A kind of acoustic signal processing method and device for spatialized soundTechnical field
The present invention relates to Audio Signal Processing field, more particularly, to a kind of Audio Signal Processing for spatialized soundMethod and device.
Background technology
So-called its key of sound spatialization audio system is the space attribute for embodying sound source and acoustic irradiation.Sound hasThere are many attributes, its most basic attribute includes and the wavelength of time correlation, frequency, and the loudness related to loudness, acoustic pressureThe parameters such as level.Why human auditory can judge this sequence of acoustic characteristic that most basic is a little that the mankind gather aroundThere are a pair of ears.Two ears are located at the left side and right side on head respectively.Human brain by the loudness between two ears, frequency andThe difference of time carrys out location sound so that people can perceive out the position of sound source.
Positioned for the left and right of sound, there are 2 points of most obvious difference：
(1) in high-frequency range, because sound source to ears is apart from difference, loudness is made to create a difference, i.e. ILD(Interaural Level Differences, ears level error) or IID (Interaural IntensityDifferences, ears loudness are poor).
(2) in all frequency ranges, due to ears in itself apart from difference, the time is created a difference, i.e. ITD(Interaural Time Differences, ears time difference).
Under normal circumstances, loudness difference and the collective effect of time difference cause people to can interpolate that sound position.Deeper level is groundStudy carefully discovery over different frequencies, both difference play the role of different.Lord Rayleigh were in the Duplex proposed in 1907Theory points out that low frequency positioning relies on the time difference, and high frequency positioning is poor by loudness.
Spatialized sound can be picked up by using 2 microphones, then be played back by 2 loudspeakers.Such asOTRF pickup standards, i.e., using 2 single directional microphones, vibrating diaphragm is mutually 110 degree of angle at a distance of 17 centimetres.17 centimetres of differenceAway from make it that the time creates a difference；110 degree of angular compliance list directional microphone causes loudness to create a difference.During playback, 2 are usedLoudspeaker independently plays back the sound of 2 microphone records, you can by the audio reproduction of spatialization.Exactly because with 2 microphones pairSpatialized sound is picked up, while have recorded loudness differences and time difference so that current live performance recording has non-Normal real spatial impression.
Spatialized sound can also be simulated by artificially producing loudness difference and time difference.Such case is picked up using monophonicIt is very common in sound or music making.For example, same sound, deliver to 2 loudspeakers, but the loudness of two loudspeakers ifDifference, then it is artificial to generate volume difference.Similarly, if the loudness of 2 loudspeakers is identical, but existence time is poor, then peopleFor generate the time difference.
The method for producing loudness difference at present is easily achieved.Such as on sound console or Digital Audio Workstation software" acoustic image control " and " balance control " can be realized and the loudness ratio of two sound channels in two-channel is controlled.Acoustic image control is generally usedIn monophonic passage.Controlled by acoustic image, monophonic sounds can be assigned to two buses by a certain percentage.Thereby produce soundDegree is poor；Balance control is generally used for stereo channel.It can change the loudness ratio of stereo left and right passage by balancing control,Also it is poor that loudness is generated to a certain extent.
But the method for current generation time difference is more complicated, Major Difficulties have it is following some：
(1) method of generation time difference is to use carryover effects device.But current most of carryover effects devices are still with effectProcessor form is present, and for reflection of the simulated sound in space, the least unit of time delay is ms, with SteinbergExemplified by ModDelay, its time delay shows that least unit is 1ms, and the least unit that can be inputted is 0.1ms.Relative to for producingThe exemplary value such as 0.26ms of raw time difference, precision have seemed very low.
(2) common delay disposal device conveniently, intuitively can not show and adjust the time delay between the passage of left and right.WithExemplified by Avid Pro Tools Mod Delay II delay disposal devices, the time delay of its left passage and right passage is only by twoVertical control is completed, and rather than the acoustic image control for loudness difference, is only completed by a control.
Due to above reason, when carrying out spatialization to sound during common music making, due to only applyingLoudness is poor, causes final result to be unable to reach telepresenc and the sense of reality possessed by the spatialized sound of 2 microphone pickups.
The content of the invention
It is an object of the present invention to overcome the above-mentioned drawbacks of the prior art and provide a kind of spatialization effect is good,The simple and effective acoustic signal processing method and device for spatialized sound of algorithm.
The purpose of the present invention can be achieved through the following technical solutions：
A kind of acoustic signal processing method for spatialized sound, this method are specially：
Receive pending audio signal, obtain sound position control parameter in space, loudness difference control intensity and whenBetween difference control intensity, spatialization processing is carried out to pending audio signal using independent control pattern or overall-in-one control schema pattern.
The independent control pattern is specially：
According to position control parameter, loudness difference control intensity and the time difference control intensity point of the sound of acquisition in spaceIt is other that pending audio signal is controlled, obtain the audio signal after spatialization.
The overall-in-one control schema pattern is specially：
Existed according to position control parameter, loudness difference control intensity and the time difference control intensity of the sound of acquisition in spaceOverall-in-one control schema is carried out to pending audio signal in 2 dimension control interfaces, obtains the audio signal after spatialization.
In 2 dimension control interfaces, X-axis controls the position of sound spatialization,
Following state be present in Y-axis：
A) lower position only loudness difference in Y-axis pole influences sound, b) Y-axis pole top position only time difference influence sound, c) Y-axisCentre position, loudness difference and time difference joint effect sound, d) Y-axis other positions, then the transition between tri- kinds of states of a, b, c.
The method for producing loudness difference is realized by changing the loudness ratio of two sound channels in two-channel；
The method of generation time difference adds delay disposal device to realize by some sound channel in two-channel.
A kind of audio signal processor for spatialized sound, including：
Pending audio signal input module, for receiving pending audio signal；
Position control parameter input module, for obtaining the position control parameter of sound in space；
Loudness difference control intensity input module, for obtaining loudness difference control intensity；
Time difference controls intensity input module, for obtaining time difference control intensity；
Control mode switch module, for according to input instruction switching control pattern；
Individual control module, for position control parameter, the loudness difference control intensity according to the sound of acquisition in spacePending audio signal is controlled respectively with time difference control intensity；
Overall-in-one control schema module is strong for being controlled according to position control parameter, the loudness difference of the sound of acquisition in spaceDegree and time difference control intensity carry out overall-in-one control schema in 2 dimension control interfaces to pending audio signal.
The pending audio signal input module includes left channel audio signal input block and right channel audio signalInput block.
Existing " acoustic image control " and " balance control " on sound console or Digital Audio Workstation software is only capable of passing through soundDegree difference to sound carry out spatialization, compare with real world and lacked the time difference so that the sound of spatialization lack telepresenc withSpatial impression, the result of sound spatialization is only embodied between 2 loudspeakers.Although it can be increased using reverberation and other effects deviceSpatial impression, but the spatial impression as caused by the time difference in real world can not be simulated.The present invention compared with prior art, hasFollowing beneficial effect：
(1) this method uses loudness difference and time difference simultaneously, can simulate the result of sound localization in real world, andControl mode is still based on traditional " acoustic image control " so that this method both can be with completely compatible existing sound localizationMethod, localization of sound can also be strengthened by increasing the time difference.
(2) additionally, due to can with independent control loudness difference and the time difference influence sound intensity, can also by using onlyTime difference obtains some special effects.
(3) result of real world can be more nearly, is greatly enriched using the sound after this method processing, spatializationThe expressive ability of works.
Brief description of the drawings
Fig. 1 is the schematic flow sheet of the present invention；
Fig. 2 realizes schematic diagram for the present invention in Max platforms；
Fig. 3 is Max platforms comb~algorithm filter principle schematic；
Fig. 4 is the schematic diagram for applying the inventive method in Max platforms under independent control pattern；
Fig. 5 is the schematic diagram for applying the inventive method in Max platforms under overall-in-one control schema pattern.
Embodiment
The present invention is described in detail with specific embodiment below in conjunction with the accompanying drawings.The present embodiment is with technical solution of the present inventionPremised on implemented, give detailed embodiment and specific operating process, but protection scope of the present invention is not limited toFollowing embodiments.
The embodiment of the present invention provides a kind of acoustic signal processing method for spatialized sound, and this method includes simultaneouslyTo loudness difference and the simulation of time difference, and by certain algorithm, make user with the concept of traditional " acoustic image control ", it is auxiliaryWith intuitively state modulator so that user to loudness difference and the time difference easily controlled, allow sound spatialization controlMore directly, and make its result truer.
The present invention solves the difficult point of the prior art mentioned in background technology using following methods：
(1) using as the small as possible audio processing algorithms of minimum delay unit, now for audio frequency process might notIt is conventional delay effect device.Such as employed in the present embodiment and be originally used for making under the Max platforms of the companies of Cycling ' 74The effect device comb of comb-filter effects~.
(2) using certain algorithm be transformed into common control device the time delay between the passage of left and right, and be aided withExtra parameter.Traditional " acoustic image control " is remained in this example, and is aided with loudness difference control intensity and time difference control by forceSpend the two additional parameters.
As shown in figure 1, the inventive method is specially：Pending audio signal is received, obtains the position of sound in spaceControl parameter, loudness difference control intensity and time difference control intensity, are treated using independent control pattern or overall-in-one control schema patternHandle audio signal and carry out spatialization processing.
Wherein, independent control pattern is specially：Controlled according to position control parameter, the loudness difference of the sound of acquisition in spaceIntensity processed and time difference control intensity are controlled to pending audio signal respectively, obtain the audio signal after spatialization.
Overall-in-one control schema pattern is specially：Controlled according to position control parameter, the loudness difference of the sound of acquisition in spaceIntensity and time difference control intensity carry out overall-in-one control schema in 2 dimension control interfaces to pending audio signal, obtain spatializationAudio signal afterwards.In 2 dimension control interfaces, X-axis controls the position of sound spatialization, and Y-axis has following state：
A) lower position only loudness difference in Y-axis pole influences sound, b) Y-axis pole top position only time difference influence sound, c) Y-axisCentre position, loudness difference and time difference joint effect sound, d) Y-axis other positions, then the transition between tri- kinds of states of a, b, c.
The above method is verified and implemented under the Max software platforms of the companies of Cycling ' 74.Max is a kind of figureChange programming language, it is widely used for such performance data by composer, concert performer, artist, software designer and research staffProcessing, audio frequency process and various audio units.Due to its flexibility and opening, present Max has been depicted as Development of InteractiveThe all-purpose language of music performance software.
It is illustrated in figure 2 using Max to realize the flow of the above method, its middle port 1,2 is used to receive pending audioSignal, port 3 are used to obtain the position control parameter of sound in space, and port 4 is used to obtain loudness difference control intensity, port5 in acquisition time difference control intensity.Because Max is generally controlled with MIDI relevant devices, therefore the data area of port 3,4,57bit, i.e., 0~127 are arranged to, complies fully with midi standard.
In the present embodiment, the method for producing loudness difference can be by changing the loudness ratio of two sound channels in two-channel come realIt is existing.Benjamin Bauer draw prediction loudness difference and sound orientation angle according to two-way speaker model and spherical headformThe formula of degree：
Wherein φ is the sound orientation angle that people feels, θ is the angle between two-way speaker；L and R is each soundThe loudness of road loudspeaker.
Calculated according to above formula, when the two-way speaker angle that hearer uses is 60 degree, if the sound of left and right acoustic channelsFor degree difference in 30dB, the sensation that people position to sound is in ultra-Left or ultra-Right loudspeaker.Continuing increase loudness difference will not be moreObvious positioning change.But understand in view of conventional " acoustic image control " and voice signal is fully allocated to a passage, in this methodStill the conventional method is used.Method of the volume using cosine wave table (cycle~) control loudness (*~) is controlled, L channel ringsDegree control uses 0 degree to 90 degree phase, and the control of R channel loudness uses 270 degree to 360 degree phases, and interchannel phase difference is270 degree.The position data of 3 inputs is converted into the acquiescence of 0 degree to 90 degree change by expr $ f1/127*0.25-0.125 firstCentre position, i.e., 45 degree；4 inputs control the intensity of its change, are converted into by expr $ f1/127 by the 0 to 127 of midi standard0 to 1, and it is multiplied by the position phase that an expr is drawn, you can realize the intensity that loudness difference influences.The actual control numerical value drawnBy delivered to after $ 120 and line~smoothing processing cycle~, output 0 to 1 data be used for control loudness.
In the present embodiment, the method for generation time difference can add delay disposal device in the sound channel of some in two-channel and comeRealize.Specific time delay can be drawn by Woodworth formula：
Wherein r is a radius, and c is speed of sound, and θ is the sound orientation angle that people feels.
Calculated according to above formula, as a diameter of 8.75cm of hearer, two-way speaker angle is 60 degree, if leftThe time difference of R channel, the sensation that people position to sound was in ultra-Left or ultra-Right loudspeaker in about 260us.It is poor with loudnessUnlike, if continuing to increase the time difference, people can feel some other effects：
When the time difference reaching 10ms or so, sound positioning can be felt feel not in undelayed loudspeaker, hearerTo the sound of delay loudspeaker；
When the time difference is in 10~30ms, sound positioning is felt still in undelayed loudspeaker, but can increase certain" presence " and volume.Wherein 15~25ms is generally used for making stereophonic effect for monophonic sounds；
When the time difference is in 30~50ms, people may feel that slight carryover effects；
When the time difference is in more than 50ms, people may feel that obvious delay effect.
The maximum delay numerical value used in this method is 10ms, therefore can obtain most strong positioning and feel.In order to carryThe time precision of high latency, in this example using comb~, the algorithm filter principle is as shown in Figure 3.
It is 10ms, initial delay time 0ms that comb~parameter 100010 in this method, which specifies maximum delay time,Primary signal a=0, postpones signal intensity b=1, postponed signal intensity c=0.Due to comb~carried out to time of delaySmoothing processing, thus the time delay inputted be not necessarily accurate to the time of sampled point, therefore obtained the higher delay of precisionTime.5 inputs are maximum delay time, i.e., the intensity that the time difference influences, and the numerical value first will by expr f1/127*10The 0 to 127 of midi standard is converted into 0 to 10, and corresponding 0 to 10ms time delay, then the final time of L channel delay is firstFirst pass through expr abs ($ f1/127-0.5) * 2* ($ f1>64), judging whether acoustic image positions are more than that 64 (acoustic image positions are openedBegin to the right), and 0 to 1 will be converted into from 3 acoustic image positions, then be multiplied with maximum delay time obtained above, that is, obtain a left sideThe final delay time of sound channel.The final time of R channel delay as L channel, simply judges whether acoustic image positions are less than64, therefore use expr abs ($ f1/127-0.5) * 2* ($ f1<64).
Fig. 4 is the example that algorithm above is applied to sound spatialization in Max, uses independent control pattern.Wherein3rd, acoustic image positions, loudness difference intensity and time difference intensity are corresponded to 4,5 respectively, its excursion meet the 0 of midi standard to127。
Fig. 5 is the example that algorithm above is applied to sound spatialization in Max, uses overall-in-one control schema pattern.2 dimensionsIn control interface, X-axis control sound spatialization position, Y-axis below pole only loudness difference influence sound, Y-axis above pole onlyTime difference influences sound.When Y-axis is in centre, loudness difference and time difference joint effect sound.In other positions, then this 3Transition between kind state.