre-processing filter unit, be used to be reduced in relevance, and produce the sensation of reality between the binaural audio signal of multichannel audio input signal; The virtual speaker filter cell, the binaural audio signal that is used for exporting from described pre-processing filter unit is converted to the virtual sound source in the precalculated position; The signal correction filter cell is used to proofread and correct at the remaining multichannel audio input signal except described dual-channel audio input signal and from the characteristics of signals between the described binaural audio signal of described virtual speaker filter cell output; And, addition unit, the multi-channel audio signal that is used for being added in mutually from described virtual speaker filter cell and the output of described signal correction filter cell will output to the signal of first sound channel, and is added in the signal that will output to second sound channel from the multi-channel audio signal of described virtual speaker filter cell and the output of described signal correction filter cell mutually.

Above-mentioned and/or other aspects of the present invention can also realize by a kind of stereo generation device is provided, described stereo generation device is used for the reproduction of multichannel audio input signal is exported as binaural audio signal, described device comprises: the pre-processing filter unit is used for postponing in groups the predetermined frequency component of the binaural audio signal that (group-delay) select at described multichannel audio input signal; The virtual speaker filter cell is used for selected binaural audio signal from described pre-processing filter unit output is converted to virtual sound source in the precalculated position; The signal correction filter cell is used to proofread and correct at the residue multi-channel audio signal except selected binaural audio signal with from output level and time delay between the selected binaural audio signal of described virtual speaker filter cell output; And, addition unit, the multi-channel audio signal that is used for being added in mutually from described virtual speaker filter cell and the output of described signal correction filter cell will output to the signal of first sound channel, and is added in the signal that will output to second sound channel from the multi-channel audio signal of described virtual speaker filter cell and the output of described signal correction filter cell mutually.

Above-mentioned and/or other aspects of the present invention also can realize by a kind of stereo generation device is provided, described stereo generation device is used for realizing two convolution with matrix structure of pre-sizing by calculating ears synthesizer and the crosstalk canceller relevant with binaural signal in advance, described device comprises: delay cell is used for predetermined separately length of delay and postpones first and second channel input signal; Gain unit is used for being adjusted at each output level of first and second channel input signals that described delay cell postpones; First addition unit is used for described first channel input signal of addition and described gain and postpones controlled second sound channel signal; First filter cell is used to adjust the frequency characteristic from the signal of described first addition unit output; Second addition unit is used for described second channel input signal of addition and described gain and postpones controlled first sound channel signal; And second filter cell is used to adjust the frequency characteristic from the signal of described second addition unit output.

Above-mentioned and/or other aspects of the present invention can also realize by a kind of stereo generation device is provided, described stereo generation device is used for the multichannel audio input signal is reproduced as the dual track output signal, described device comprises: virtual ring is around filter cell, be used to be reduced in relevance, and described two surround channel audio signals are converted to virtual sound source in the precalculated position between two surround channel audio signals of described multichannel audio input signal; Wide stereo generation unit is used for two preceding channel audio signal at described multichannel audio input signal being produced as the stereophonic signal of widening with crosstalk canceller by the convolution ears are synthetic; And, the signal correction filter cell is used to proofread and correct at the remaining multichannel audio input signal except two surround channel signals and described two preceding channel audio signal with from output level and the time delay of described virtual ring between the channel audio signal of filter cell and described wide stereo generation unit output.

Above-mentioned and/or other aspects of the present invention also can realize by a kind of stereo generation device is provided, described stereo generation device comprises: first filter cell, be used for being received in the surround audio signal of at least 5 input audio signals, and produce with respect to the virtual sound source of listening point in the precalculated position; Second filter cell is used for being received in the remaining audio signal of described at least 5 input audio signals, and compensates delay and the gain inequality that is caused in described surround audio signal around filter cell by described virtual ring; And output unit is used to make up first those surround audio signals selected and remaining audio signal producing left output signal, and makes up second those surround audio signals of selecting and remaining audio signal to produce right output signal.

Above-mentioned and/or other aspects of the present invention can also realize by a kind of stereo generation device is provided, described stereo generation device is used for the multichannel audio input signal is reproduced as the dual track output signal, described device comprises: virtual ring is around filter cell, be used for being reduced in relevance between the binaural audio signal of described multichannel audio input signal producing the sensation of reality, and described binaural audio signal is converted to virtual sound source in the precalculated position; The signal correction filter cell is used to proofread and correct at the remaining multichannel audio input signal except described dual-channel audio input signal with from the characteristics of signals of virtual ring between the binaural audio signal of filter cell output; And, addition unit, be used for being added in mutually from described virtual ring and will output to the signal of first sound channel, and be added in the signal that around the multi-channel audio signal of filter cell and the output of described signal correction filter cell, will output to second sound channel from described virtual ring mutually around the multi-channel audio signal of filter cell and the output of described signal correction filter cell.

Above-mentioned and/or other aspects of the present invention also can realize by a kind of stereo production method is provided, described stereo production method is used for virtual effect is applied to binaural signal, and described method comprises: with the frequency band division of first and second sound channel signals is high frequency band and low-frequency band; Each of the described first and second sound channel low band signal of branch sample (decimating); Produce virtual sound source and this virtual sound source of output on the precalculated position by the relevance that is reduced between the branch sample signal (decimated signal) separately; Carry out interpolation with respect to described first and second sound channel signals of exporting as virtual sound source; Described first and second sound channel signals that are interpolated of low-pass filtering; And, described first sound channel signal that is low pass filtering of addition and high frequency first sound channel signal that is delayed, and described second sound channel signal that is low pass filtering of addition and the described high frequency that is delayed second sound channel signal.

Above-mentioned and/or other aspects of the present invention also can realize by a kind of stereo production method is provided, described stereo production method is used for virtual effect is applied to binaural signal, and described method comprises: feel to carry out preliminary treatment filtering by what lower the relevance between first and second sound channel signals and produce reality; With the frequency band division of first and second sound channel signals of described pretreated filtering is high frequency band and low-frequency band; Each of the described first and second sound channel low band signal of branch sample; By with each minute sample signal output as carrying out virtual speaker filtering at the virtual sound source in precalculated position; First and second sound channel signals with respect to the virtual speaker filtering of exporting as virtual sound source are carried out interpolation; Described first and second sound channel signals that are interpolated of low-pass filtering; And, first sound channel signal of the described low-pass filtering of addition and high frequency first sound channel signal that is delayed, and second sound channel signal of the described low-pass filtering of addition and the described high frequency that is delayed second sound channel signal.

Above-mentioned and/or other aspects of the present invention can also realize by a kind of stereo production method is provided, described stereo production method is used for the reproduction of multichannel audio input signal is exported as dual track, described method comprises: be reduced in the relevance between the binaural audio signal in the described multichannel audio input signal, and produce the sensation of reality; Described binaural audio signal is converted to virtual sound source in the precalculated position; And, adjust remaining multi-channel audio signal except described binaural audio signal according to the output level of described binaural audio signal and time delay, and the signal of being adjusted is output as binaural signal.

Above-mentioned and/or other aspects of the present invention can also realize by a kind of stereo production method is provided, described stereo production method is used for left back and right back generation virtual speaker the listener, and described method comprises: adjust the gain and the delay of L channel input signal; Adjust the gain and the delay of R channel input signal; Described L channel input signal of addition and described gain and postpone controlled right-channel signals to obtain first sum signal; Adjust the frequency characteristic of first sum signal and the result is outputed to left speaker; Described R channel input signal of addition and described gain and postpone controlled left channel signals to obtain second sum signal; And, adjust the frequency characteristic of described second sum signal and the result is outputed to right loud speaker.

Description of drawings

By below in conjunction with description of drawings embodiment, these and/or other aspect of the present invention will become apparent and is more readily understood, wherein:

Fig. 1 is the block diagram of the traditional stereo generation of expression system;

Fig. 2 is the block diagram that be used for by dual track reproduce the stereo generation device of multi-channel audio signal of expression according to the embodiment of the invention;

Fig. 3 is expression according to the virtual ring of the stereo generation device of Fig. 2 of the embodiment of the invention schematic diagram around filter cell;

Fig. 4 is expression according to the virtual ring of Fig. 3 of the embodiment of the invention schematic diagram around the pre-processing filter unit of filter cell;

Fig. 5 is expression according to the virtual ring of Fig. 3 of the another embodiment of the present invention schematic diagram around the pre-processing filter unit of filter cell;

Fig. 6 is expression according to the virtual ring of Fig. 3 of the embodiment of the invention detailed maps around the virtual speaker filter cell of filter cell;

Fig. 7 is the design block diagram of expression according to the virtual speaker filter cell of Fig. 6 of the embodiment of the invention;

Fig. 8 is the Approximate Design block diagram of expression according to the virtual speaker filter cell of Fig. 6 of the embodiment of the invention;

Fig. 9 is the block diagram of expression according to the virtual speaker filter cell of Fig. 6 of the embodiment of the invention;

Figure 10 is the approximate schematic diagram of expression according to the virtual speaker filter cell of Fig. 6 of another embodiment of the present invention;

Figure 11 is the block diagram of expression according to the virtual speaker filter cell of Fig. 6 of another embodiment of the present invention;

Figure 12 is expression according to the virtual ring of the stereo generation device of Fig. 2 of the another embodiment of the present invention block diagram around filter cell;

Figure 13 is expression according to the virtual ring of the stereo generation device of Fig. 2 of the another embodiment of the present invention block diagram around filter cell;

Figure 14 is the more detailed block diagram of expression according to the signal correction filter cell of the stereo generation device of Fig. 2 of the embodiment of the invention;

Figure 15 is the block diagram that be used for by dual track reproduce the stereo generation device of multi-channel audio signal of expression according to another embodiment of the present invention; And

Figure 16 is the more detailed block diagram of expression according to the signal correction filter cell of the stereo generation device of Figure 15 of the embodiment of the invention.

Embodiment

Describe the embodiment of the total design of the present invention now in detail, its example is expressed in the accompanying drawings.In whole accompanying drawings, similarly drawing reference numeral is represented similar elements.By described embodiment is described referring to accompanying drawing, so that explain the total design of the present invention.

Fig. 2 is the block diagram that be used for by dual track reproduce the stereo generation device of multi-channel audio signal of expression according to the embodiment of the invention.

The stereo generation device of representing among Fig. 2 comprises that multi-channel audio signal 100, virtual ring are around filter cell 200, signal correction filter cell 300, first addition unit 401, second addition unit 402, left channel loudspeaker 500 and right channel loudspeaker 600.

Multi-channel audio signal 100 comprises left channel signals (L), center channel signal (C), low-frequency effect sound channel signal (LFE), right-channel signals (R), left surround channel signal (Ls) and right surround channel signal (Rs).Though in the present embodiment 5.1 sound channels are illustrated as example, it will be apparent to those skilled in the art that present embodiment can be applied to other multi-channel signal, for example 6.1 sound channels and 7.1 sound channels.

Virtual ring has the input that is used at the left surround channel signal (Ls) and the right surround channel signal (Rs) of described multi-channel audio signal around filter cell 200.

Virtual ring is reduced in the relevance of right and left rings between sound channel signal Ls and Rs of input when left back of the listener and right back produces reality sensation and virtual sound source around filter cell 200.Referring now to Fig. 3～7, describe this operation in detail.

Signal correction filter cell 300 has the input of the left channel signals (L), center channel signal (C), low-frequency effect sound channel signal (LFE) and the right-channel signals (R) that are used at described multi-channel audio signal.

(Ls, in output Rs), output gain changes and the generation time delay around sound channel signal in the right and left rings of exporting around filter cell 200 by virtual ring.Signal correction filter cell 300 can (Ls, output gain Rs) and time delay be adjusted the gain and the time delay of left channel signals (L), center channel signal (C), low-frequency effect sound channel signal (LFE), right-channel signals (R) around sound channel signal according to described right and left rings.

401 additions of first addition unit are from the left channels of sound signal of virtual ring around filter cell 200 and 300 outputs of signal correction filter cell, and 402 additions of second addition unit are from the right side sound channel signal of virtual ring around filter cell 200 and 300 outputs of signal correction filter cell.Then, the left-side signal after left channel loudspeaker 500 output additions, the right-side signal after right channel loudspeaker 600 output additions.

As mentioned above, if input signal is 6.1 channel audio signal, then comprises the back surround channel, and have described 5.1 sound channels.In this case, can in described stereo generation device, comprise with virtual ring around identical another virtual ring of filter cell 200 around filter, and back surround channel audio signal can be divided into two parts and be input to described additional virtual ring around filter.

If input signal is 7.1 channel audio signal, then comprise two after surround channel, and have described 5.1 sound channels.In this case, can in described device, comprise with described virtual ring around identical another virtual ring of filter cell 200 around filter, and to described additional virtual ring surround channel audio signal after filter is imported described two.

Fig. 3 is expression according to the virtual ring of the stereo generation device of Fig. 2 of the embodiment of the invention schematic diagram around filter cell 200 (not marking in Fig. 3).

Virtual ring comprises pre-processing filter unit 220 and virtual speaker filter cell 280 around filter cell 200.

Pre-processing filter unit 220 is reduced in the relevance between the right surround channel signal (Rs) of the left surround channel signal (Ls) of input and input, so that can improve surround channel localization of sound and actual sensation.

When the relevance of right and left rings between several road signal Ls and Rs is high, do not produce audiovideo in listener's left and right sides rear portion side, be used as mirage phantom (phantom) audiovideo but alternatively produce audiovideo in listener's central rear.And, obscure owing to front/rear, therefore described audiovideo sounds as starting from listener's front side, makes thus to be difficult to the perception surrounding effect.

Therefore, pre-processing filter unit 220 is reduced in right and left rings (Ls, the Rs) relevance between, and produce the reality sensation is so that can produce the surround channel effect of nature around sound channel signal.To illustrate in greater detail pre-processing filter unit 220 referring to Figure 4 and 5.

Virtual speaker filter cell 280 receives the signal from 220 outputs of pre-processing filter unit, and virtual sound source is arranged in listener's left back and right back portion, so that can produce stereo perception.To illustrate in greater detail virtual speaker filter cell 280 referring to Fig. 6 and 7.

Fig. 4 is expression according to the virtual ring of Fig. 3 of the embodiment of the invention schematic diagram around the pre-processing filter unit 220 (not having mark in Fig. 4) of filter cell (being the drawing reference numeral 200 among Fig. 2).

By using asymmetrical each other a plurality of delay cells, a plurality of gain unit and a plurality of addition unit to realize pre-processing filter unit 220.

That is, pre-processing filter unit 220 comprises that first delay cell 221, second delay cell 222, the 3rd delay cell 223, the 4th delay cell 224, first gain unit 225, second gain unit 226, first addition unit 227, second addition unit 228, first filter 229, second filter 230, the 3rd filter 231, the 4th filter 232, the 5th delay cell 233, the 6th delay cell 234, the 3rd gain unit 235, the 4th gain unit 236, third phase add unit 237 and the 4th addition unit 238.Pre-processing filter unit 220 also can comprise the 5th gain unit 239 and the 6th gain unit 240.

First delay cell 221 postpones preset time (i.e. first scheduled time) with left surround channel signal Ls.In the present embodiment, first delay cell 221 can be Z by having ^-mLLThe delay filter of transfer function realize.

Second delay cell 222 postpones preset time (i.e. second scheduled time) with right surround channel signal Rs.In the present embodiment, second delay cell 222 can be Z by having ^-mRRThe delay filter of transfer function realize.

First delay cell 221 and second delay cell 222 are asymmetric each other, and promptly described scheduled delay differs from one another.In other words, first scheduled time is different with second scheduled time.

The 3rd delay cell 223 is with left surround channel signal Ls delay scheduled time (i.e. the 3rd scheduled time).In the present embodiment, the 3rd delay cell 223 can be Z by having ^-mLRThe delay filter of transfer function realize.

The 4th delay cell 224 is with right surround channel signal Rs delay scheduled time (i.e. the 4th scheduled time).In the present embodiment, the 4th delay cell 224 can be by having Z ^-mRLThe delay filter of transfer function is realized.

The 3rd delay cell 223 and the 4th delay cell 224 are asymmetric each other, and promptly described scheduled delay differs from one another.In other words, the 3rd scheduled time is different with the 4th scheduled time.

First gain unit 225 changes the output gain of the 3rd delay cell 223, and second gain unit 226 changes the output gain of the 4th delay cell 224.

The output of second addition unit 228 additions, first delay cell 221 and second gain unit 226.The output of first addition unit 227 additions, second delay cell 222 and first gain unit 225.

Here, the output gain of first gain unit 225 and second gain unit 226 left surround channel signal Ls that will be delayed respectively and the right surround channel signal Rs that is delayed reduces predetermined amplitude.These first and second gain units 225 and 226 prevent the mixing of the audio signal of described two sound channels.

The output signal of first filter, 229 filtering, second addition unit 228, the output signal of second filter, 230 filtering, first addition unit 227.First and second filters 229 and 230 output signal are imported into virtual speaker filter cell 280 (referring to Fig. 3).As mentioned above, the 5th and the 6th gain unit 239 and 240 can gain respectively and adjusts the output signal of first and second filters 229 and 230.But the 5th and the 6th gain unit 239 and 240 does not need must be included in the pre-processing filter unit 220.The output signal of first and second filters 229 and the 230 or the 5th and the 6th gain unit 239 and 240 has the relevance of reduction betwixt.

The 5th delay cell 233 is with the output signal delay scheduled time (i.e. the 5th scheduled time) of the first and the 3rd filter 229 and 231.In the present embodiment, the 5th delay cell 233 can be by having Z ^-mLLsThe delay filter of transfer function is realized.

The 6th delay cell 234 is with the output signal delay scheduled time (i.e. the 6th scheduled time) of the second and the 4th filter 230 and 232.In the present embodiment, the 6th delay cell 234 can be Z by having ^-mRRsThe delay filter of transfer function realize.The 5th delay cell 233 and the 6th delay cell 234 other side are asymmetric, and promptly described scheduled delay differs from one another.In other words, the 5th scheduled time and the 6th scheduled time differ from one another.

According to current embodiment of the present invention, first to the 4th filter 229 to 232 can be a low pass filter.

The 3rd gain unit 235 changes the output gain of the 5th delay cell 233, and the 4th gain unit 236 changes the output gain of the 6th delay cell 234.

Third phase adds the output signal and the left surround channel signal (Ls) of unit 237 additions the 3rd gain unit 235, the output signal and the right surround channel signal (Rs) of the 4th addition unit 238 additions the 4th gain unit 236.

Fig. 5 is expression according to the virtual ring of Fig. 3 of the another embodiment of the present invention schematic diagram around the pre-processing filter unit 220 of filter cell (being the drawing reference numeral 200 among Fig. 2).

The pre-processing filter unit 220 of Fig. 5 has those the similar features with the pre-processing filter unit 220 of Fig. 4.But the pre-processing filter unit 220 of Fig. 5 can be applied to the full range band filter of artificial echo (artificial reverberator) so that the reverberation of artificially reproduction space (reverberation) characteristic produces more natural wide stereophonic effect by use.And described full range band filter has the characteristic that postpones predetermined frequency component, and by using this characteristic, makes it possible to produce stereophonic effect with respect to monophonic signal.

In the pre-processing filter unit 220 shown in Fig. 5, each of left surround channel signal (Ls) and right surround channel signal (Rs) is applied to two full range band filters.That is, left surround channel signal (Ls) is converted into a plurality of reverberation sound by two left full range band filters of series connection.And right surround channel signal (Rs) is converted into a plurality of reverberation sound by two right full range band filters of series connection.Therefore, can use reverberation sound to lower relevance between left surround channel signal Ls and right surround channel signal Rs.

Explanation at first, now is used for the processing of full range band filtering left side surround channel signal (Ls).In left full range band filter, first to the 4th adder 255,253,260 and 258 is connected respectively to the input and the output of first and second delay cells 251 and 256.Input signal is fed forward to the second and the 4th adder 253 and 258 of using attenuation coefficient (GL) to form by the first and the 3rd multiplier 262 and 267 respectively.(the second and the 4th multiplier 254 and 259 that GL) forms is fed back to the first and the 3rd adder 255 and 260 by the use attenuation coefficient respectively for the second and the 4th adder 253 and 258 addition output.

The structure of described two right full range band filters can be with described two left full range band filters of left surround channel signal Ls identical.For illustrative purposes, in Fig. 5, described two right full range band filters are disposed under described two left full range band filters.Described two right full range band filters can comprise the 5th to the 8th adder 265,263,270 and 268, third and fourth delay cell 261 and the 266, the 4th to the 8th multiplier 272,264,267 and 269.

Here, when input signal was monophonic signal, in order to make described monophonic signal become stereophonic signal, the length of delay with the 4th delay cell 251,256,261 and 266 differently was set to L0, L1, R0 and R1 respectively.The length of delay of two delay cells of connecting in each sound channel has the L0 of relation＞L1, R0＞R1 or L0＜L1, R0＜R1.This is by the asymmetric reduction relevance that maximizes, as in the pre-processing filter unit 220 of above-mentioned Fig. 4.

And the yield value of the multiplier of filter can have identical value, and ought can differently be provided with in case of necessity.For example, as shown in Figure 5, first multiplier 262 and second multiplier 254 can have respectively value GL and-GL.And in order to prevent out-phase (out-of-phase) phenomenon, described attenuation coefficient (GL and GR) can have identical symbol or opposite symbol (sign), and still, the gain of two filters of Lian Jieing explicitly is made into and has identical symbol.

Fig. 6 is expression according to the virtual ring of Fig. 3 of the embodiment of the invention detailed maps around the virtual speaker filter cell 280 of filter cell (being the drawing reference numeral 200 among Fig. 2).

The virtual speaker filter cell shown in Fig. 6 280 will be as above (Ls Rs) is converted to respectively virtual sound source in left back of listener and right back portion around sound channel signal referring to the right and left rings of described 220 outputs from the pre-processing filter unit of Figure 4 and 5.

Virtual speaker filter cell 280 has such structure, and wherein, (Ls is Rs) by four finite impulse response (FIR)s (FIR) filter K around sound channel signal for the right and left rings of 220 outputs from the pre-processing filter unit ₁₁, K ₁₂, K ₂₁, K ₂₂Convolution and addition.

Use FIR filter K ₁₁Come convolution left side surround channel signal (Ls), use FIR filter K ₁₂Come the right surround channel signal (Rs) of convolution.Described two are added and are produced as the L channel output signal then by the signal of convolution.Also use FIR filter K ₂₁Come convolution left side surround channel signal (Ls), and also use FIR filter K ₂₂Come the right surround channel signal (Rs) of convolution.Then two signals by convolution are added and produce as the R channel output signal.These left and right acoustic channels output signals are added to the output signal of signal correction filter cell 300 described later (referring to Fig. 1) respectively, and produce the last output signal of dual track.

Fig. 7 is the design block diagram of expression according to the virtual speaker filter cell 280 of Fig. 6 of the embodiment of the invention.

At first, virtual speaker filter cell 280 comprises: ears composite filter B ₁₁, B ₁₂, B ₂₁, B ₂₂, they are implemented as head related transfer function (HRTF) matrix between virtual sound source and virtual listening person; And cross-talk is eliminated filter C ₁₁, C ₁₂, C ₂₁, C ₂₂, they are implemented as the HRTF inverse of a matrix matrix between virtual listening person and dual track outgoing position.

Described ears composite filter B ₁₁, B ₁₂, B ₂₁, B ₂₂Be designed as follows.By using HRTF to realize described ears composite filter B as the transmission sound function between sound source and virtual listening person's (perhaps actual listener) eardrum ₁₁, B ₁₂, B ₂₁, B ₂₂

Described HRTF comprises the information of the characteristic in the space that is used to indicate transfer voice, comprises the shape of the auricle of time difference (ITD) and listener between grade differential between the sense of hearing (ILD), the sense of hearing.Particularly, described HRTF comprises the information about auricle, and it has crucial influence to sound localization up and down.Because it is very difficult having the modeling of the auricle of complicated shape, therefore obtain described HRTF by the measurement of using artificial head usually.Usually, with respect to the preceding center of artificial head, between 90 degree and 110 degree, arrange circulating loudspeaker.Therefore, for virtual speaker being positioned between 90 degree and 110 degree, between the left side at the preceding center of described artificial head and right 90 degree and 110 degree, measure HRTF.

Suppose corresponding to being respectively B at the sound source between the left side 90 of described artificial head degree and 110 degree and the left ear of described artificial head and the HRTF in the path between the auris dextra ₁₁And B ₂₁, and corresponding to being respectively B at the sound source between the right side 90 of described artificial head degree and 110 degree and the left ear of described artificial head and the HRTF in the path between the auris dextra ₁₂And B ₂₂

If export the synthetic output signal of described ears by earphone, then the listener feels the audiovideo that produces between 90 degree on the left side at preceding center and right side and 110 degree.When by the headphone reproduction signal, the synthetic performance of showing the best of described ears.

But, if come reproducing signal by two loud speakers, then between two loud speakers and two ears cross-talk takes place, make the positioning performance variation.That is,, occur in the cross-talk phenomenon between the dual track though only hear described L channel sound and only hear described right channel sound at auris dextra at left ear.As a result, also in auris dextra, hear described L channel sound, and also in left ear, hear described right channel sound.Therefore, the sensation variation of location is not to such an extent as to be positioned at audiovideo on the accurate point.

Therefore, cross-talk cancellation filter cell C ₁₁, C ₁₂, C ₂₁And C ₂₂Be designed to eliminate cross-talk.Design hereto should be measured the HRTF between listener (it is corresponding to virtual listening person) and two loud speakers.

Suppose that at the loud speaker in the precalculated position (it can be measured by artificial head) that is positioned at the listener left side and the left ear and the HRTF between the auris dextra of described artificial head be respectively H ₁₁And H ₂₁, and be respectively H at the loud speaker in the precalculated position on the right side of described artificial head and the left ear and the HRTF between the auris dextra of described artificial head ₁₂And H ₂₂, then cross-talk is eliminated the inverse matrix that electric-wave filter matrix (C (z)) is designed to HRTF, as equation 1 below:

$\left[\begin{array}{cc}{C}_{11}\left(z\right)& C_{12}\left(z\right)\\ C_{21}\left(z\right)& C_{22}\left(z\right)\end{array}\right]={\left[\begin{array}{cc}{H}_{11}\left(z\right)& H_{12}\left(z\right)\\ H_{21}\left(z\right)& H_{22}\left(z\right)\end{array}\right]}^{-1}---\left(1\right)$

Described ears composite filter matrix is positioned at virtual speaker the position of left and right sides circulating loudspeaker.Described cross-talk is eliminated the cross-talk of electric-wave filter matrix elimination between described two loud speakers (being virtual speaker) and listener's two ears.Therefore, by calculate the matrix K (z) of virtual speaker filter cell 280 as two electric-wave filter matrix that multiply each other in the equation 2 below

$\left[\begin{array}{cc}{K}_{11}\left(z\right)& K_{12}\left(z\right)\\ K_{21}\left(z\right)& K_{22}\left(z\right)\end{array}\right]=\left[\begin{array}{cc}{C}_{11}\left(z\right)& C_{12}\left(z\right)\\ C_{21}\left(z\right)& C_{22}\left(z\right)\end{array}\right]\left[\begin{array}{cc}{B}_{11}\left(z\right)& B_{12}\left(z\right)\\ B_{21}\left(z\right)& B_{22}\left(z\right)\end{array}\right]---\left(2\right)$

As finding out in Fig. 6, virtual speaker filter cell 280 comprises four filters, and carries out convolution operation four times.Therefore, virtual speaker filter cell 280 needs a large amount of calculating when high on the rank of filter.

The current trend of digital media products is to comprise the stereophonic loud-speaker system that is mounted.In the portable equipment such as portable media player (PMP) and PDA(Personal Digital Assistant) and television set, two loud speakers are arranged close to each otherly.

Therefore, arranged to such an extent that compare listener's distance when nearer, K each other when two loud speakers ₁₁(z) and K ₁₂(z) owing to cross-talk elimination characteristic has high relevance, and K ₂₁(z) and K ₂₂(z) also has high relevance.

Therefore, when two loud speakers are asymmetricly arranged with respect to the listener, can as following expression 3, suppose the virtual speaker filter coefficient:

$K_{12}\left(z\right)\≅{\text{\α}}_1{z}^{{-\mathrm{\β}}_1}{K}_{11}\left(z\right)\text{,}$ $K_{21}\left(z\right)\text{\≅}{\mathrm{\α}}_2{z}^{{-\mathrm{\β}}_2}{K}_{22}\left(z\right)---\left(3\right)$

Here, yield value (α) is differential between two HRTF, and length of delay (β) is that the delay between two HRTF is poor.From obtain differential (α) between two HRTF in the difference between the maximum of the impulse response of two HRTF between loud speaker and listener's the ears or the difference between root mean square (RMS) value.When the cross-correlation (cross-correlation) of the impulse response of two HRTF between loud speaker and ears when function becomes maximum, obtain delay poor (β) between two HRTF from this time.In another embodiment, can be by determining described yield value (α) with respect to the difference between the maximum of impulse response of two filters of lattice shape (lattice) structure of design in advance, and length of delay (β) can be defined as time when the cross-correlation function with respect to the impulse response of two filters of the lattice structure that designs in advance becomes maximum.

When using equation 3, virtual speaker filter cell 280 (see figure 3)s can be expressed as the block diagram of Fig. 8.In addition, can once more the block diagram of Fig. 8 be expressed as the block diagram of Fig. 9.

Fig. 9 is the block diagram of expression according to virtual speaker filter cell 280 (see figure 3)s of Fig. 6 of the embodiment of the invention.Referring to Fig. 9, the left channel signals (Y that first gain unit 412 uses the first predetermined gain value adjustment to be transfused to _L) gain.

Right-channel signals (the Y that second gain unit 416 uses the second predetermined gain value adjustment to be transfused to _R) gain.

First delay cell 414 uses the first predetermined delay value to postpone the left channel signals (Y that is adjusted by gain in first gain unit 412 _L).

Second delay cell 418 uses the second predetermined delay value to postpone the right-channel signals (Y that is adjusted by gain in second gain unit 416 _R).

Left channel signals (the Y that the first addition unit 419-1 addition is transfused to _L) and the right-channel signals (Y that gained and postpone to adjust by second gain unit 416 and second delay cell 418 _R).

Right-channel signals (the Y that the second addition unit 419-2 addition is transfused to _R) and the left channel signals (Y that gained and postpone to adjust by first gain unit 412 and first delay cell 414 _L).

First filter cell 422 has the contrary HRTF form as the HRTF of the transmission sound function between loud speaker and listener's ears, and is adjusted at the frequency characteristic of the signal that mixes among the first addition unit 419-1.Output signal (the S of first filter cell 422 _L) be output to left speaker.

Second filter cell 424 has the contrary HRTF form as the HRTF of the transmission sound function between loud speaker and listener's ears, and is adjusted at the frequency characteristic of the signal that mixes among the second addition unit 419-2.Output signal (the S of second filter cell 424 _R) be output to right loud speaker.

Therefore, the virtual speaker filter cell 280 of Fig. 9 comprises two gain units 412 and 416, two delay cell 414 and 418 and two filters 422 and 424.

The result, when carrying out four convolution with respect to four filters in the structure at the virtual speaker filter cell 280 of Fig. 6 and 7, only carry out convolution twice with respect to two filters in the virtual speaker filter cell 280 of the present embodiment of Fig. 8 and 9, make the size that can reduce amount of calculation and memory.

In addition, when two loud speakers were arranged symmetrically with respect to the listener, the virtual speaker electric-wave filter matrix became K ₁₁(z)=K ₂₂(z) and K ₂₁(z)=K ₁₂(z).Therefore, the virtual speaker electric-wave filter matrix can be expressed as following expression 4:

$K_2\left(z\right)\≅{\mathrm{\αz}}^{-\mathrm{\β}}{K}_1\left(z\right)---\left(4\right)$

By using expression formula 4, can be with the block diagram of virtual speaker expression matrix in Figure 10, representing.Figure 10 is the approximate schematic diagram according to virtual speaker filter cell 280 (see figure 3)s of Fig. 6 of another embodiment of the present invention.To come calculated gains value (α) and length of delay (β) with mode identical in the virtual speaker filter cell 280 of Fig. 9.The block diagram of Figure 10 can be expressed as again the block diagram of Figure 11.Figure 11 is the block diagram of expression according to virtual speaker filter cell 280 (see figure 3)s of Fig. 6 of another embodiment of the present invention.

Referring to Figure 11, first and second filter cells 512 and 514 are adjusted the frequency characteristic of the left and right sound track signals of input respectively.

First and second gain units 522 and 526 use the predetermined gain value to adjust the gain of the output signal of first and second filter cells 512 and 514 respectively.

First and second delay cells 524 and 528 use predetermined length of delay to postpone the signal of being adjusted by gain respectively in first and second gain units 522 and 526.

The output signal that the output signal of the first addition unit 529-1 addition, first filter cell 512 and the quilt of second delay cell 528 gain and postpone to adjust.

The output signal that the output signal of the second addition unit 529-2 addition, second filter cell 514 and the quilt of first delay cell 524 gain and postpone to adjust.

The virtual ring of Figure 12 and Figure 13 presentation graphs 2 is around other embodiment of filter cell 200.

Usually, be low-frequency band for the influential frequency band in the location of virtual sound source.And in having short wavelength's very high frequency band, cross-talk is eliminated the degradation of filter, and can not remove the cross-talk component.Therefore, in filter cell 200, the signal processing of only carrying out low-frequency band is as follows at the virtual ring of Fig. 2.That is, by using low pass filter and high pass filter that input signal is divided into two frequency bands.High-frequency signal by high pass filter is by signal processing, and the signal by low pass filter is by a minute sample.Divide the signals sampling frequency of sample (decimated) to be lowered.Therefore, the delay filter coefficient of pre-processing filter unit 220 is lowered, and the FIR exponent number of virtual speaker filter cell 280 is lowered, and makes to reduce the amount of calculation and the amount of memory of virtual ring around filter 200 widely.

Figure 12 is expression according to the virtual ring of Fig. 2 of the another embodiment of the present invention block diagram around filter cell 200.Referring to Figure 12, (Ls is Rs) by the sensation of pre-processing filter unit 220 with reduction relevance and generation reality for first and second sound channel signals.Each of first and second sound channel signals of pretreated filtering by high pass filter (HPF) 512 and 518 and low pass filter (LPF) 514 and 516 be divided into high frequency band and low-frequency band.At this moment, pass through two LPF 514 and 516 low band signal of exporting by minute sample unit 524 and about 526 minutes samples respectively, so that reduce sample frequency.And, by delay cell 522 and 528, will postpone preset time by the high-frequency band signals of two HPF 512 and 518 outputs, so that the path of high-frequency band signals and low band signal is synchronous.Therefore, each is output as dual track virtual sound source in the precalculated position by the signal of minute sample by virtual speaker filter cell 280.Here, divide the signal of sample owing to hanging down the FIR filter order that sample frequency reduces virtual speaker filter cell 280.Be used for interpolation from the binaural signal of virtual speaker filter cell 280 output by interpolater 542 and 544.At this, interpolater 542 and 544 will be adjusted into original sample frequency by the sample frequency that minute sample reduces.The signal that is interpolated is low pass filtering by LPF 552 and 554 then.

At last, first and second adders 562 and 564 additions are respectively from first and second sound channel signals that are low pass filtering of LPF 552 and 554 outputs and respectively from HPF 512 and 518 outputs and high frequency first and second sound channel signals that postpone delay cell 522 and 528.

At this, the filtering of full range band signal is carried out in pre-processing filter unit 220.

Therefore, produce space sense for the full range band signal.And, because only locate virtual sound source, therefore many speed (multi-rate) of only handling low band signal can be handled and be applied to virtual speaker filter cell 280 with respect to low band signal.

Can use any one of embodiment of Figure 4 and 5 to realize pre-processing filter unit 220, and can use Fig. 6,9 and 11 embodiment any one realize virtual speaker filter cell 280.

Figure 13 is expression according to the virtual ring of Fig. 2 of the another embodiment of the present invention block diagram around filter cell 200.Referring to Figure 13, respectively by HPF 612 and 618 and LPF 614 and 616 first and second sound channel signals are divided into high-frequency band signals and low band signal.Respectively by minute sample unit 624 and about 626 minutes samples each by two LPF 614 and 616 low band signal of exporting.And, be delayed the scheduled time by two HPF 612 and 618 high-frequency band signals of exporting, so that the path of high-frequency band signals and low band signal is synchronous.In the signal of sample branch, reduce relevance by pre-processing filter unit 220 and virtual speaker filter cell 280, and low band signal is exported as the binaural signal that is converted into the virtual sound source with precalculated position.

Pass through interpolater 642 and 644 interpolations from the binaural signal of virtual speaker filter cell 280 outputs.Described interpolated signal is by LPF 652 and 654 low-pass filtering.

At last, first and second sound channel signals that are low pass filtering of first and second adders 662 and 664 additions with from HPF 612 and 618 outputs and high frequency first and second sound channel signals that delay cell 622 and 628, postpone.

Can use any one of embodiment of Figure 4 and 5 to realize pre-processing filter unit 220, and can use Fig. 6,9 and 11 embodiment any one realize virtual speaker filter cell 280.

Figure 14 is the more detailed block diagram of expression according to the signal correction filter cell 300 of Fig. 2 of the embodiment of the invention.

The signal correction filter cell 300 of Figure 14 comprises and has predetermined gain value (Ga, Gb, Gc, gain unit 710,720,730 Gd) and 740 and have a predetermined delay value (Z ^-Δ) delay cell 715,725,735 and 745.

Change the output gain of left channel signals (L) by gain unit 710, and postpone left channel signals (L) by delay cell 715.

Change the output gain of center channel signal (C) by gain unit 720, and postpone center channel signal (C) by delay cell 725.

Change the output gain of LFE sound channel signal (LFE) by gain unit 730, and postpone LFE sound channel signal (LFE) by delay cell 735.

Change the output gain of right-channel signals (R) by gain unit 740, and postpone right-channel signals (R) by delay cell 745.

The first addition unit 700-1 addition is from the signal of delay cell 715,725 and 735 outputs.

The second addition unit 700-2 addition is from the signal of delay cell 725,735 and 745 outputs.

If right and left rings around sound channel signal by virtual ring around filter cell 200, then right and left rings is around those changes from the primary signal of the stereo generation device that is imported into Fig. 2 of the output gain of sound channel signal and time delay.Therefore, according to the characteristic of virtual ring, adjust the output gain and the time delay of L channel (L), center channel (C), LFE sound channel (LFE) and R channel (R) signal around filter cell 200.At this, " according to the characteristic of virtual ring " around filter do not represent by the change in input signal determine right and left rings around the output gain of sound channel signal and the time Yanzhong change.On the contrary, this expression by virtual ring around the element of filter cell 200 determine the output gain that causes by signal correction filter cell 300 and the time Yanzhong change.

At this, by virtual ring relatively around the RMS value of the input signal of filter cell 200 and output signal determine gain unit 710,720,730 and 740 yield value (Ga, Gb, Gc, Gd).By using virtual ring around the impulse response of filter cell 200 or postpone to obtain the length of delay (Z of delay cell 715,725,735 and 745 by the use group ^-Δ).For example, can be according to the FIR filter (K of previous embodiment ₁₁) group postpone to determine time delay value.

Figure 15 is the block diagram that be used for by dual track reproduce the stereo generation device of multi-channel audio signal of expression according to another embodiment of the present invention.

Stereo generation device shown in Figure 15 comprises that multichannel audio input signal 800, signal correction filter cell 810, wide stereo generation unit 820, virtual ring are around filter cell 830, first and second addition unit 850 and 860, left channel loudspeaker 890-1 and right channel loudspeaker 890-2.

Described multi-channel audio signal 800 comprises left channel signals (L), center channel signal (C), low-frequency effect sound channel signal (LFE), right-channel signals (R), left surround channel signal (Ls) and right surround channel signal (Rs).

Virtual ring can be similar to the virtual ring of Fig. 2 around filter cell 200 around filter cell 830.

Wide stereo generation unit 820 reception left and right sound track signals (L, input R), and produce the stereophonic signal of widening.Wide stereo generation unit 820 comprises: widen filter, be used to carry out the convolution of the synthetic and crosstalk canceller of a left side/right ears; And panorama (panorama) filter, be used to carry out described convolution of widening the direct filter (direct filter) of filter and a left side/right side.The described filter of widening produces left and right sound track signals as at an arbitrary position virtual sound source according to the HRTF that measures in the precalculated position (L R), and removes the cross-talk of virtual sound source according to the filter coefficient of having used HRTF.About directly filter be adjusted at the sound-source signal of stereo channels and remove between the virtual sound source of cross-talk such as gain and the characteristics of signals postponing.

Signal correction filter cell 810 is received in the center channel (C) in the multichannel audio input signal 800 and the signal of LFE sound channel.

Change thus the right and left rings exported around filter cell 830 by virtual ring around sound channel signal (Ls, Rs) and left and right sound track signals (L, output gain R) and the time delay exported by wide stereo generation unit 820.Signal correction filter cell 810 according to from virtual ring around the right and left rings of filter cell 830 output around sound channel signal (Ls, Rs) and the left and right sound track signals of leniently stereo generation unit 820 output (L, output gain R) and time delay are adjusted the gain and the time delay of center channel signal (C) and LFE sound channel signal (LFE).

850 additions of first addition unit are from the left channel signals of virtual ring around filter cell 830, signal correction filter cell 810 and 820 outputs of wide stereo generation unit.860 additions of second addition unit are from the right-channel signals of virtual ring around filter cell 830, signal correction filter cell 810 and 820 outputs of wide stereo generation unit.Then, by the left signal after the left channel loudspeaker 890-1 output addition, by the right signal after the right channel loudspeaker 890-2 output addition.

Figure 16 is the more detailed block diagram of expression according to the signal correction filter cell 810 of Figure 15 of the embodiment of the invention.

The signal correction filter cell 810 of Figure 15 comprises and has predetermined gain value (Ga, gain unit 910 Gb) and 920 and have a predetermined delay value (Z ^-Δ) delay cell 915 and 925.

Change the output gain of center channel signal (C) by gain unit 910, and in delay cell 915, postpone center channel signal (C).

Change the output gain of LFE sound channel signal (LFE) by gain unit 920, and in delay cell 925, postpone LFE sound channel signal (LFE).

The first addition unit 900-1 addition is from the signal of delay cell 915 and 925 outputs.The second addition unit 900-2 also addition from delay cell 915 and 925 output signals.

At this, by virtual ring relatively around the RMS value of the input signal of filter cell 830 and output signal determine gain unit 910 and 920 yield value (Ga, Gb).By using virtual ring around the impulse response of filter cell 830 or postpone to obtain the length of delay (Z of delay cell 915 and 925 by the use group ^-Δ).

Be understood that, though the ears with reference to listener and listener or virtual listening person have illustrated embodiments of the invention, can use the device of the embodiment of the invention to be created in stereo around the listening point of stereo generation system and/or virtual ring system for winding.Described listening point can refer to the position of listener optimum stereo effect, and this can use for example above-mentioned artificial head to be similar to.Therefore, when the device as said various embodiment moved, in fact the listener needn't be positioned at listening point.

The total design of the present invention also can be embodied as the computer-readable code on computer readable recording medium storing program for performing.Described computer readable recording medium storing program for performing be can store can be subsequently by any data storage device of the data of computer system reads.The example of described computer readable recording medium storing program for performing comprises read-only memory (ROM), random-access memory (ram), CD-ROM, tape, floppy disk, optical data storage spare and carrier wave (carrier waver) (such as the transfer of data by the internet).Described computer readable recording medium storing program for performing also can be distributed on the computer system of network-coupled, so that store and carry out described computer-readable code in the mode that distributes.And, can easily explain function program, code and the code segment of the design that is used to realize that the present invention is total by the skilled programmer in the technical field under the total design of the present invention.

Each embodiment of total according to the present invention described above design can use dual track to export and reproduce multi-channel audio signal, and by only using dual track output, can realize the stereo perception of multi-channel speaker system.

And, being associated around the channel audio input signal with right and left rings, left back and right back generation virtual speaker by the listener can provide stereo perception to the listener effectively.

And, even, also can improve described localization of sound, and can produce actual sound, so that can provide more improve stereo to the listener when when the relevance of right and left rings between the channel audio input signal is high.

Though illustrated and illustrated some embodiment of the design that the present invention is total, but those skilled in the art can understand, under the situation of principle that does not break away from the total design of the present invention and spirit, can change in these embodiments, scope of the present invention is limited in appended claim and their equivalents.

The cross reference that relates to the application

The application requires on December 13rd, 2005 at korean patent application 10-2005-0122433 number of Korea S Department of Intellectual Property application and the U.S. Provisional Application the 60/719th of application on September 22nd, 2005, No. 191 rights and interests, it openly is included as a whole at this as a reference by reference.

Claims (47)

1. a stereo generation device is used for the multichannel audio input signal is reproduced as the dual track output signal, and described device comprises:

The pre-processing filter unit is used to be reduced in the relevance between the binaural audio signal of multichannel audio input signal, and produces the sensation of reality;

The virtual speaker filter cell, the binaural audio signal that is used for exporting from described pre-processing filter unit is converted to the virtual sound source in the precalculated position;

The signal correction filter cell is used to proofread and correct at the remaining multichannel audio input signal except described dual-channel audio input signal and from the characteristics of signals between the described binaural audio signal of described virtual speaker filter cell output; And,

Addition unit, the multi-channel audio signal that is used for being added in mutually from described virtual speaker filter cell and the output of described signal correction filter cell will output to the signal of first sound channel, and is added in the signal that will output to second sound channel from the multi-channel audio signal of described virtual speaker filter cell and the output of described signal correction filter cell mutually.

2. according to the device of claim 1, wherein, described pre-processing filter unit comprises:

First delay cell was used for one period of the first channel audio signal delay control at described dual-channel audio input signal; And

Second delay cell was used for second period of rising tone channel audio signal delay at described dual-channel audio input signal.

3. according to the device of claim 2, wherein, described pre-processing filter unit comprises:

The 3rd delay cell was used for three periods of the first channel audio signal delay control at described dual-channel audio input signal;

The 4th delay cell was used for the 4th period of rising tone channel audio signal delay at described dual-channel audio input signal;

First gain unit is used to adjust the output gain of the 3rd delay cell;

Second gain unit is used to adjust the output gain of the 4th delay cell;

First addition unit is used for the output of addition first delay cell and the output of second gain unit; And

Second addition unit is used for the output of addition second delay cell and the output of first gain unit.

4. according to the device of claim 3, wherein, described pre-processing filter unit comprises:

First filter is used for the output signal of low-pass filtering first addition unit;

Second filter is used for the output signal of low-pass filtering second addition unit;

The 5th delay cell is used for five periods of output signal delay control with first filter;

The 6th delay cell is used for six periods of output signal delay control with second filter;

The 3rd gain unit is used to adjust the output gain of the 5th delay cell;

The 4th gain unit is used to adjust the output gain of the 6th delay cell;

Third phase adds the unit, is used for the output signal of addition first channel audio signal and the 3rd gain unit; And

The 4th addition unit is used for the output signal of addition second channel audio signal and the 4th gain unit.

5. according to the device of claim 4, wherein, described first to the 6th period differs from one another.

6. according to the device of claim 1, wherein, described virtual speaker filter cell comprises:

The ears synthesis unit is used for first channel audio signal and rising tone channel audio conversion of signals from the output of described pre-processing filter unit are the virtual sound source in the precalculated position; And

Crosstalk canceller is used to eliminate the cross-talk phenomenon from the signal of described ears synthesis unit output.

7. according to the device of claim 1, wherein, described virtual speaker filter cell comprises:

Delay cell is used for separately predetermined delay value and postpones first and rising tone channel audio input signal;

Gain unit is used for being adjusted at each output gain of described first and second channel audio input signals that described delay cell postpones;

First addition unit is used for described first channel audio input signal of addition and gain and postpones controlled second sound channel signal;

First filter cell is used to adjust the frequency characteristic from the signal of described first addition unit output;

Second addition unit is used for addition described rising tone channel audio input signal and gain and postpones controlled first sound channel signal; And

Second filter cell is used to adjust the frequency characteristic from the signal of described second addition unit output.

8. according to the device of claim 1, wherein, described virtual speaker filter cell comprises:

First and second filter cells are used to adjust the frequency characteristic of first and second sound channel signals;

Delay cell is used for predetermined separately length of delay and postpones the output signal of described first and second filter cells;

Gain unit is used for being adjusted at each output level of signal that described delay cell postpones;

First addition unit is used for the gain of the output signal of addition first filter cell and second filter cell and postpones controlled output signal; And

Second addition unit is used for the gain of the output signal of described second filter cell of addition and described first filter cell and postpones controlled output signal.

9. according to the device of claim 8, wherein, by with the impulse response separately that is associated at two head related transfer functions (HRTF) between loud speaker and the listener's ears between the maximum difference determine the gain of described gain unit.

10. according to the device of claim 8, wherein, the time when becoming maximum by the cross-correlation function when the impulse response that is associated with two HRTF between loud speaker and listener's ears is determined the delay of delay cell.

11., wherein, determine to gain by the difference between the maximum of the impulse response that is associated at two filters of lattice structure with design in advance according to the device of claim 8.

12. according to the device of claim 8, wherein, the time the when cross-correlation function by the impulse response that is associated when two filters of lattice structure with design in advance becomes maximum is determined to postpone.

13. according to the device of claim 1, wherein, described signal correction filter cell comprises:

Gain unit is used to adjust the gain of the multichannel audio input signal except described dual-channel audio input signal; And

Delay cell is used for the multichannel audio input signal delay scheduled time except described dual-channel audio input signal.

14. according to the device of claim 13, wherein, by the relatively output signal of virtual speaker filter cell and the gain that described two channel audio input signals are determined described gain unit.

15. according to the device of claim 13, wherein, root mean square (RMS) value of the output signal by more described virtual speaker filter cell and the RMS value of described two channel audio input signals are determined the gain of described gain unit.

16., wherein, postpone to determine the described scheduled time according to the group of crosstalk canceller according to the device of claim 13.

17. according to the device of claim 1, wherein, described addition unit comprises:

First addition unit is used for being added in mutually the signal that will output to first sound channel from the multichannel audio input signal of described virtual speaker filter cell and the output of described signal correction filter cell; And

Second addition unit is used for being added in mutually the signal that will output to second sound channel from the multichannel audio input signal of described virtual speaker filter cell and the output of described signal correction filter cell.

18. a stereo generation device is used for the multichannel audio input signal is reproduced as binaural audio signal output, described device comprises:

The pre-processing filter unit is used for postponing in groups the predetermined frequency component of the binaural audio signal selected at described multichannel audio input signal;

The virtual speaker filter cell is used for selected binaural audio signal from described pre-processing filter unit output is converted to virtual sound source in the precalculated position;

The signal correction filter cell is used to proofread and correct in the residue multi-channel audio signal except selected binaural audio signal and selected output level and time delay between the binaural audio signal of described virtual speaker filter cell output; And,

Addition unit, the multi-channel audio signal that is used for being added in mutually from described virtual speaker filter cell and the output of described signal correction filter cell will output to the signal of first sound channel, and is added in the signal that will output to second sound channel from the multi-channel audio signal of described virtual speaker filter cell and the output of described signal correction filter cell mutually.

19. according to the device of claim 18, wherein, in described pre-processing filter unit, with each connect explicitly " n " individual full band pass filter of described first and second sound channels.

20. according to the device of claim 19, wherein, each of described full band pass filter comprises:

Delay cell is used for the delayed audio signal scheduled time with input;

First gain unit is used to adjust the gain of input audio signal;

First addition unit is used for the output of described first gain unit of addition and the output of described delay cell;

Second gain unit is used to adjust the output gain of described first addition unit; And

Second addition unit is used for the output signal and the described input audio signal of described second gain unit of addition.

21. according to the device of claim 20, wherein, the gain of described first gain unit and described second gain unit equates, but has opposite symbol.

22. a stereo generation device is used for carrying out two convolution with matrix structure of pre-sizing by calculating the ears synthesizer and the crosstalk canceller that are associated with binaural signal in advance, described device comprises:

Delay cell is used for predetermined separately length of delay and postpones first and second channel input signal;

Gain unit is used for being adjusted at each output level of first and second channel input signals that described delay cell postpones;

First addition unit is used for described first channel input signal of addition and described gain and postpones controlled second sound channel signal;

First filter cell is used to adjust the frequency characteristic from the signal of described first addition unit output;

Second addition unit is used for described second channel input signal of addition and described gain and postpones controlled first sound channel signal; And,

Second filter cell is used to adjust the frequency characteristic from the signal of described second addition unit output.

23. a stereo generation device is used for the multichannel audio input signal is reproduced as the dual track output signal, described device comprises:

Virtual ring is used to be reduced in the relevance between two surround channel audio signals of described multichannel audio input signal around filter cell, and described two surround channel audio signals are converted to virtual sound source in the precalculated position;

Wide stereo generation unit is used for synthesizing two stereophonic signals that preceding channel audio signal is used as widening that are created in described multichannel audio input signal with crosstalk canceller by the convolution ears; And,

The signal correction filter cell is used to proofread and correct at the remaining multichannel audio input signal except two surround channel signals and described two preceding channel audio signal with from output level and the time delay of described virtual ring between the channel audio signal of filter cell and described wide stereo generation unit output.

24. the device according to claim 23 also comprises:

Addition unit, be used for being added in mutually the signal that will export by first sound channel from the multi-channel audio signal of virtual speaker filter cell, described signal correction filter cell and described wide stereo generation unit output, and addition will be by the signal of second sound channel output.

25. according to the device of claim 23, wherein, described signal correction filter cell comprises:

Gain unit is used to adjust the gain of the multi-channel audio signal except described two surround channel audio signals and described two preceding channel audio signal; And

Delay cell is used for the multi-channel audio signal delay scheduled time except described two surround channel audio signals and described two preceding channel audio signal.

26. device according to claim 25, wherein, by being compared with described two surround channel audio input signals and described two preceding sound channel signals, the output signal of described virtual speaker filter cell and described wide stereo generation unit determines the gain of gain unit.

27., wherein, determine the gain of gain unit by the RMS value of the output signal of described virtual speaker filter cell and described wide stereo generation unit is compared with the RMS value of remaining channel audio signal according to the device of claim 25.

28. a stereo generation device comprises:

First filter cell is used for being received in the surround audio signal of at least 5 input audio signals, and produces with respect to the virtual sound source of listening point in the precalculated position;

Second filter cell is used for being received in the remaining audio signal of described at least 5 input audio signals, and compensates delay and the gain inequality that is caused in described surround audio signal around filter cell by described virtual ring; And,

Output unit is used to make up first those surround audio signals selected and remaining audio signal producing left output signal, and makes up second those surround audio signals of selecting and remaining audio signal to produce right output signal.

29. the stereo generation device according to claim 28 also comprises:

Left speaker is used to export left output signal; And

Right loud speaker is used to export right output signal.

30. stereo generation device according to claim 29, wherein, described left speaker and described right loud speaker are arranged to each other at a distance of first preset distance, and described left and right sides loud speaker be arranged to described listening point at a distance of second preset distance, make described second preset distance greater than described first preset distance.

31. according to the stereo generation device of claim 28, wherein, described surround audio signal comprises right and left rings around signal, the remaining audio signal comprises left signal, right signal, central signal and low-frequency effect signal.

32. a stereo generation device is used for the multichannel audio input signal is reproduced as the dual track output signal, described device comprises:

Virtual ring is around filter cell, is used for being reduced in relevance between the binaural audio signal of described multichannel audio input signal producing the sensation of reality, and described binaural audio signal is converted to virtual sound source in the precalculated position;

The signal correction filter cell is used to proofread and correct at the remaining multichannel audio input signal except described dual-channel audio input signal with from the characteristics of signals of virtual ring between the binaural audio signal of filter cell output; And,

Addition unit, be used for being added in mutually from described virtual ring and will output to the signal of first sound channel, and be added in the signal that around the multi-channel audio signal of filter cell and the output of described signal correction filter cell, will output to second sound channel from described virtual ring mutually around the multi-channel audio signal of filter cell and the output of described signal correction filter cell.

33. according to the stereo generation device of claim 32, wherein, described virtual ring comprises around filter cell:

Delay cell is used for predetermined separately length of delay and postpones first and second channel input signal;

Gain unit is used for being adjusted at each output level of first and second channel input signals that described delay cell postpones;

First addition unit is used for described first channel input signal of addition and described gain and postpones controlled second sound channel signal;

First filter cell is used to adjust the frequency characteristic from the signal of described first addition unit output;

Second addition unit is used for described second channel input signal of addition and described gain and postpones controlled first sound channel signal; And,

Second filter cell is used to adjust the frequency characteristic from the signal of described second addition unit output.

34. a stereo production method is used for virtual effect is applied to binaural signal, described method comprises:

With the frequency band division of first and second sound channel signals is high frequency band and low-frequency band;

Each of the described first and second sound channel low band signal of branch sample;

By being reduced in that the relevance of dividing separately between the sample signal produces virtual sound source and at precalculated position output virtual sound source;

Carry out interpolation with respect to described first and second sound channel signals of exporting as virtual sound source;

Described first and second sound channel signals that are interpolated of low-pass filtering; And

Described first sound channel signal that is low pass filtering of addition and high frequency first sound channel signal that is delayed, and described second sound channel signal that is low pass filtering of addition and the described high frequency that is delayed second sound channel signal.

35. according to the method for claim 34, wherein, the generation of described virtual sound source comprises:

Feel to carry out preliminary treatment filtering by being reduced in the relevance between the branch sample signal separately and producing reality; And

By branch sample signal output conduct is separately carried out virtual speaker filtering at the virtual sound source in precalculated position.

36. a stereo production method is used for virtual effect is applied to binaural signal, described method comprises:

Carry out preliminary treatment filtering by the relevance that lowers between first and second sound channel signals, and produce the sensation of reality;

With the frequency band division of first and second sound channel signals of described pretreated filtering is high frequency band and low-frequency band;

Each of the described first and second sound channel low band signal of branch sample;

Carry out virtual speaker filtering by the output conduct at the sample of the branch separately signal of the virtual sound source in precalculated position;

Carried out interpolation with respect to what export by first and second sound channel signals of virtual speaker filtering as virtual sound source;

Described first and second sound channel signals that are interpolated of low-pass filtering; And

Described first sound channel signal that is low pass filtering of addition and high frequency first sound channel signal that is delayed, and described second sound channel signal that is low pass filtering of addition and the described high frequency that is delayed second sound channel signal.

37. a stereo production method is used for the multichannel audio input signal is reproduced as the dual track output signal, described method comprises:

Be reduced in the relevance between the binaural audio signal in the described multichannel audio input signal, and produce the sensation of reality;

Described binaural audio signal is converted to virtual sound source in the precalculated position; And,

Adjust remaining multi-channel audio signal except described binaural audio signal according to the output level of the binaural audio signal after the described conversion and time delay, and, with the signal output adjusted as binaural signal.

38. the method according to claim 37 also comprises:

Behind the controlled signal of output, addition will output to the signal of first sound channel, and addition will output to the signal of second sound channel.

39. according to the method for claim 37, wherein, the generation of the reduction of the relevance between the described binaural audio signal in described multichannel audio input signal and reality sensation comprises:

Carry out the first delay operation by first channel audio signal being postponed first scheduled time;

Postpone operation by second scheduled time of rising tone channel audio signal delay being carried out second;

Postpone operation by three scheduled times of the first channel audio signal delay control being carried out the 3rd;

Postpone operation by the 4th scheduled time of rising tone channel audio signal delay being carried out the 4th;

Carry out first addition by addition by first predetermined gain being multiply by described first output and described second each value that obtains that postpones the output of operation that postpones operation;

Carry out second addition by addition by second predetermined gain being multiply by described second output and the described the 3rd each value that obtains that postpones the output of operation that postpones operation;

Carry out the 5th delay operation by filtering first signal and with described five scheduled times of the first filtering signal delay control, described first signal is to obtain by output and the 4th output that postpones to operate that addition first postpones to operate;

Carry out the 6th delay operation by the filtering secondary signal and with described six scheduled times of the second filtering signal delay control, described secondary signal is to obtain by output and the 3rd output that postpones to operate that addition second postpones to operate; And

By the addition the described the 5th and the output of the 6th delay operation and the audio signal of first and second sound channels are carried out third and fourth addition respectively.

40., wherein, multiply by different gains respectively with the output signal of described the 4th addition with the described the 3rd according to the method for claim 39.

41. according to the method for claim 39, wherein, described first postpones operation and the described the 6th postpones to operate asymmetric each other.

42., wherein, eliminate electric-wave filter matrix by multiply each other ears composite filter matrix and cross-talk and carry out and described binaural audio signal is converted to virtual sound source in the precalculated position according to the method for claim 37.

43. a stereo production method is used for left back and right back generation virtual speaker the listener, described method comprises:

Adjust the gain and the delay of L channel input signal;

Adjust the gain and the delay of R channel input signal;

Described L channel input signal of addition and described gain and postpone controlled right-channel signals to obtain first sum signal;

Adjust the frequency characteristic of first sum signal, and the result is outputed to left speaker;

Described R channel input signal of addition and described gain and postpone controlled left channel signals to obtain second sum signal; And,

Adjust the frequency characteristic of described second sum signal, and the result is outputed to right loud speaker.

44. according to the method for claim 43, wherein, by determining described gain in the maximum difference between the impulse response separately that is associated with two head related transfer functions (HRTF) between loud speaker and listener's ears.

45. according to the method for claim 43, wherein, the time when becoming maximum by the cross-correlation function when the impulse response that is associated with two HRTF between loud speaker and listener's ears is determined to postpone.

46., wherein, determine to gain by the difference between the maximum of the impulse response that is associated at two filters of lattice structure with design in advance according to the method for claim 43.

47. according to the method for claim 43, wherein, the time the when cross-correlation function by the impulse response that is associated when two filters of lattice structure with design in advance becomes maximum is determined to postpone.

Images