CN102637432B - Self-adaptive measuring method for dual-aural cue perceptual characteristic in three-dimensional audio coding - Google Patents
Self-adaptive measuring method for dual-aural cue perceptual characteristic in three-dimensional audio coding Download PDFInfo
- Publication number
- CN102637432B CN102637432B CN2012100743217A CN201210074321A CN102637432B CN 102637432 B CN102637432 B CN 102637432B CN 2012100743217 A CN2012100743217 A CN 2012100743217A CN 201210074321 A CN201210074321 A CN 201210074321A CN 102637432 B CN102637432 B CN 102637432B
- Authority
- CN
- China
- Prior art keywords
- value
- change
- reversion
- gain
- current
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Images
Landscapes
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
In order to solve the problems of overlarge number of sound channels and overlarge data size in three-dimensional audio frequency, all the sound channels are mixed downwards and a dual-aural cue for expressing spatial information is extracted so that the coding rate of the three-dimensional audio frequency can be effectively reduced. The invention discloses a self-adaptive measuring method for a dual-aural cue perceptual characteristic in three-dimensional audio coding. The just perception difference of the dual-aural cue is located through self-adaptively adjusting the variable step size of the dual-aural cue according to a test result by using a circling approach method so that the quantization coding of spatial audio frequency is guided.
Description
Technical field
The invention belongs to field of acoustics, but relate in particular to the just self-adapting testing method of difference in perception of binaural cue.
Background technology
In the end of the year 2009, three-dimensional movie " A Fanda " is climbed up top box-office value in more than 30 country in the whole world, and at the beginning of 2010 9 months, accumulative total box office, the whole world is over 2,700,000,000 dollars.Why " A Fanda " can obtain the box office achievement of splendidness like this, is that it has adopted brand-new three-dimensional special effect making technology to bring the effect of the shock on people's sense organ.Gorgeous picture and sound effect true to nature that " A Fanda " represents have not only shaken spectators, the asserting of " film enter three-dimensional epoch " that also made industry have.Moreover, it also will expedite the emergence of technology and the standard of more relevant video display, recording, broadcasting aspect.In the international consumption electronic product exhibition of holding at Las Vegas, US in January, 2010, the TV new product that each colour TV giant reveals has one after another brought the new expectations of people---the three-dimensional new focus that has become each large colour TV manufacturer competition of the whole world.Want to reach better audiovisual experience, the three-dimensional sound field auditory effect of synchronizeing with 3 d video content need to be arranged, could really reach audiovisual impression on the spot in person.Early stage three dimensional audio systems (as the Ambisonics system), due to its complex structure, is had relatively high expectations to collection and playback apparatus, is difficult to promote practical.Japanese NHK company has released 22.2 sound channel systems in recent years, can pass through 24 three-dimensional sound fields that loudspeaker reproduction is original.MPEG in 2011 set about formulating the international standard of three-dimensional audio, when reaching certain code efficiency, wish to reduce three-dimensional sound field by fewer loudspeaker or earphone, so that can be by this Technique Popularizing to the ordinary family user.Three-dimensional audio frequency and video technology has become the study hotspot of multimedia technology field and the important directions further developed as can be seen here.
In three-dimensional audio, number of channels increases and causes the problem that data volume is increased sharply to solve by efficient three-dimensional audio coding.At first to remove statistics and the perception redundancy of each sound channel of three-dimensional audio, utilizing the binaural cue of Ear detection, multi-channel signal is carried out to efficient parameter expression, reach the purpose of removing the spatial information redundancy, can when realizing high compression efficiency, keep space audio preferably.
Binaural cue is being passed on the information of the spatial perception in sound signal.In audio coding decoding, the quantization encoding of binaural cue can meet the surround sound effect of the rear sound signal of decoding.Simultaneously, the quantizing bit number taken due to binaural cue is few, and therefore better application is arranged in the field of space audio.
In field of acoustics, the binaural cue of general indication comprises three: intensity difference at two ears (ILD), ears mistiming (ITD) and the ears degree of correlation (IC).ILD and ITD characterize the information that space sound field direction feeling is known, and the IC sign is the width of spatial sound picture.The sensing range of people's ear is limited, and the frequency range of mankind's energy perception is between 20Hz-20kHz.Different individualities is discrepant in the frequency perception.In order better to utilize binaural cue, further compressed sensing redundancy, widen the application of space audio, and research people's ear perception dependency characteristic to binaural cue under each frequency is necessary.
Just but difference in perception (JND) is an important parameter of research binaural cue perception dependency characteristic, what its represented is that people's ear changes lucky non value to binaural cue, just can weigh the threshold of perception current of binaural cue under each frequency by the size of JND, instruct Audio Perceptual Coding.And the size of JND can only obtain by a large amount of subjective testing experiments.In existing research, the method for testing of taking is generally to make a standard pronunciation and a large amount of test tones, and constantly criterion sound and test tone find lucky undistinguishable value repeatedly to allow the auditor.Such method of testing has been brought a large amount of workloads, and easily produces auditory fatigue and the sense of hearing is obscured, and has increased the weight of tester's burden, and in fact is difficult to hold the degree of just distinguishing.
Summary of the invention
The present invention is directed to the deficiencies in the prior art, the self-adapting testing method of binaural cue apperceive characteristic in a kind of three-dimensional audio coding is provided.
The self-adapting testing method that technical scheme of the present invention is binaural cue apperceive characteristic in a kind of three-dimensional audio coding comprises following steps:
Step 2, change change step SP in real time according to the judged result that adopts step 1 gained cycle tests to be tested
dvalue, implementation is as follows,
Put positive exact figures N
r=0 is original state, when judged result is correct, by wrong number N
wbe set to 0, by positive exact figures N
radd 1, until N continuous
rafter=N1 judged result is correct, reduce change step SP
dvalue, and by positive exact figures N
rwith wrong number N
wsetting to 0, judge whether reversion to occur, is to preserve current reversion number of times and change step SP
dcurrent value also enters step 3, otherwise returns to step 1 according to change step SP
dcurrent value generates cycle tests; When the judged result mistake, by positive exact figures N
rbe set to 0, by wrong number N
wadd 1, work as N continuous
wafter=N2 judged result mistake, increase change step SP
dvalue, and by positive exact figures N
rwith wrong number N
wsetting to 0, judge whether reversion to occur, is to preserve current reversion number of times and change step SP
dcurrent value also enters step 3, otherwise returns to step 1 according to change step SP
dcurrent value generates cycle tests; Wherein, N1 and N2 are respectively the default SP that reduces
dvalue frequency threshold value and increase SP
dthe value frequency threshold value;
Described reversion, refer to change step SP
dbeing changed to and reducing or, from reducing to become increase, be called once reversion from increase, is current reversion number of times from performing step for the first time 2 to current reversion number of times altogether;
Step 3, the reversion frequency threshold value L according to default, if current reversion number of times reaches L time, enter step 4, if current reversion number of times does not reach L time, returns to step 1 according to change step SP
dcurrent value generates cycle tests;
Step 4, the change step SP during to nearest t reversion
dvalue be averaged, but the proper difference in perception mean value that obtains testing under test frequency X, t is the preset times value.
And, in step 2, N continuous
rafter=N1 judged result is correct, reduce change step SP
dthe value implementation for making SP
d=SP
d/ gain-step; N continuous
wafter=N2 judged result mistake, increase change step SP
dthe value implementation for making SP
d=SP
d* gain+step; Wherein, gain is the index running parameter, and step is the linear change parameter.
And the value of establishing gain has g1, g2, g3, tetra-kinds of g4, the value of step has s1, s2, s3, tetra-kinds of s4; If current reversion number of times is designated as reversals, for reversals arranges R1, R2, R3, tetra-kinds of critical values of R4, R1<R2<R3R4=L; For change step SP
dvalue SP is set
1, SP
2, SP
3three kinds of critical values, SP
1<SP
2<SP
3; In step 3, if current reversion number of times does not reach L time, carry out returning to step 1 after the value of following processing adjustment index running parameter gain and linear change parameter s tep,
Judging whether reversals>R1, is to judge whether reversals>R2, otherwise makes gain=g1, step=s1; If reversals>R2, judge whether reversals>R3, otherwise judge whether SP
d>SP
1, be to make gain=g1, step=s1, otherwise make gain=g2, step=s2;
If reversals>R3, judge whether reversals>R4, otherwise judge whether SP
d>SP
2, be to make gain=g2, step=s2, otherwise make gain=g3, step=s3;
If reversals=R4, enter step 4, otherwise judge whether SP
d>SP
3, be to make gain=g3, step=s3, otherwise make gain=g4, step=s4.
The present invention utilizes the frequencydependence characteristic of auditory perceptual characteristic and binaural cue, test self-adaptation by reality and adjust the binaural cue change step, but take the proper difference in perception of the method Localization binaural unit clue of progressively approaching, thereby reduced testing time, improved measuring accuracy.
The accompanying drawing explanation
Fig. 1 is the process flow diagram of the self-adapting testing method of the embodiment of the present invention.
Fig. 2 is the aural training process flow diagram of the embodiment of the present invention.
Fig. 3 is the adjustment process flow diagram of the change step of the embodiment of the present invention.
Fig. 4 is the change step adaptive change schematic diagram of the embodiment of the present invention.
Fig. 5 is the test macro block diagram of the embodiment of the present invention.
Embodiment
The present invention, mainly based on the space psychologic acoustics, considers the frequencydependence characteristic of auditory perceptual characteristic and spatial parameter, but experimental technique and the system of the proper difference in perception of a kind of adaptive testing binaural cue of proposition.This method has taken into full account each individual different apperceive characteristic, but approaches the proper difference in perception of binaural cue by the method that self-adaptation is adjusted step-length.The result obtained by the present invention is science more, more accurate.
Method provided by the invention can be used the computer software technology realization flow.Referring to Fig. 1, embodiment be take binaural cue ILD and flow process of the present invention is carried out to a concrete elaboration as example, as follows:
Definition SP
reffor the ILD value of reference sound, SP
testfor the ILD value of test tone, SP
dfor the target JND value of representative test, be change step, SP
test=SP
ref+ SP
d.
The concrete implementation process of embodiment is described as follows:
Take the corresponding time is combined into a cycle tests by this with reference to sound and test tone as interval.Front and back order random alignment, such as:
SP
ref=0, SP
d=4, so SP
test=SP
ref+ SP
d, according to SP
refvalue and SP
testvalue generates accordingly with reference to sound and test tone.If a is the reference sound to generate, establish the test tone of b for generating, can first play a, then certain interval of time broadcasting b, also can first play b, then certain interval of time broadcasting a, the playing sequence of a and b is random.The corresponding time at interval is generally fixed value, for example gets 0.5 second.
If generate for the first time cycle tests, change step SP
dadopt default initial value, adopt afterwards the change step SP after step 2 is readjusted
dvalue.
Can adopt the cycle tests generated for the first time to carry out aural training to the auditor, judge according to training result whether auditor's hearing meets testing standard.If meet testing standard, but then carry out the just formal audiometry of difference in perception of binaural cue, adopt cycle tests to carry out the result of official testing for step 2.As shown in fig. 1, idiographic flow can be, and configuration adaptive testing experiment parameter, import default SP
refand SP
dvalue, generate the listen test sequence, determines whether audiometry first, is to carry out aural training, and undesirable process ends, meet and enter step 2.Also can adopt in advance other cycle testss to carry out aural training, while carrying out technical solution of the present invention without carrying out aural training.
For the sake of ease of implementation, it is as follows that the present invention advises proposing judging whether auditor's hearing meets the concrete mode of testing standard,
Training stage SP
dvalue remains unchanged, if train total max, training the accuracy that need reach for m time is a%, by judging that the training result of nearest m time determines whether to meet testing standard, specifically judge one by one interval 1-m, 2-(m+1), 3-(m+2) ..., (max-m+1) the training accuracy of-max, mean that respectively the 1st time to the m time, the 2nd time to the m+1 time, the 3rd time to the m+2 time ... (max-m+1), to the max time, is more than or equal to a% if proceed to certain interval accuracy, meets testing standard; All be less than a% if proceed to last interval accuracy, think that auditor's hearing does not reach the audiometry standard, finish test.
The embodiment specific embodiment is as follows:
At first, draft in advance the total max=50 of training, training the accuracy that need reach for m=10 time is 80%.By judging that the training result of nearest 10 times determines whether to meet testing standard.Concrete training result judges interval 1-10,2-11,3-12..., 41-50.Minimum test 10 times, test at most 50 times.Specifically can provide referring to Fig. 2 the flow process of nearest m accuracy determination methods in the training stage: the training audiometry, the training of judgement number of times whether>=m, otherwise return to the training audiometry, it is the accuracy that judges nearest m training result, be more than or equal to a%, meet the listen test standard, training finishes, and enters formal audiometry; Be less than a%, the training of judgement sum whether>max, otherwise return to the continuation aural training, be to illustrate not meet the listen test standard, training finishes.
Then, the training accuracy in real-time judge each interval as above.Be more than or equal to 80% if proceed to certain interval accuracy, can finish training, think and reach the audiometry standard, proceed to step 3.If frequency of training to 50 time accuracy all be less than 80%, think that auditor's hearing does not reach the audiometry standard, finish listen test.
Step 2, change change step SP in real time according to the judged result that adopts step 1 gained cycle tests to be tested
dvalue.
While generally being tested, according to the sequence of current cycle tests judgement perception preference for orientation one side.Embodiment specifies left side, with reference to sound, an orientation is arranged, and test tone is than take back any sound of reference sound; During test, can be shuffle with reference to sound and test tone, which takes back two sections sounds playing before and after the tester judges, if select be test tone select judged result correct, otherwise judged result mistake.Concrete test is embodied as prior art, and technical scheme of the present invention directly adopts the judged result that cycle tests is tested to adjust change step.For the sake of ease of implementation, the present invention further provides the real-time change change step SP of embodiment
dthe concrete mode of value as follows:
Put positive exact figures N
r=0 is original state, when judged result is correct, by wrong number N
wbe set to 0, by positive exact figures N
radd 1, until N continuous
rafter=N1 judged result is correct, reduce change step SP
dvalue, and by positive exact figures N
rwith wrong number N
wsetting to 0, judge whether reversion to occur, is to preserve current reversion number of times and change step SP
dcurrent value also enters step 3, otherwise returns to step 1 according to change step SP
dcurrent value generates cycle tests; When the judged result mistake, by positive exact figures N
rbe set to 0, by wrong number N
wadd 1, work as N continuous
wafter=N2 judged result mistake, increase change step SP
dvalue, and by positive exact figures N
rwith wrong number N
wsetting to 0, judge whether reversion to occur, is to preserve current reversion number of times and change step SP
dcurrent value also enters step 3, otherwise returns to step 1 according to change step SP
dcurrent value generates cycle tests.
Wherein, N1 and N2 are respectively and reduce SP
dvalue frequency threshold value and increase SP
dthe value frequency threshold value, reduce SP
dthe required minimum of value judges it is number of times and increase SP continuously
dbe worth the required minimum no number of times of judgement continuously.N1 and N2 can according to circumstances oneself preset numerical value by those skilled in the art.Namely continuously correct number of times is less than N1, and errors number is less than N2 continuously, change step SP
dvalue all constant.
Described reversion, refer to change step SP
dvariation tendency while changing (comprise from increase be changed to reduce or from reducing to become increases), be called once and reverse, from performing step for the first time 2, to current reversion number of times altogether, be current reversion number of times.
As shown in Figure 3, in idiographic flow, can while being correct, at first judge whether N in judged result
w>0, be to establish wrong number N
w=0, positive exact figures N
r=1.Otherwise direct N
r=N
r+ 1, at N
rall keep SP during<N1
dbe worth constant, the cycle tests (SP that also remains unchanged
test=SP
test), continue to judge that whether the result of listen test is correct next time, until N
rduring=N1, SP
d=SP
d/ gain-step, according to SP
test=SP
ref+ SP
dgenerate new cycle tests, establish N
r=0.Can, when judged result is mistake, at first determine whether N
r>0, be to establish positive exact figures N
r=0, mistake is counted N
w=1, otherwise direct N
w=N
w+ 1, at N
wall keep SP during<N2
dbe worth constant, the cycle tests (SP that also remains unchanged
test=SP
test), continue to judge that whether the result of listen test is correct next time, until N
wduring=N2, SP
d=SP
d* gain+step, according to SP
test=SP
ref+ SP
dgenerate new cycle tests, establish N
w=0.Reduced like this operation that sets to 0 of redundancy, efficiency is higher.Because N1 may be 1, at N
w>0 and establish wrong number N
w=0, positive exact figures N
rafter=1, judging whether N1=1, is to make SP
d=SP
d/ gain-step, according to SP
test=SP
ref+ SP
dgenerate new cycle tests, establish N
r=0, otherwise keep SP
dbe worth constant, the cycle tests (SP that also remains unchanged
test=SP
test), continue to judge that whether the result of listen test is correct next time.Same, because N2 may be 1, at N
r>0 and establish wrong number N
w=1, positive exact figures N
rafter=0, judging whether N2=1, is to make SP
d=SP
d* gain+step, according to SP
test=SP
ref+ SP
dgenerate new cycle tests, establish N
w=0, otherwise keep SP
dbe worth constant, the cycle tests (SP that also remains unchanged
test=SP
test), continue to judge that whether the result of listen test is correct next time.
The embodiment specific embodiments is:
Default N1=3, N2=1, the sequence of taking back in the cycle tests generated according to step 1 judgement perception orientation.Put N
r=0 is original state, when correct judgment, by positive exact figures N
radd 1, keep change step SP
dcurrent value proceeds to step 2.Until N continuous
r=3 correct judgments, reduce test value, i.e. SP
d=SP
d/ gain-step, make with reference to sound SP
refvalue and test tone SP
testbe worth more approaching.When N is arranged
w=1 misjudgment, increase test value, i.e. SP
d=SP
d* gain+step, make with reference to sound SP
refvalue and test tone SP
testit is large that the difference of value becomes.
Gain and step are for making SP
dtwo parameters that value increases or reduces, wherein, gain is the index running parameter, can allow SP
dvalue is by index variation, and step is the linear change parameter, can allow SP
dvalue is by linear change.Gain and step can adopt default fixed value, in order to make test value, approach fast desired value, can also change in real time the value of gain and step according to current reversion number of times, so just can pass through SP
d=SP
d/ gain-step and SP
d=SP
d* these two formulas of gain+step are regulated SP
dthe value size.
Step 3, according to default reversion frequency threshold value L, if current reversion number of times reversals reaches L time, enter step 4, if current reversion number of times reversals does not reach L time, returns to step 1 according to change step SP
dcurrent value generates cycle tests.Those skilled in the art can set the L value voluntarily as the case may be.
The reversion frequency threshold value L=12 of embodiment, specific embodiment is: when inverted population reach 12 times, proceed to step 5.If the reversion number of times does not reach 12 times, proceed to step 1, according to current SP
dinformation regenerates sequence and proceeds listen test.
For the purpose of the self-adaptation regulated efficiency further is provided, in embodiment, the value of gain has g1, g2, and g3, tetra-kinds of g4, the value of step has s1, s2, s3, tetra-kinds of s4.If current reversion number of times is designated as reversals, embodiment is that reversals arranges R1, R2, and R3, tetra-kinds of critical values of R4, R1<R2<R3R4=L will change the value of gain and step when reaching a certain critical value.And be change step SP
dvalue SP is set
1, SP
2, SP
3three kinds of critical values, SP
1<SP
2<SP
3, reference change step-length SP
dcurrent value realizes adaptive change.As the adaptive change flow process of test value in the test process of embodiment that Fig. 4 is provided, N continuous
rneed to reduce SP after=N1 judged result is correct
dduring value, or N continuous
wneed to increase SP after=N2 judged result mistake
dduring value, carry out following steps:
Input current SP
dvalue, by self-adaptive processing,
Judging whether reversals>R1, is to judge whether reversals>R2, otherwise execution step A1, gain=g1, step=s1, perform step afterwards at 2 o'clock according to gain=g1, step=s1 and carry out Up calculating or Down calculating, and Up (increases SP
dvalue is calculated): SP
d=SP
d* gain+step or Down (reduce SP
dvalue is calculated): SP
d=SP
d/ gain-step;
If reversals>R2, judge whether reversals>R3, otherwise judge whether SP
d>SP
1be jump procedure A1, perform step afterwards at 2 o'clock and carry out according to gain=g1, step=s1 that Up calculates or Down calculates, otherwise execution step A2, be gain=g2, step=s2, perform step afterwards at 2 o'clock according to gain=g2, step=s2 and carry out Up calculating or Down calculating;
If reversals>R3, judge whether reversals>R4, otherwise judge whether SP
d>SP
2be jump procedure A2, perform step afterwards at 2 o'clock and carry out according to gain=g2, step=s2 that Up calculates or Down calculates, otherwise execution step A3, be gain=g3, step=s3, perform step afterwards at 2 o'clock according to gain=g3, step=s3 and carry out Up calculating or Down calculating;
If reversals>R4, adaptive change flow process completes, and enters step 4, otherwise judges whether SP
d>SP
3be jump procedure A3, perform step at 2 o'clock after then and carry out according to gain=g3, step=s3 that Up calculates or Down calculates, otherwise execution step A4, be gain=g4, step=s4, perform step afterwards at 2 o'clock according to gain=g4, step=s4 and carry out Up calculating or Down calculating.
The embodiment process relate to data the following table is:
| R1 | R2 | R3 | R4 | g1 | | g3 | g4 | |
| 1 | 3 | 6 | 12 | 0.6 | 1 | 1 | 1 | |
| s1 | s2 | s3 | s4 | | SP2 | SP3 | ||
| 0 | 0.5 | 0.3 | 0.1 | 3 | 1.5 | 0.8 |
Be R1=1, R2=3, R3=6, R4=L=12, g1=0.6, g2=g3=g4=1, s1=0, s2=0.5, s3=0.3, s4=0.1, SP
1=3, SP
2=1.5, SP
3=0.8.Concrete change step SP
dshift gears with SP
refand test frequency is relevant, those skilled in the art can need to set as the case may be other modes, or the mode provided with reference to the embodiment of the present invention arranges numerical value in table voluntarily.
Step 4, the value of the change step SPd during to nearest t reversion is averaged, but the proper difference in perception mean value that obtains testing under test frequency X, t is the preset times value, those skilled in the art can set the t value voluntarily as the case may be.
For example during t=4, if test always has 7 reversions after finishing, calculate the SP of the 4th, 5,6,7 times
dmean value.
Method provided by the invention, also can utilize modularity design technology to be embodied as corresponding system.As shown in Figure 5, can set up the adaptive testing system, comprise cycle tests generation module, change step adjusting module in system, finish judge module, test and finish module, cycle tests generation module, change step adjusting module, end judge module, test finish module and connect successively, and the cycle tests generation module is connected with the end judge module;
The cycle tests generation module, under a test frequency, according to default reference sound binaural cue SP
refwith change step SP
dgenerate cycle tests, described test frequency is designated as X; The Real-time Generation of cycle tests is, is combined into a cycle tests with reference to sound and test tone with the time interval of presetting, during combination before and after the order random alignment, test tone binaural cue SP
test=SP
ref+ SP
d; While generating cycle tests for the first time, change step SP
dadopt default initial value.The cycle tests generated can adopt existing test platform to play and collect judged result, judged result input change step adjusting module.
The change step adjusting module, change change step SP in real time according to the judged result that adopts cycle tests generation module gained cycle tests to be tested
dvalue, implementation is as follows,
Put positive exact figures N
r=0 is original state, when judged result is correct, and just wrong number N
wbe set to 0, positive exact figures N
radd 1, keep change step SP
dcurrent value, until N continuous
rafter=N1 judged result is correct, reduce change step SP
dvalue, and by positive exact figures N
rwith wrong number N
wsetting to 0, judge whether reversion to occur, is to preserve current reversion number of times and change step SP
dcurrent value end of input judge module, otherwise will with change step SP
dcurrent value input test sequence generation module generates new cycle tests; When the judged result mistake, by positive exact figures N
rbe set to 0, by wrong number N
wadd 1, work as N continuous
wafter=N2 judged result mistake, increase change step SP
dvalue, and by positive exact figures N
rwith wrong number N
wsetting to 0, judge whether reversion to occur, is to preserve current reversion number of times and change step SP
dcurrent value end of input judge module, otherwise will with change step SP
dcurrent value input test sequence generation module generates new cycle tests; Wherein, N1 and N2 are respectively the default SP that reduces
dvalue frequency threshold value and increase SP
dthe value frequency threshold value;
Described reversion, refer to change step SP
dbeing changed to and reducing or, from reducing to become increase, be called once reversion from increase, is current reversion number of times from performing step for the first time 2 to current reversion number of times altogether.
Finish judge module, for the reversion frequency threshold value L according to default, if current reversion number of times reaches L time, the change step SP during by nearest t reversion
dthe value input test finish module, if current reversion number of times does not reach L time, by change step SP
dcurrent value input test sequence generation module generates new cycle tests;
Test finishes module, the change step SP during to nearest t reversion
dvalue be averaged, but the proper difference in perception mean value that obtains testing under test frequency X, t is the preset times value.
Specific embodiment described herein is only to the explanation for example of the present invention's spirit.Those skilled in the art can make various modifications or supplement or adopt similar mode to substitute described specific embodiment, but can't depart from spirit of the present invention or surmount the defined scope of appended claims.
Claims (1)
1. the self-adapting testing method of binaural cue apperceive characteristic in three-dimensional audio coding, is characterized in that, comprises following steps:
Step 1, under a test frequency, according to default reference sound binaural cue SP
refwith change step SP
dgenerate cycle tests, described test frequency is designated as X; The Real-time Generation of cycle tests is, is combined into a cycle tests with reference to sound and test tone with the time interval of presetting, during combination before and after the order random alignment, test tone binaural cue SP
test=SP
ref+ SP
d; Perform step for the first time at 1 o'clock, change step SP
dadopt default initial value;
Step 2, change change step SP in real time according to the judged result that adopts step 1 gained cycle tests to be tested
dvalue, implementation is as follows,
Put positive exact figures N
r=0 is original state, when judged result is correct, by wrong number N
wbe set to 0, by positive exact figures N
radd 1, until N continuous
rafter=N1 judged result is correct, reduce change step SP
dvalue, and by positive exact figures N
rwith wrong number N
wsetting to 0, judge whether reversion to occur, is to preserve current reversion number of times and change step SP
dcurrent value also enters step 3, otherwise returns to step 1 according to change step SP
dcurrent value generates cycle tests; When the judged result mistake, by positive exact figures N
rbe set to 0, by wrong number N
wadd 1, work as N continuous
wafter=N2 judged result mistake, increase change step SP
dvalue, and by positive exact figures N
rwith wrong number N
wsetting to 0, judge whether reversion to occur, is to preserve current reversion number of times and change step SP
dcurrent value also enters step 3, otherwise returns to step 1 according to change step SP
dcurrent value generates cycle tests; Wherein, N1 and N2 are respectively the default SP that reduces
dvalue frequency threshold value and increase SP
dthe value frequency threshold value;
Described reversion, refer to change step SP
dbeing changed to and reducing or, from reducing to become increase, be called once reversion from increase, is current reversion number of times from performing step for the first time 2 to current reversion number of times altogether;
In N continuous
rafter=N1 judged result is correct, reduce change step SP
dthe value implementation for making SP
d=SP
d/ gain – step; N continuous
wafter=N2 judged result mistake, increase change step SP
dthe value implementation for making SP
d=SP
d* gain+step; Wherein, gain is the index running parameter, and step is the linear change parameter;
Step 3, the reversion frequency threshold value L according to default, if current reversion number of times reaches L time, enter step 4, if current reversion number of times does not reach L time, returns to step 1 according to change step SP
dcurrent value generates cycle tests;
If the value of gain has g1, g2, g3, tetra-kinds of g4, the value of step has s1, s2, s3, tetra-kinds of s4; If current reversion number of times is designated as reversals, for reversals arranges R1, R2, R3, tetra-kinds of critical values of R4, R1<R2<R3<R4=L; For change step SP
dvalue SP is set
1, SP
2, SP
3three kinds of critical values, SP
1<SP
2<SP
3; In step 3, if current reversion number of times does not reach L time, carry out returning to step 1 after the value of following processing adjustment index running parameter gain and linear change parameter s tep,
Judge whether reversals > R1, be to judge whether reversals R2, otherwise make gain=g1, step=s1;
If reversals>R2, judge whether reversals>R3, otherwise judge whether SP
dsP
1, be to make gain=g1, step=s1, otherwise make gain=g2, step=s2;
If reversals>R3, judge whether reversals>R4, otherwise judge whether SP
dsP
2, be to make gain=g2, step=s2, otherwise make gain=g3, step=s3;
If reversals=R4, enter step 4, otherwise judge whether SP
dsP
3, be to make gain=g3, step=s3, otherwise make gain=g4, step=s4;
Step 4, the change step SP during to nearest t reversion
dvalue be averaged, but the proper difference in perception mean value that obtains testing under test frequency X, t is the preset times value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012100743217A CN102637432B (en) | 2012-03-20 | 2012-03-20 | Self-adaptive measuring method for dual-aural cue perceptual characteristic in three-dimensional audio coding |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012100743217A CN102637432B (en) | 2012-03-20 | 2012-03-20 | Self-adaptive measuring method for dual-aural cue perceptual characteristic in three-dimensional audio coding |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102637432A CN102637432A (en) | 2012-08-15 |
| CN102637432B true CN102637432B (en) | 2013-12-18 |
Family
ID=46621804
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012100743217A Expired - Fee Related CN102637432B (en) | 2012-03-20 | 2012-03-20 | Self-adaptive measuring method for dual-aural cue perceptual characteristic in three-dimensional audio coding |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102637432B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104240712B (en) * | 2014-09-30 | 2018-02-02 | 武汉大学深圳研究院 | A kind of three-dimensional audio multichannel grouping and clustering coding method and system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101408614A (en) * | 2008-11-26 | 2009-04-15 | 武汉大学 | Method and device for measuring binaural sound strong difference ILD critical apperceive characteristic |
| CN101419801A (en) * | 2008-12-03 | 2009-04-29 | 武汉大学 | Method for subband measuring correlation sensing characteristic between ears and device thereof |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8612237B2 (en) * | 2007-04-04 | 2013-12-17 | Apple Inc. | Method and apparatus for determining audio spatial quality |
-
2012
- 2012-03-20 CN CN2012100743217A patent/CN102637432B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101408614A (en) * | 2008-11-26 | 2009-04-15 | 武汉大学 | Method and device for measuring binaural sound strong difference ILD critical apperceive characteristic |
| CN101419801A (en) * | 2008-12-03 | 2009-04-29 | 武汉大学 | Method for subband measuring correlation sensing characteristic between ears and device thereof |
Non-Patent Citations (4)
| Title |
|---|
| A Method to Measure the Discriminability of Spatial Cues;Heng Wang et al;《Third International Symposium on Information Science and Engineering (ISISE),2010》;20101226;244-246 * |
| Heng Wang et al.A Method to Measure the Discriminability of Spatial Cues.《Third International Symposium on Information Science and Engineering (ISISE),2010》.2010, |
| 胡瑞敏 等.基于人耳感知特性的空间参数量化.《第六届和谐人机环境联合学术会议（HHME2010)、第19届全国多媒体学术会议（NCMT2010）、第6届全国人机交互学术会议（CHCI2010）、第5届全国普适计算学术会议（PCC2010）论文集 , 2010年》.2010, |
| 胡瑞敏 等.基于人耳感知特性的空间参数量化.《第六届和谐人机环境联合学术会议(HHME2010)、第19届全国多媒体学术会议(NCMT2010)、第6届全国人机交互学术会议(CHCI2010)、第5届全国普适计算学术会议(PCC2010)论文集, 2010年》.2010, * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102637432A (en) | 2012-08-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9055382B2 (en) | Calibration of headphones to improve accuracy of recorded audio content | |
| CN102017402B (en) | System for adjusting perceived loudness of audio signals | |
| US8612237B2 (en) | Method and apparatus for determining audio spatial quality | |
| US8238563B2 (en) | System, devices and methods for predicting the perceived spatial quality of sound processing and reproducing equipment | |
| CN102100088B (en) | Apparatus and method for generating audio output signals using object based metadata | |
| EP1979900B1 (en) | Apparatus for estimating sound quality of audio codec in multi-channel and method therefor | |
| CN105376690A (en) | Method and device of generating virtual surround sound | |
| Peng et al. | Relationship between Chinese speech intelligibility and speech transmission index in rooms based on auralization | |
| CN110933459B (en) | Event video clipping method, device, server and readable storage medium | |
| CN102637432B (en) | Self-adaptive measuring method for dual-aural cue perceptual characteristic in three-dimensional audio coding | |
| Sun et al. | Realism and immersion in the reproduction of audio-visual recordings for urban soundscape evaluation | |
| Vigeant et al. | Investigations of orchestra auralizations using the multi-channel multi-source auralization technique | |
| CN102802111B (en) | A kind of method and system for exporting surround sound | |
| CN102752701B (en) | Device and method for testing direction perception sensitivity of three-dimensional space | |
| CN105759964B (en) | Sound input method and device for virtual reality equipment and virtual reality equipment | |
| CN114205724B (en) | Hearing aid earphone debugging method, device and equipment | |
| Wu et al. | Distortion reduction via CAE and DenseNet mixture network for low bitrate spatial audio object coding | |
| Brunskog et al. | Subjective response to foot-fall noise, including localization of the source position | |
| Schaefer | Auditory assessment of multichannel audio systems | |
| Otani et al. | Auditory artifacts due to switching head-related transfer functions of a dynamic virtual auditory display | |
| Hirvonen et al. | Top-down strategies in parameter selection of sinusoidal modeling of audio | |
| Xu et al. | Identification of anthropometric measurements for individualization of head-related transfer functions | |
| Irene et al. | Prediction models of desirable levels of birdsong and water sound in a noisy environment: A laboratory experiment based on virtual reality | |
| Song et al. | Simulation of realistic background noise using multiple loudspeakers | |
| Gutierrez-Parera et al. | On the influence of headphone quality in the spatial immersion produced by Binaural Recordings |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20131218 Termination date: 20200320 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |