A method for allocating an optimum amount of bits to MPEG audio data at high speed. A signal-to-mask ratio obtained through a psychoacoustic model is divided by 6 and the resultant integer quotient is defined as an array pointer. A valid bit allocation amount is obtained at the high speed according to the defined array pointer. The obtained valid bit allocation amount is compared with a fixed bit allocation amount. An optimum bit allocation amount is obtained in accordance with the compared result. Therefore, according to the present invention, the optimum bit allocation amount to the MPEG audio data can be obtained at the high speed with no unnecessary loop repetition by using the signal-to-mask ratio obtained through the psychoacoustic model.
Description Distribute the fast method of optimum bit amount for the motion picture expert group audio dataTechnical field
In general, the present invention relates to the bit quantity of motion picture expert group (hereinafter referred to as MPEG) audio data is distributed, more specifically, relate to a kind of fast method of distributing the optimum bit amount for the MPEG audio data, wherein frequency band is that unit determines a bit allocation amounts in limited MPEG audio frame bit range by carrying out minimum computing in proper order, and this bit quantity is distributed in the multi-channel system and carries out expeditiously.
Background technology
Common MPEG audio data bit quantity distribution method contains one makes signal cover than the recruitment of (SMR) or decrease being 1 repetitive cycling, and adjust the size of coverage according to its end value, this circulating repeating reduced the speed of MEPG audio data Bit Allocation in Discrete.
On the other hand, Japan Patent is openly put down into and has been provided a kind of sound signal processing method of carrying out Bit Allocation in Discrete efficiently for 5-248972 number.This sound signal processing method is used for a sound signal is converted to a frequency signal, and the frequency signal of this conversion is carried out Bit Allocation in Discrete so that it is encoded.Used one to cover threshold values and reduce the sense of hearing noise that causes by blanketing-effect.This coverage threshold values is to determine according to the signal spectra distribution of present frame and former frame.Employing can carry out digital coding based on the coverage threshold values of auditory properties more efficiently.This digital coding more efficiently causes the increase of data compression rate.Yet said method is being different aspect the enforcement with the present invention, and it to compare effect relatively poor with the present invention.
Therefore, The present invention be directed to that the problems referred to above propose, an object of the present invention is to provide a kind of fast method of distributing the optimum bit amount for the MPEG audio data, wherein can obtain the bit allocation amounts that the following frequency band is a unit and cover size, and needn't make signal cover than increasing or reducing 1 by a division arithmetic.
Summary of the invention
According to the present invention, above-mentioned purpose and other purposes can distribute the fast method of optimum bit amount to reach to the MPEG audio data by providing a kind of, this method comprises: (a) covering ratio divided by 6 by the resulting signal of psychoacoustic model, the integer part of gained quotient is defined as an array pointer, and according to defined array pointer as follows (a-1) obtain a significant bit sendout apace to step (a-6): (a-1) give by covering by the resulting signal of psychoacoustic model and cover ratio to obtain a new signal than weighted value 1.01 in addition; (a-2) whether check resulting new signal in above-mentioned steps (a-1) is covered than greater than zero, and when should be in above-mentioned steps (a-1) resulting new signal cover than less than zero the time value of the next inferior frequency band of check; (a-3) finding in above-mentioned steps (a-1) resulting new signal when the check of above-mentioned steps (a-2) covers than greater than zero the time, resulting new signal in above-mentioned steps (a-1) is covered ratio divided by 6, the integer part of resulting quotient is defined as the array pointer, and from the signal to noise ratio that indicates, deduct resulting new signal coverage ratio in above-mentioned steps (a-1) by defined array pointer, cover noise ratio to obtain one; (a-4) check in above-mentioned steps (a-3) resulting coverage noise ratio whether greater than zero; (a-5) find that in above-mentioned steps (a-3) resulting coverage noise ratio is less than zero the time when the check of above-mentioned steps (a-4), then in above-mentioned steps (a-3), add 1 on the defined array pointer, to obtain a new array pointer, and, cover noise ratio to obtain one from covering ratio by deducting in above-mentioned steps (a-1) resulting new signal the specified signal to noise ratio of resulting new array pointer; (a-6) obtain the significant bit sendout, it is by the specified bit allocation amounts of defined array pointer in above-mentioned steps (a-3), or carrying out above-mentioned steps (a-5) the specified bit allocation amounts of resulting pointer in above-mentioned steps (a-5) afterwards, or find in above-mentioned steps (a-3) resulting coverage noise ratio corresponding bit allocation amounts greater than zero time when the check of above-mentioned steps (a-4); And (b) resulting significant bit sendout and a fixing bit allocation amounts are compared, and according to comparative result as follows (b-1) obtain an optimum bit sendout to step (b-4): (b-1) whether check resulting significant bit sendout in above-mentioned steps (a) identical with the fixed bits assignment amount, and when resulting significant bit sendout is identical with the fixed bits assignment amount in above-mentioned steps (a), end operation; (b-2) when the check of above-mentioned steps (b-1) finds that resulting significant bit sendout is greater than the fixed bits assignment amount in above-mentioned steps (a), find out an array pointer of covering noise ratio corresponding to maximum, from the array pointer that is found, deduct 1, to obtain a new array pointer, the bit allocation amounts that is indicated by resulting new array pointer is added on the fixed bits assignment amount, and from resulting addition result bit allocation amounts, deduct by the specified bit allocation amounts of array pointer that is found, to obtain the optimum bit sendout; (b-3) when the check of above-mentioned steps (b-1) finds that resulting bit allocation amounts is less than the fixed bits assignment amount in above-mentioned steps (a), find out an array pointer of covering noise ratio corresponding to minimum, on the array pointer that is found, add 1, to obtain a new array pointer, being added on the fixed bits assignment amount by the specified bit allocation amounts of resulting new array pointer, and from resulting addition result bit allocation amounts, deduct by the specified bit allocation amounts of array pointer that is found, to obtain the optimum bit sendout; (b-4) from covering noise ratio corresponding to deducting by the resulting signal of psychoacoustic model in the specified signal to noise ratio of the array pointer of resulting optimum bit sendout in above-mentioned steps (b-2) or (b-3) by one, cover noise ratio to obtain one, and turn back to above-mentioned steps (b-1).
Description of drawings
Above-mentioned purpose of the present invention and other purposes and characteristics and advantage will obtain clearer understanding by the detailed description of doing below in conjunction with accompanying drawing, in the accompanying drawings:
Fig. 1 explanation is according to the flow chart of fast bit batch operation of the present invention; And
Fig. 2 is the flow chart of explanation according to optimum bit batch operation of the present invention.
Embodiment
In MPEG II, a multi-channel system is according to a Bit String job.In order to carry out Bit Allocation in Discrete within a limited frame bit quantity scope, for a plurality of passages, must be again to each channel allocation a minimum with bit quantity the best.Minimizing of bit allocation amounts means the data compression that does not damage in the sound quality scope.
According to the present invention, minimized and best bit allocation amounts obtains by carrying out a divide operations, tests verifiedly, and when the bit allocation amounts of MPEG audio data increased or reduces by 1 bit, signal to noise ratio (snr) just increased or reduce 6dB.Basic principle of the present invention is exactly according to above-mentioned experimental results show that and by this digital value 6 is obtained bit allocation amounts as divisor.That is to say, covering than (SMR) divided by 6 by the resulting signal of psychoacoustic model, the bit allocation amounts of utilizing the quotient of gained and obtaining to be similar to is not because exist accurately 6dB change width, so each signal is covered than weighted values 1.01 in addition all in the MPEG audio data.For example, be 3 and 5 situation for quantizing progression, by their signals are separately covered than obtaining identical quotient 1 divided by 6.In this case, although quotient is identical, bit allocation amounts will be different, for this reason, each signal be covered than weighted value in addition, make the array pointer no longer identical, and here, each array pointer is defined by corresponding quotient.
The table 1 of back illustrates a Bit Allocation in Discrete table or a reference table, and wherein the pointer about Bit Allocation in Discrete obtains with said method.
Table 1 has been expressed the example of MPEG layer II.Defined array pointer N covers ratio divided by 6 quotients that obtain by handle by the signal of psychoacoustic model gained.Array pointer N indicates the bit allocation amounts of one frequency band, a signal to noise ratio (snr), a coverage noise ratio (MNR) and a Bit Allocation in Discrete desired value.The group pointer indicates each time frequency band group of same allocation step in the MPEG Bit Allocation in Discrete table.Inferior frequency band group is divided into 4 groups according to emissivity.
Now, the fast method of distributing best bit quantity for the MPEG audio data according to of the present invention will be described on the basis of mentioned above principle and table 1.
With reference to figure 1, the there shows the flow chart of explanation according to fast bit batch operation of the present invention.In order to carry out the fast bit batch operation, the quotient that is obtained by division arithmetic is defined as an array pointer that indicates signal to noise ratio, from covering ratio by subtraction signal the specified signal to noise ratio of array pointer, judges that above-mentioned difference is negative or positive number.Result according to judging determines whether to add 1 bit on bit allocation amounts.
For each time frequency band is all obtained a bit allocation amounts, the batch operation of this fast bit only need be done three kinds of computings altogether: division, subtraction and condition judgment.Therefore, the batch operation of this fast bit is simple.
Illustrate in greater detail the batch operation of this fast bit below with reference to Fig. 1.
In first step T1, by by means of psychoacoustic model resulting current demand signal cover than on weighted value 1.01 in addition, obtain a new signal and cover ratio, also be SMR=SMR+1.01.In step T2, whether resulting new signal is covered than greater than zero among the check first step T1.If the check of the second step T2 is found the resulting new signal of first step T1 and is covered than less than zero, then checks the value of next time frequency band again.
If the check of the second step T2 is found the resulting new signal of first step T1 and is covered ratio greater than zero, then in third step T3 it divided by 6, and the integer part of gained quotient (int) is defined as an array pointer, then, in third step T3 by from covering ratio by deducting the resulting new signal of first step T1 the specified signal to noise ratio of defined array pointer N, obtain one and cover noise ratio, that is, and MNR=SNR[N]-SMR.
In the 4th step T4, whether the resulting coverage noise ratio of check third step T3 is greater than zero.If the coverage noise ratio that the check of the 4th step T4 discovery third step T3 obtains is greater than zero, then the batch operation of this fast bit advances to the 6th step T6.
If the check of the 4th step T4 finds that the resulting coverage noise ratio of third step T3 less than zero, then obtains a new array pointer N by add 1 on the defined array pointer of third step T3 N in the 5th step T5, also, N=N+1.Then, in the 5th step T5,, obtains a coverage noise ratio by from covering ratio by deducting the resulting new signal of first step T1 the specified signal noise ratio of the new array pointer N of gained, that is, and MNR=SNR[N]-SMR.
After having carried out the 5th step T5, if perhaps the check of the 4th step T4 finds that the resulting coverage noise ratio of third step T3 is greater than zero, then handle perhaps is defined as effective bit allocation amounts to the resulting bit allocation amounts of the 5th step T5 by the specified bit allocation amounts of the defined array pointer of third step T3 N in the 6th step T6.
As mentioned above, the batch operation of this fast bit does not need circulation to repeat.Therefore, can obtain bit allocation amounts with very high speed.
On the other hand, be the situation of fixing for bit quantity to be allocated, must handle to such an extent that approach the maximum allocated bit quantity most by the resulting significant bit sendout of fast bit batch operation shown in Figure 1.Explain this optimum bit batch operation below with reference to Fig. 2.
Referring to Fig. 2, the there shows the flow chart of an explanation according to optimum bit batch operation of the present invention.In first step U1, whether the resulting significant bit sendout of fast bit batch operation shown in Figure 1 of upchecking is identical with fixing bit allocation amounts, if it is identical with the fixed bits assignment amount by the resulting significant bit sendout of fast bit batch operation shown in Figure 1 that the check of first step U1 is found, also be, if this significant bit sendout is best, then the optimum bit batch operation finishes.
If the check of first step U1 is found, greater than the fixed bits assignment amount, then in the second step U2, find out an array pointer of covering noise ratio corresponding to maximum by the resulting significant bit sendout of fast bit batch operation shown in Figure 1.In the second step U2, obtain a new array pointer then by from the array pointer that finds, deducting 1, be that N=N-1 and handle are added on the fixed bits assignment amount by this specified bit allocation amounts of new array pointer that obtains, the second step U2 also deducts the specified bit allocation amounts of array pointer that is found by above-mentioned from the bit allocation amounts that obtains like this.In a word, the optimum bit sendout here obtains like this:
BITSï¼BITS+bits[N]-bits[N+1]
If the check of first step U1 is found, by the resulting significant bit sendout of fast bit batch operation shown in Figure 1 less than the fixed bits assignment amount, then in third step U3, find out an array pointer of covering noise ratio corresponding to minimum, then in third step U3, by adding that from the array pointer that finds 1 obtains a new array pointer, promptly N=N+1 is also being added on the fixed bits assignment amount by this specified bit allocation amounts of new array pointer that obtains.Third step U3 also deducts the specified bit allocation amounts of array pointer that is found by above-mentioned from the bit allocation amounts that obtains like this.In a word, the optimum bit distribution here obtains the most like this.
BITSï¼BITS+bits[N]-bits[N-1]
In the 4th step U4, by from by corresponding to second or third step U2 or U3 deduct by the resulting signal of psychoacoustic model in the specified noise acoustic ratio of the array pointer of resulting optimum bit sendout and cover ratio, cover noise ratio and obtain one, that is MNR=SNR (N)-SMR.In the 4th step U4, the batch operation of this optimum bit turns back to first step U1 to repeat above steps then.
According to top explanation, clearly as can be seen, can obtain optimum bit sendout apace according to the present invention to the MPEG audio data, cover the repetition that recently circulates and do not need to utilize by the resulting signal of psychoacoustic model.If the significant bit sendout is different from the fixed bits assignment amount, then will make it approach maximum assignable bit quantity most by handling the significant bit sendout rightly, therefore, Bit Allocation in Discrete edge in the multi-channel system between each passage becomes bigger, its effect is the Bit Allocation in Discrete that can more effectively carry out to audio data, have, the present invention can be applied to adopt coded system and other related systems of MPEG sound signal compression algorithm again.
Although the preferred embodiment of the present invention disclosed herein is for the present invention is described, but those skilled in the art that can see, under the situation that does not depart from category of the present invention disclosed in the accompanying claims and spirit, might make various modifications, augment and substitute.
Claims (1)1. fast method of distributing the optimum bit amount for the motion picture expert group audio data, it comprises the steps:
(a) covering ratio divided by 6 by the resulting signal of psychoacoustic model, the integer part of gained quotient is defined as an array pointer, and obtaining a significant bit sendout apace according to defined array pointer, wherein said step (a) further comprises the steps:
(a-1) give by covering by the resulting signal of psychoacoustic model and cover ratio to obtain a new signal than weighted value 1.01 in addition;
(a-2) whether check resulting new signal in above-mentioned steps (a-1) is covered than greater than zero, and when should be in above-mentioned steps (a-1) resulting new signal cover than less than zero the time value of the next inferior frequency band of check;
(a-3) finding in above-mentioned steps (a-1) resulting new signal when the check of above-mentioned steps (a-2) covers than greater than zero the time, resulting new signal in above-mentioned steps (a-1) is covered ratio divided by 6, the integer part of resulting quotient is defined as the array pointer, and from the signal to noise ratio that indicates, deduct resulting new signal coverage ratio in above-mentioned steps (a-1) by defined array pointer, cover noise ratio to obtain one;
(a-4) check in above-mentioned steps (a-3) resulting coverage noise ratio whether greater than zero;
(a-5) find that in above-mentioned steps (a-3) resulting coverage noise ratio is less than zero the time when the check of above-mentioned steps (a-4), then in above-mentioned steps (a-3), add 1 on the defined array pointer, to obtain a new array pointer, and, cover noise ratio to obtain one from covering ratio by deducting in above-mentioned steps (a-1) resulting new signal the specified signal to noise ratio of resulting new array pointer; With
(a-6) obtain the significant bit sendout, it is by the specified bit allocation amounts of defined array pointer in above-mentioned steps (a-3), or carrying out above-mentioned steps (a-5) the specified bit allocation amounts of resulting pointer in above-mentioned steps (a-5) afterwards, or find in above-mentioned steps (a-3) resulting coverage noise ratio corresponding bit allocation amounts greater than zero time when the check of above-mentioned steps (a-4); And
(b) resulting significant bit sendout and a fixing bit allocation amounts are compared, and obtain an optimum bit sendout according to comparative result, wherein said step (b) further comprises the steps:
(b-1) whether check resulting significant bit sendout in above-mentioned steps (a) is identical with the fixed bits assignment amount, and when resulting significant bit sendout is identical with the fixed bits assignment amount in above-mentioned steps (a), end operation;
(b-2) when the check of above-mentioned steps (b-1) finds that resulting significant bit sendout is greater than the fixed bits assignment amount in above-mentioned steps (a), find out an array pointer of covering noise ratio corresponding to maximum, from the array pointer that is found, deduct 1, to obtain a new array pointer, the bit allocation amounts that is indicated by resulting new array pointer is added on the fixed bits assignment amount, and from resulting addition result bit allocation amounts, deduct by the specified bit allocation amounts of array pointer that is found, to obtain the optimum bit sendout;
(b-3) when the check of above-mentioned steps (b-1) finds that resulting bit allocation amounts is less than the fixed bits assignment amount in above-mentioned steps (a), find out an array pointer of covering noise ratio corresponding to minimum, on the array pointer that is found, add 1, to obtain a new array pointer, being added on the fixed bits assignment amount by the specified bit allocation amounts of resulting new array pointer, and from resulting addition result bit allocation amounts, deduct by the specified bit allocation amounts of array pointer that is found, to obtain the optimum bit sendout; With
(b-4) from covering noise ratio corresponding to deducting by the resulting signal of psychoacoustic model in the specified signal to noise ratio of the array pointer of resulting optimum bit sendout in above-mentioned steps (b-2) or (b-3) by one, cover noise ratio to obtain one, and turn back to above-mentioned steps (b-1).
CN95121183A 1994-12-30 1995-12-29 Method for allocating optimum amount of bits to MPEG audio data at high speed Expired - Fee Related CN1114272C (en) Applications Claiming Priority (2) Application Number Priority Date Filing Date Title KR2019940038919U KR970003559Y1 (en) 1994-12-30 1994-12-30 Instrument core prevents vehicle heat deformation KR40389/1994 1994-12-31 Publications (2) Family ID=19405114 Family Applications (1) Application Number Title Priority Date Filing Date CN95121183A Expired - Fee Related CN1114272C (en) 1994-12-30 1995-12-29 Method for allocating optimum amount of bits to MPEG audio data at high speed Country Status (2) Families Citing this family (1) * Cited by examiner, â Cited by third party Publication number Priority date Publication date Assignee Title US8781822B2 (en) 2009-12-22 2014-07-15 Qualcomm Incorporated Audio and speech processing with optimal bit-allocation for constant bit rate applications Citations (2) * Cited by examiner, â Cited by third party Publication number Priority date Publication date Assignee Title US4516258A (en) * 1982-06-30 1985-05-07 At&T Bell Laboratories Bit allocation generator for adaptive transform coder US5040217A (en) * 1989-10-18 1991-08-13 At&T Bell Laboratories Perceptual coding of audio signalsRetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4