CN100339886C - Coding device capable of detecting transient position of sound signal and its coding method - Google Patents
Coding device capable of detecting transient position of sound signal and its coding method Download PDFInfo
- Publication number
- CN100339886C CN100339886C CNB031103685A CN03110368A CN100339886C CN 100339886 C CN100339886 C CN 100339886C CN B031103685 A CNB031103685 A CN B031103685A CN 03110368 A CN03110368 A CN 03110368A CN 100339886 C CN100339886 C CN 100339886C
- Authority
- CN
- China
- Prior art keywords
- sub
- data
- sampled data
- frequency
- reference sample
- 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
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
The present invention relates to a coding device which comprises a multiphase filter bank, a transient state detector and a coding processing unit. The coding device firstly carries out a sub band coding step and generates a plurality of sub band samples according to an input signal, and each sub band sample has a plurality of frequency sub bands. The coding device next carries out a selecting step and selects a plurality of sub band samples as reference sampled data, and the coding device determines the length of a data block of a window according to the sum of the energy of the frequency sub bands of the reference sampled data in a preset frequency range. The coding device finally carries out a transform coding step, the frequency sub bands are transformed by a preset conversion calculation method according to the data of the window determined by the selected step, and an output signal is generated.
Description
Technical field
The invention provides a kind of scrambler, refer to a kind of scrambler that can detect the transient position of voice signal especially.When can also further judging Frequency Domain Coding, uses by scrambler of the present invention the block length of window data.
Background technology
There are many scramblers to adopt special coding algorithm according to human auditory system's characteristic at present, can be with more than the digital audio signal data compression to ten times, as MP3, AAC, WWA and Dolby Digital etc., these scramblers have adopted technology such as consciousness coding, Frequency Domain Coding, form switching and dynamic bit distribution to eliminate content unnecessary in the original sound signal data.
Consciousness coding be by eliminate general human auditory system impression less than sound signal data compress.In general, the human sound frequency that can hear is about 20Hz between the 20kHz, and the general mankind of the sound of other frequency domains be experience less than.On the other hand, human auditory system also can produce the shielding (mask) of the sense of hearing in some cases, and can't tell the noise of quantification, for example when the outstanding especially sound of a volume or tone color occurs, its contiguous tiny sound can difficultly be discovered, and does not therefore need all sound details are all compiled into when coding.
Frequency Domain Coding is a kind of method that can effectively eliminate unnecessary data, to there be the time domain data of very strong correlation to be transformed in the almost incoherent frequency domain of each element, remove except that content unnecessary in the data, generally can be divided into transition coding or subband (subband) coding.The frequency spectrum resolution of transition coding is higher, and the resolution of sub-band coding is low but efficient is higher, so these two kinds of codings can be combined into a compound filter, at the different frequency place different resolutions is arranged.Yet Frequency Domain Coding has a significant phenomenon to be called forward echo (preechoes), for instance, if occur very big sound after general quiet suddenly, may make quantization error increase.In transition coding and sub-band coding, all can produce this phenomenon, cause data the forward echo of sound after changing back time domain, to occur.
A kind of method of eliminating forward echo be with error limitation in a less time period, the other parts of sound and forward echo are separated, forward echo is resulted among the shadow zone.In a less time period, need to use less piece to carry out frequency domain transform error limitation, this method is called form and switches, when signal stabilization, use bigger piece to carry out Frequency Domain Coding, and when signal has significantly transient state (Transient), carry out Frequency Domain Coding with regard to using less piece.The shortcoming that form switches needs more figure place when being the expression identical data, because along with the increase of coded data quantity needs more information.
Scrambler whether have good coding quality, with the very big relation that is assigned between each subband or coefficient.For dividing coordination effectively, must constantly analyze input signal, and, will be assigned to people's the most effective zone of the sense of hearing than multidigit according to the model that human auditory system's knowledge is set up, just need not distribute or only distribute bits of coded seldom in the insensitive zone of people's ear.Because signal does not stop to change, people's auditory system also has different reactions to signal under different condition, the technology that dynamic bit that Here it is is distributed.Good position allocative decision needs accurate psychoacoustic model (psychoacoustic model).
Please refer to Fig. 1, Fig. 1 is the synoptic diagram of known MPEG layer-3 sound signal encoding.At first, pulse code modulated (pulse code modulation, PCM) input signal 10 is divided into 32 wide frequency subbands (frequencysubbands) by multiphase filter group (polyphase filter bank) 12, multiphase filter group 12 easily analysis frequency to time relation, but wide frequency subband can not reflect human auditory system's auditory properties exactly, in addition, contiguous frequency subband has more lap, (modified discrete cosine transform MDCT) 14 compensates so the output of multiphase filter group 12 need be used the correction discrete cosine transform.Revise discrete cosine transform 14 and further frequency subband is done segmentation, obtaining frequency spectrum resolution preferably, and can be with some overlapping eliminating that is produced by multiphase filter group 12.Revise the form piece that discrete cosine transform 14 comprises two different lengths, be respectively long piece and one six short block of sampling of one 18 sampling because continuous transfer form piece have 50 percent overlapping, so the length of piece is respectively 36 and 12.When voice signal was stablized, long piece had higher frequency resolution degree to reach compressibility preferably, and short block then provides temporal analytical density preferably.Because the temporal analytical density of long piece is lower, if in the piece of handling transition effect takes place, (Quantization Noise) can be diffused into whole because of quantizing noise, make the less signal of energy can't cover quantizing noise and produce distortion, as forward echo because of own shielding effect (Mask) is low.For avoiding forward echo, known MPEG sound signal encoding applied mental acoustic model 16 detects transient state (Transient) position of voice signal, avoids forward echo to use short block to revise discrete cosine transform 14.After using the technology of Frequency Domain Coding to be transformed into frequency domain input signal 10, then carry out quantification program 18, come quantized data according to psychoacoustic model 16, carry out canned program 20 then, with the output signal 22 of output data bit stream (bitstream) after the data encapsulation.
From the above, when carrying out Frequency Domain Coding, for avoiding forward echo, it is a kind of technology commonly used that form switches, and the mechanism that at this moment detects the voice signal transient position is just very important.Known MPEG sound signal encoding applied mental acoustic model 16 detects the transient position of voice signal, though it is very accurate, but because psychological acoustic model 16 is quite complicated, required cost is also very high, if use expensive psychoacoustic model 16, be quite uneconomic because use the form switching to need to detect the transient position of voice signal.
Summary of the invention
Therefore fundamental purpose of the present invention provides a kind of scrambler that detects the voice signal transient position.On the other hand, the present invention also provides a kind of scrambler and coding method of using the block length of window data when judging Frequency Domain Coding, to address the above problem.
The present invention provides a kind of scrambler, is used for input signal is encoded to output signal.This scrambler comprises the multiphase filter group, is used for producing a plurality of sub-band samples according to this input signal, and different sub-band samples is corresponding to the waveform input signal of different periods, and comprises a plurality of frequency subbands in each sub-band samples; Transient detector is connected to this multiphase filter group, and the block length with deciding window data includes a plurality of weighted values in this window data, and this transient detector comprises the subband selector switch, is used for selecting these a plurality of sub-band samples as the reference sampled data; Energy calculator is connected to this subband selector switch, is used for calculating the energy summation of this reference sample data medium frequency subband; Zonal device is connected between this subband selector switch and this energy calculator, is used for these reference sample data are divided into the array sub-sampled data, and each group sub-sampled data comprises at least one sub-band samples; And comparer, be connected to this energy calculator, be used for the output valve and first critical value of energy calculator are made comparisons, represent the signal of the block length of window data according to this comparative result output; And coding processing unit, be connected to this multiphase filter group and this transient detector, be used for will these a plurality of frequency subbands multiply by a plurality of weighted values in this transient state window data to produce weighted results, produce this output signal with default conversion calculus method according to this weighted results again.
The present invention provides a kind of coding method in addition, is used for input signal is encoded to output signal.This coding method includes carries out the sub-band coding step, and producing a plurality of sub-band samples according to this input signal, different sub-band samples is corresponding to the waveform input signal of different periods, and comprises a plurality of frequency subbands in each sub-band samples; Select step,, include a plurality of weighted values in this window data so that the window data corresponding to default block length to be provided; And include in this selection step: in these a plurality of sub-band samples, select a plurality of sub-band samples, and decide the block length of this window data according to the energy summation of the frequency subband of these reference sample data in the predeterminated frequency scope as the reference sampled data; And carry out the transition coding step, a plurality of weighted values that these a plurality of frequency subbands be multiply by the window data that this selection step determined to be producing weighted results, and produce this output signal with default conversion calculus method according to this weighted results.
Relative known technology, the invention provides the block length of the window data of using when a kind of scrambler and coding method can be used to determine to revise discrete cosine transform, utilize the contained energy value of the sub-band samples medium frequency subband that produced in the process of scrambler for judging whether sound signal data transient state takes place, need lower cost more than known applied mental acoustic model, meet economic benefit.
Description of drawings
Fig. 1 is the synoptic diagram of known MPEG layer-3 sound signal encoding.
Fig. 2 is the synoptic diagram of the scrambler of the embodiment of the invention.
Fig. 3 is the synoptic diagram of the sub-band samples of present embodiment.
Fig. 4 is the process flow diagram that scrambler detects the transient position method of voice signal in the embodiment of the invention.
The reference numeral explanation
10 input signals, 12 multiphase filter groups
14 revise discrete cosine transform 16 psychoacoustic models
18 quantification programs, 20 canned programs
22 output signals, 30 scramblers of the present invention
32 transient detector, 34 coding processing unit
36 subband selector switchs, 38 energy calculator
40 zonal devices, 42 comparers
50 reference sample data
Embodiment
Please refer to Fig. 2, Fig. 2 is the synoptic diagram of the scrambler 30 of the embodiment of the invention.Scrambler 30 is used for the input signal 10 of pulse code modulated is encoded to the output signal 22 of bit stream.Scrambler 20 comprises multiphase filter group 12, transient detector 32 and coding processing unit 34.Multiphase filter group 12 produces a plurality of sub-band samples according to this input signal 10, and different sub-band samples is corresponding to input signal 10 waveforms of different periods, and comprises a plurality of frequency subbands in each sub-band samples.Coding processing unit 34 can be revised discrete cosine transform to these a plurality of frequency subbands.Transient detector 32 is connected between multiphase filter group 12 and the coding processing unit 34, can determine the block length of employed window data when coding processing unit 34 is revised discrete cosine transform.Transient detector 32 comprises subband selector switch 36, energy calculator 38, zonal device 40 and comparer 42.Subband selector switch 36 can select the sub-band samples of part in these a plurality of sub-band samples as the reference sampled data in the predeterminated frequency scope, then energy calculator 38 can be calculated contained energy value in the reference sample data, afterwards this energy value is given comparer 42 and is made comparisons with critical value.When if the gross energy of reference sample data surpasses this critical value, the situation that just in the reference sample data, may have transient state, then by zonal device 40 the reference sample data are divided into the wide sub-sampled data of array again, and each group sub-sampled data comprises a sub-band samples at least, this moment, the energy difference of the frequency subband of two adjacent groups sub-sampled data in the predeterminated frequency scope was calculated in energy calculator 38 meetings, this energy difference was sent to comparer 42 again and made comparisons with predetermined critical value.If this energy difference during greater than predetermined critical value, then can determine coding processing unit 34 to use the window data of short blocks to revise discrete cosine transform, so repetitiousness is finished all possible sub-sampled data combination up to zonal device 42.If the energy difference of the sub-sampled data of two adjacent groups is still less than predetermined critical value at this moment, then can determine coding processing unit 34 to use the window data of long piece to revise discrete cosine transform.
Please refer to Fig. 3, Fig. 3 is the synoptic diagram of the sub-band samples of present embodiment.Multiphase filter group 12 is exported 18 sub-band samples in a period t1, contain 32 frequency subbands in each sub-band samples.Each frequency subband in 34 pairs of overlapping periods of coding processing unit is revised discrete cosine transform, just 36 sub-band samples.Transient detector 32 is done to detect with decision coding processing unit 34 at the position that voice signal transient state takes place and should be used which kind of form piece to revise discrete cosine transform.So-called predeterminated frequency scope is commonly referred to as between the frequency between subband and coding restriction subband, and subband selector switch 36 can select the frequency subband in this frequency range to be used as reference sample data 50.By subband can be rule of thumb or experiment value select first subband or the subband of high frequency more.In the present embodiment, the frequency by subband is approximately 4kHz.Coding restriction subband just must decide according to coding rule.Because bit rate (bitrate) and bandwidth (bandwidth) all have its restriction, scrambler 30 must be given up the information of part high-frequency sub-band, and the data of the frequency subband that is rejected are just no longer listed consideration in.Suppose not have information to be rejected, then last subband restriction subband of encoding exactly.After reference sample data 50 are selected, energy calculator 38 can calculate energy value contained in the reference sample data 50, judge whether reference sample data 50 are continued to detect by comparer 42 again, zonal device 40 can be divided into reference sample data 50 the wide sub-sampled data of array again, energy calculator 38 can be calculated the energy difference of two adjacent groups sub-sampled data then, by the block length of comparer 42 decision window data.For instance, at first energy calculator 38 is calculated the gross energy of all frequency subbands in the reference sample data 50 that subband selector switchs 36 select, if gross energy is greater than-60dB, the situation that then may have transient state in the reference sample data takes place, by zonal device 40 sub-band samples in the reference sample data 50 is divided into six groups of wide sub-sampled data, then giving comparer 42 by the energy difference of energy calculator 38 calculating two adjacent groups sub-sampled data compares, if the energy difference of two sub-sampled data is not greater than 20dB, represent the also situation of non-transient generation in fact between this two this sub-sampled data, zonal device 40 can be divided into the sub-band samples in the reference sample data 50 3 groups of wide sub-sampled data again, and the energy difference of calculating the two adjacent groups sub-sampled data by energy calculator 38 again this moment is given comparer 42 and judged whether greater than 12dB.If greater than 12dB, then represent to contain in the data situation of transient state, therefore judge and should use the short block form; If not greater than 12dB, then use long piece form.
Please refer to Fig. 4, Fig. 4 is in the embodiment of the invention, and scrambler 30 detects the process flow diagram of the method for voice signal transient position.The coding method of present embodiment can detect the transient position of voice signal.The sub-band coding step is at first carried out in the coding method of present embodiment, produces a plurality of sub-band samples according to input signal 10, and different sub-band samples is corresponding to input signal 10 waveforms of different periods, and comprises a plurality of frequency subbands in each sub-band samples.Then select step, block length with the window data of decision next step required use, contain a plurality of weighted values in the window data, selecting the method for step is in these a plurality of sub-band samples, select a plurality of sub-band samples as the reference sampled data, and decide the block length of this window data according to the energy summation of the frequency subband of reference sample data in the predeterminated frequency scope.Carry out the transition coding step at last, these a plurality of frequency subbands be multiply by a plurality of weighted values of selecting the window data that step determined producing weighted results, and use according to weighted results and to revise discrete cosine transform and produce output signal.And the detailed step that detects the voice signal transient position is as follows:
Step 110: the transient position that begins to detect voice signal;
Step 120: calculate to select as with reference to the gross energy of the frequency subband in the sampled data whether greater than predetermined critical value, if, then carry out step 130, if not, then carry out step 170;
Step 130: the reference sample data are divided into the wide sub-sampled data of array, each group sub-sampled data comprises more than one sub-band samples, calculate all energy value of frequency subband in the predeterminated frequency scope in each group sub-sampled data, then carry out step 140;
Step 140: whether the energy difference of judging the two adjacent groups sub-sampled data greater than predetermined critical value, if, then carry out step 160, if not, then carry out step 150;
Step 150: judge whether the reference sample data can also be divided into different sub-sampled data, if, then get back to step 130, if not, then carry out step 170;
Step 160: contain transient position in the reference sample data, the window data signal of short block is used in output, carry out step 180;
Step 170: do not contain transient position in the reference sample data, the window data signal of long piece is used in output, carry out step 180;
Step 180: output judged result, the transient position of detection of end voice signal.
Relative known technology, the invention provides the block length of the window data of using when a kind of scrambler and coding method can be used to determine to revise discrete cosine transform, utilize the contained energy value of the sub-band samples medium frequency subband that produced in the process of scrambler for judging whether sound signal data transient state takes place, need lower cost more than known applied mental acoustic model, meet economic benefit.
The above only is the preferred embodiments of the present invention, and all similar variation and improvement of doing according to claims of the present invention all should belong to the covering scope of patent of the present invention.
Claims (18)
1. a coding method is used for input signal is encoded to output signal, and this method includes:
Carry out the sub-band coding step, producing a plurality of sub-band samples according to this input signal, different sub-band samples is corresponding to the waveform input signal of different periods, and comprises a plurality of frequency subbands in each sub-band samples;
Select step,, include a plurality of weighted values in this window data so that the window data corresponding to default block length to be provided;
And include in this selection step:
In these a plurality of sub-band samples, select a plurality of frequency subbands as the reference sampled data according to a predeterminated frequency scope, and decide the block length of this window data by the energy summation of the frequency subband of these reference sample data in this predeterminated frequency scope is compared with one first critical value; And
Carry out the transition coding step, a plurality of weighted values that a plurality of frequency subbands in these reference sample data be multiply by the window data that this selection step determined to be producing weighted results, and produce this output signal with default conversion calculus method according to this weighted results.
2. coding method as claimed in claim 1, wherein when carrying out this selection step, if the energy summation of the frequency subband in these reference sample data then compares step in addition greater than first critical value, it comprises:
These reference sample data are divided into the array sub-sampled data, and each group sub-sampled data comprises at least one frequency subband; And
Calculate the energy magnitude difference of the frequency subband in the two adjacent groups sub-sampled data, if this difference greater than second critical value, then when this transition coding step, is used the window data of short block length; And
If the energy magnitude difference of the frequency subband in the two adjacent groups sub-sampled data is less than or equal to this second critical value, then carry out another time comparison step, and the sub-sampled data in relatively once before the frequency subband that sub-sampled data contained in this comparison step is different from.
3. coding method as claimed in claim 2 is wherein if the energy summation of frequency subband in these reference sample data during less than this first critical value, then when this transition coding step, is used the window data of a long block length.
4. coding method as claimed in claim 1, wherein this input signal is to be pulse code modulated signal.
5. coding method as claimed in claim 1, wherein this output signal is to be coding stream.
6. coding method as claimed in claim 1 should default conversion calculus method be for revising discrete cosine transform wherein.
7. a scrambler is used for input signal is encoded to output signal, and it comprises:
The multiphase filter group is used for producing a plurality of sub-band samples according to this input signal, and different sub-band samples is corresponding to the waveform input signal of different periods, and comprises a plurality of frequency subbands in each sub-band samples;
Transient detector, be connected to this multiphase filter group, with the block length that decides window data, include a plurality of weighted values in this window data, this transient detector comprises: the subband selector switch is used for selecting a plurality of frequency subbands as the reference sampled data from a plurality of sub-band samples according to a predeterminated frequency scope; Energy calculator is connected to this subband selector switch, is used for calculating the energy summation of this reference sample data medium frequency subband; Zonal device is connected between this subband selector switch and this energy calculator, is used for these reference sample data are divided into the array sub-sampled data, and each group sub-sampled data comprises at least one frequency subband; And comparer, be connected to this energy calculator, be used for the output valve and first critical value of energy calculator are made comparisons, represent the signal of the block length of window data according to this comparative result output; And
Coding processing unit, be connected to this multiphase filter group and this transient detector, be used for a plurality of frequency subbands in these reference sample data be multiply by a plurality of weighted values in this transient state window data to produce weighted results, produce this output signal with default conversion calculus method according to this weighted results again.
8. scrambler as claimed in claim 7, wherein this energy calculator can be calculated the energy magnitude difference of two adjacent groups sub-sampled data medium frequency subband, again the result is sent to this comparer and second critical value is made comparisons.
9. scrambler as claimed in claim 8, wherein this zonal device is according to the comparative result of this comparer, these reference sample data are divided into the sub-sampled data of array in addition, and contained frequency subband is different from frequency subband contained in the previous sub-sampled data in each group sub-sampled data.
10. scrambler as claimed in claim 7, wherein this input signal is to be pulse code modulated signal.
11. scrambler as claimed in claim 7, wherein this output signal is to be coding stream.
12. scrambler as claimed in claim 7 should default conversion calculus method be for revising discrete cosine transform wherein.
13. a method that detects voice signal transient state when carrying out sound signal encoding, this method comprises:
(a) produce a plurality of sub-band samples according to this voice signal, different sub-band samples is corresponding to the sound signal waveform of different periods, and comprises a plurality of frequency subbands in each sub-band samples;
(b) in these a plurality of sub-band samples, select a plurality of frequency subbands as the reference sampled data according to a predeterminated frequency scope, and according to the energy summation of the frequency subband of this reference sample data computation in this predeterminated frequency scope;
(c) if the energy summation of the frequency subband in these reference sample data greater than first critical value, is divided into the array sub-sampled data with these reference sample data, each group sub-sampled data comprises at least one frequency subband, and execution in step (d);
On the contrary,, judge then in these reference sample data not have voice signal transient state that processing finishes if the energy summation of the frequency subband in these reference sample data is not more than this first critical value;
(d) calculate the energy magnitude difference of the frequency subband in the two adjacent groups sub-sampled data, and judge this voice signal sound intermediate frequency transient state part according to this difference.
14. method as claimed in claim 13, when wherein judging voice signal transient state part when carrying out step (d) and according to this difference, if this difference is greater than second critical value, judge that then pairing sound signal waveform is the waveform of transient state between these two groups of sub-sampled data, and if this difference less than this second critical value, then these reference sample data are divided into the sub-sampled data that array differs from step (c), carry out step (d) once more.
15. transient detector that is arranged in the voice coder, be used for detecting the voice signal of importing this scrambler and whether comprise transient state, this voice coder comprises the multiphase filter group, be used for producing a plurality of sub-band samples according to this input signal, different sub-band samples is corresponding to the waveform input signal of different periods, and comprise a plurality of frequency subbands in each sub-band samples, this transient detector is connected to this multiphase filter group, and comprises:
The subband selector switch is used for selecting a plurality of frequency subbands as the reference sampled data according to a predeterminated frequency scope from these a plurality of sub-band samples;
Energy calculator is connected to this subband selector switch, is used for calculating the energy summation of this reference sample data medium frequency subband;
Zonal device is connected between this subband selector switch and this energy calculator, is used for these reference sample data are divided into the array sub-sampled data, and each group sub-sampled data comprises at least one frequency subband; And
Comparer is connected to this energy calculator, is used for the output valve and first critical value of energy calculator are made comparisons, and judges according to this comparative result whether this voice signal of this scrambler of input comprises transient state.
16. transient detector as claimed in claim 15, wherein this energy calculator can be calculated the energy magnitude difference of two adjacent groups sub-sampled data medium frequency subband, again the result is sent to this comparer and second critical value is made comparisons.
17. transient detector as claimed in claim 16, wherein this zonal device is according to the comparative result of this comparer, these reference sample data are divided into the sub-sampled data of array in addition, and contained frequency subband is different from frequency subband contained in the previous sub-sampled data in each group sub-sampled data.
18. transient detector as claimed in claim 15, wherein this voice signal is to be pulse code modulated signal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB031103685A CN100339886C (en) | 2003-04-10 | 2003-04-10 | Coding device capable of detecting transient position of sound signal and its coding method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB031103685A CN100339886C (en) | 2003-04-10 | 2003-04-10 | Coding device capable of detecting transient position of sound signal and its coding method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1536559A CN1536559A (en) | 2004-10-13 |
| CN100339886C true CN100339886C (en) | 2007-09-26 |
Family
ID=34319676
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB031103685A Expired - Fee Related CN100339886C (en) | 2003-04-10 | 2003-04-10 | Coding device capable of detecting transient position of sound signal and its coding method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100339886C (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101246689B (en) * | 2004-09-17 | 2011-09-14 | 广州广晟数码技术有限公司 | Audio encoding system |
| US8204261B2 (en) * | 2004-10-20 | 2012-06-19 | Fraunhofer-Gesellschaft Zur Foerderung Der Angewandten Forschung E.V. | Diffuse sound shaping for BCC schemes and the like |
| DE102006047197B3 (en) * | 2006-07-31 | 2008-01-31 | Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. | Device for processing realistic sub-band signal of multiple realistic sub-band signals, has weigher for weighing sub-band signal with weighing factor that is specified for sub-band signal around subband-signal to hold weight |
| CN101308655B (en) * | 2007-05-16 | 2011-07-06 | 展讯通信（上海）有限公司 | Audio coding and decoding method and layout design method of static discharge protective device and MOS component device |
| CN101308651B (en) * | 2007-05-17 | 2011-05-04 | 展讯通信（上海）有限公司 | Detection method of audio transient signal |
| US8630848B2 (en) | 2008-05-30 | 2014-01-14 | Digital Rise Technology Co., Ltd. | Audio signal transient detection |
| EP2214165A3 (en) * | 2009-01-30 | 2010-09-15 | Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. | Apparatus, method and computer program for manipulating an audio signal comprising a transient event |
| JP5799707B2 (en) * | 2011-09-26 | 2015-10-28 | ソニー株式会社 | Audio encoding apparatus, audio encoding method, audio decoding apparatus, audio decoding method, and program |
| EP2954650A1 (en) * | 2013-02-05 | 2015-12-16 | Interdigital Patent Holdings, Inc. | Pulse-shaped orthogonal frequency division multiplexing |
| CN106683687B (en) * | 2016-12-30 | 2020-02-14 | 杭州华为数字技术有限公司 | Abnormal sound classification method and device |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5197087A (en) * | 1989-07-19 | 1993-03-23 | Naoto Iwahashi | Signal encoding apparatus |
| CN1093843A (en) * | 1993-02-02 | 1994-10-19 | 索尼公司 | The method of high efficient coding and decoding and device |
| CN1125010A (en) * | 1993-04-22 | 1996-06-19 | 弗兰克·乌达尔·莱昂哈德 | Method and system for detecting and generating transient conditions in auditory signals |
| JPH08286699A (en) * | 1995-04-14 | 1996-11-01 | Tech Res & Dev Inst Of Japan Def Agency | Transient sound frequency analyzing methdo and its device |
| CN1208489A (en) * | 1995-12-01 | 1999-02-17 | 数字剧场系统股份有限公司 | Multi-channel predictive subband coder using psychoacoustic adaptive bit allocation |
| CN1312977A (en) * | 1998-05-27 | 2001-09-12 | 微软公司 | Scalable audio coder and decoder |
| CN1361594A (en) * | 2000-12-25 | 2002-07-31 | 松下电器产业株式会社 | Equipment and method for coding frequency signal and computer program products |
-
2003
- 2003-04-10 CN CNB031103685A patent/CN100339886C/en not_active Expired - Fee Related
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5197087A (en) * | 1989-07-19 | 1993-03-23 | Naoto Iwahashi | Signal encoding apparatus |
| CN1093843A (en) * | 1993-02-02 | 1994-10-19 | 索尼公司 | The method of high efficient coding and decoding and device |
| CN1125010A (en) * | 1993-04-22 | 1996-06-19 | 弗兰克·乌达尔·莱昂哈德 | Method and system for detecting and generating transient conditions in auditory signals |
| JPH08286699A (en) * | 1995-04-14 | 1996-11-01 | Tech Res & Dev Inst Of Japan Def Agency | Transient sound frequency analyzing methdo and its device |
| CN1208489A (en) * | 1995-12-01 | 1999-02-17 | 数字剧场系统股份有限公司 | Multi-channel predictive subband coder using psychoacoustic adaptive bit allocation |
| CN1312977A (en) * | 1998-05-27 | 2001-09-12 | 微软公司 | Scalable audio coder and decoder |
| CN1361594A (en) * | 2000-12-25 | 2002-07-31 | 松下电器产业株式会社 | Equipment and method for coding frequency signal and computer program products |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1536559A (en) | 2004-10-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10446162B2 (en) | System, method, and non-transitory computer readable medium storing a program utilizing a postfilter for filtering a prefiltered audio signal in a decoder | |
| KR101278805B1 (en) | Selectively using multiple entropy models in adaptive coding and decoding | |
| US7684981B2 (en) | Prediction of spectral coefficients in waveform coding and decoding | |
| US7693709B2 (en) | Reordering coefficients for waveform coding or decoding | |
| US7613603B2 (en) | Audio coding device with fast algorithm for determining quantization step sizes based on psycho-acoustic model | |
| US6240380B1 (en) | System and method for partially whitening and quantizing weighting functions of audio signals | |
| US6253165B1 (en) | System and method for modeling probability distribution functions of transform coefficients of encoded signal | |
| CN101030373B (en) | System and method for stereo perceptual audio coding using adaptive masking threshold | |
| US6029126A (en) | Scalable audio coder and decoder | |
| JP4676139B2 (en) | Multi-channel audio encoding and decoding | |
| EP1080579B1 (en) | Scalable audio coder and decoder | |
| KR100348368B1 (en) | A digital acoustic signal coding apparatus, a method of coding a digital acoustic signal, and a recording medium for recording a program of coding the digital acoustic signal | |
| US20040181403A1 (en) | Coding apparatus and method thereof for detecting audio signal transient | |
| EP2054882A2 (en) | Arbitrary shaping of temporal noise envelope without side-information | |
| KR19990041072A (en) | Stereo Audio Encoding / Decoding Method and Apparatus with Adjustable Bit Rate | |
| CN100339886C (en) | Coding device capable of detecting transient position of sound signal and its coding method | |
| KR20030068716A (en) | Method for compressing audio signal using wavelet packet transform and apparatus thereof | |
| Malvar | Enhancing the performance of subband audio coders for speech signals | |
| Mandal et al. | Digital Audio Compression | |
| Malvar | Perceptual Audio Coding |
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: 20070926 Termination date: 20200410 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |