CN100527849C - Audio-frequency decoding method, audio-frequency decoder and mooth filtering-wave method therein - Google Patents
Audio-frequency decoding method, audio-frequency decoder and mooth filtering-wave method therein Download PDFInfo
- Publication number
- CN100527849C CN100527849C CNB2005101110847A CN200510111084A CN100527849C CN 100527849 C CN100527849 C CN 100527849C CN B2005101110847 A CNB2005101110847 A CN B2005101110847A CN 200510111084 A CN200510111084 A CN 200510111084A CN 100527849 C CN100527849 C CN 100527849C
- Authority
- CN
- China
- Prior art keywords
- sampled value
- interpolation
- filter
- territory
- decoding unit
- 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
Abstract
The audio decoder includes following parts: input buffer in use for receiving external essential stream (ES); decoding unit in use for decoding ES to obtain sampled value in spatial (S) domain; carrying out discrete cosine transform (DCT) and adding window process for sampled value in S domain so as to obtain filter for PCM sampled value; output buffer in use for outputting PCM sampled value. The decoding unit also receives interpolation input to filter from underflow signal fed back from the output buffer. In the invention, smoothed filter carries out before IDCT transform; mute or playback signal are inserted in sampled values in frequency domain, and smoothed filtering is carried out by multiplexing synthesized filter. Thus, the invention not only filters out high frequency interference, but also gets quit of need of additional filter and buffer; moreover, the invention saves circuit area, and reduces cost.
Description
Technical field
The present invention relates to audio-frequency decoding method and audio decoder, relate in particular to a kind of audio decoder of The disposal of gentle filter and smooth filtering method in the audio decoder of carrying out.
Background technology
The digital audio/video compress technique is very important research topic in the multimedia field, and the defined audio frequency and video compress technique of mpeg standard is used widely in media product.Mpeg standard is set up by the Motion Picture Experts Group (Moving Pictures Experts Group) of the ISO of International Standards Organization, its audio standard comprises three kinds of different encryption algorithms, it is the MPEG ground floor, the 3rd layer of the MPEG second layer and MPEG, three kinds of encryption algorithm complexity differences, thereby compression ratio is also inequality.Audio signal can be sampled and encode usually according to a kind of algorithm wherein.Audio decoder is used for the sampled value of having encoded that receives is decoded, and generally includes buffer, resolver, reconfiguration unit and filter etc.Buffer is transported to audio decoder with database stream from internal memory, and resolver is resolved the data flow of master data flow pattern, and the data that parsing obtains are sent into reconfiguration unit, obtains sampled value by the reconfiguration unit decoding.S territory sampled value when the sampled value of this moment is coding after the discrete cosine transform (DCT), need carry out inverse discrete cosine transform (IDCT) to it during decoding, obtain digital pulse coding modulation (PCM) sampling point, at last the PCM sampling point is converted to analog signal, be i.e. the sound that can be accepted for people's class.The general using filter carries out IDCT to be handled, and finishes the conversion of data flow from frequency domain to time domain.
Sometimes may in transfer of data, produce some mistake, or in the decode procedure of audio decoder, go out mistake or the like, at this moment, often need to take some measures the signal of sheltering mistake.Insert mute signal in the PCM sampling point that normally behind idct transform, obtains.But mute signal can make the signal of output discontinuous, and shown in the circle among Fig. 1 11, high-frequency percussion is introduced in the transition that the mute signal that inserts in time domain can produce waveform at boundary, makes the sound of the output shrillness that becomes, and people's ear is caused damage.The audio decoder that also has is the PCM sampled value that obtains behind idct transform when makeing mistakes, the repeat playing preceding last signal of makeing mistakes.As the circle among Fig. 2 21 places pointing out that this method can produce transition at the boundary of waveform equally, and make the sound of the output shrillness that becomes, make us unhappy.
For this High-frequency Interference of filtering, the audio decoder that has adds a series of reprocessing before the laggard line output of idct transform is play, and wherein comprise and insert mute signal or replay signal, and the operation of this High-frequency Interference of filtering.Adopt this processing mode need in post-processing module, additionally increase the one-level low pass filter usually and come this level and smooth filtering of special disposal, and need the additional configuration corresponding buffers.As shown in Figure 3, comprise one and door 31 in the post-processing module, smoothing filter 33 and level and smooth filtered buffer 35.Decoder 30 is resolved and is carried out idct transform, obtains the PCM sampled value.When system discovery was made mistakes, decoder was set as low level to mute signal 36 or replay signal 37, and mute signal or replay signal are by replacing PCM sampling point 38 under the error situation with door.The PCM sampling point temporarily leaves in the smothing filtering buffer, exports after smoothing filter filtering.
Audio decoder shown in Figure 3 carries out smothing filtering in time domain to the sampled value under the error situation, though to a certain extent filtering High-frequency Interference owing to increased extra filter and corresponding buffers, increased hardware, cost can't be reduced.
Summary of the invention
The objective of the invention is to overcome the defective that exists in the prior art, a kind of audio-frequency decoding method and audio decoder are provided, it can carry out The disposal of gentle filter, and the High-frequency Interference that filtering mute signal or replay signal are brought does not need to increase extra hardware for smothing filtering simultaneously.
According to an aspect of the present invention, provide a kind of audio decoder, comprising:
In order to receive the input buffer of outside elementary stream ES;
Described elementary stream ES is decoded, obtain the decoding unit of S territory sampled value;
Described S territory sampled value is carried out inverse discrete cosine transform and windowing (window) processing, obtain the filter of PCM sampled value; And
In order to exporting the output buffer of described PCM sampled value,
Wherein, described decoding unit also receives the underflow signal of output buffer feedback to described filter input interpolation, and described filter carries out the conversion of frequency domain to time domain to described interpolation and described S territory sampled value.
In one embodiment of the invention, decoding unit is provided with interpolation counter and interpolating unit.Output buffer is checked through underflow condition, then feeds back an interpolated signal and gives decoding unit.Decoding unit is imported interpolation by interpolating unit to filter under the control of interpolation counter, filter carries out the conversion of frequency domain to time domain to interpolation, finishes smothing filtering.
Interpolating unit for example comprises multiplexer, has two input values at least, and an input value is a S territory sampled value, and an input value is an interpolation.Multiplexer carries out gating to two input values under the control of interpolation counter.The interpolation counter generally is in reset mode, and when receiving the interpolated signal that output buffer sends here, the gating interpolation is as input, and begins counting simultaneously; When count value equals filter channel and counts, reset again.After resetting, multiplexer again the S territory sampled value of gating decoder as input.
In one embodiment of the invention, filter for example is the subband synthesis filter of 32 passages, when the response speed of resources such as bus, processor goes wrong, can be together to the S territory sampled value of decoder output and interpolation, carry out idct transform and windowing process, finish the conversion of S territory sampled value and interpolation, obtain the PCM sampled value from frequency domain to time domain.The interpolation counter for example is made of one 5 register and one 5 adder, can carry out 32 countings.
According to a further aspect in the invention, provide a kind of audio-frequency decoding method, comprising:
The outside elementary stream ES that receives is decoded, obtain S territory sampled value;
Described S territory sampled value is carried out inverse discrete cosine transform and windowing process, output PCM sampled value;
When the data volume of the PCM sampled value that detects output reaches underflow condition, provide interpolation;
Described interpolation is carried out inverse discrete cosine transform and windowing process with described S territory sampled value.
According to a further aspect in the invention, smooth filtering method in a kind of audio decoder is provided, described audio decoder comprises the outside elementary stream ES that receives decoded and obtains the decoding unit of S territory sampled value, described S territory sampled value is carried out inverse discrete cosine transform and windowing process, obtain the filter of PCM sampled value, and in order to export the output buffer of described PCM sampled value, described method comprises:
1. when the data volume of output buffer reduces to underflow condition, send underflow signal and give decoding unit;
2. decoding unit send interpolation to give filter under the triggering of underflow signal.
3. filter S territory sampled value and interpolation that decoding unit is sent into carried out idct transform and windowing process together.
The present invention carried out smothing filtering before the IDCT of audio decoder conversion, sampled value is inserted quiet or replay signal when the frequency domain, and multiplexing composite filter carries out smothing filtering.Thereby not only filtering High-frequency Interference, and exempted demand to extra smoothing filter and buffer, saved circuit area, reduced cost.
Description of drawings
Following with reference to accompanying drawing describing to exemplary embodiment of the present invention, to the elaboration of the embodiment of the invention, be to disclose feature of the present invention place, but do not limit the present invention in conjunction with the following drawings for further, same-sign is represented respective element or step among the embodiment among the figure, wherein:
Fig. 1 is prior art sound intermediate frequency decoder inserts mute signal under error situation a oscillogram.
Fig. 2 is prior art sound intermediate frequency decoder inserts replay signal under error situation a oscillogram.
Fig. 3 is for carrying out the audio decoder structural representation of smothing filtering in the prior art.
Fig. 4 A can carry out the structural representation of the audio decoder of smothing filtering for the present invention.
Fig. 4 B is the structural representation of a decoding unit in the audio decoder shown in Fig. 4 A.
Fig. 5 is the principle schematic of the output buffer of audio decoder of the present invention.
Fig. 6 is the interpolating unit structural representation of the decoding unit in the one embodiment of the invention.
Fig. 7 is the oscillogram after audio decoder of the present invention inserts interpolation.
Fig. 8 is the flow chart of audio-frequency decoding method of the present invention.
Fig. 9 is the flow chart of smooth filtering method in the audio decoder of the present invention.
Embodiment
Referring to Fig. 4 A and Fig. 4 B, audio decoder of the present invention comprises input buffer 41, decoding unit 43, filter 45 and output buffer 47.
Input buffer receives the ES stream (elementary stream) that demultiplexing module (not shown) is sent here, and sends into decoding unit as audio decoder and outside interface.Decoding unit comprises parsing module 431 and reconstructed module (reconstruction) 433.Parsing module is resolved ES stream, and decode headers information is determined header parameter, for other decoder module provides control word information.The frame structure of reconstructed module analysis of MPEG AudioLayerl/Layer2 is decoded to the bit distribution information of frame data; Carry out then the decoding of scale factor (scale factor), again get sampled point, sampling point is carried out the inverse quantization computing; Counter at last (de-normalize) operation of standardizing.The signal of reconstructed module output is input to the synthon band filter.Output buffer is at the uniform velocity sent PCM sampled value data and is given post-processing module or directly give playback equipment through after the analog-to-digital conversion, and playback equipment is play with the stabilized speed of 48K/S.For keeping the stable output at the uniform velocity of playback equipment, need make the data-signal in the output buffer keep certain data volume size.
Referring to Fig. 5, be provided with overflow pointer 51 and underflow pointer 52 in the output buffer 47, disposed almost_full (overflow) and almost_empty (underflow) signal-arm accordingly.The problem that send out signal wrong, audio-visual asynchronous, decoding error etc. may occur is controlled by CPU, when CPU sends the operation that control signal has been ended error unit, may cause that the response speed of CPU goes wrong.Also there are other situations may cause that the response speed of CPU goes wrong.When the response speed of resources such as bus, processor went wrong, underflow condition may appear in the data in the output buffer, and at this moment, output buffer can feed back to decoding unit 43 to this almost_empty signal-arm.After decoding unit receives the almost_empty signal-arm, send interpolation to give filter 45.Input buffer and output buffer can be first-in first-out buffer (FIFO).
Shown in Fig. 4 B, be provided with interpolation counter 437 and interpolating unit 435 in the decoding unit 43.Output buffer is checked through underflow condition, then feeds back a underflow signal almost_empty and gives decoding unit.Decoding unit is imported interpolation by interpolating unit to filter under the control of interpolation counter, filter carries out the conversion of frequency domain to time domain to the S territory sampled value that interpolation and decoding unit are sent into, thereby finishes smothing filtering.
Referring to Fig. 6, according to one embodiment of the present of invention, interpolating unit 435 for example comprises multiplexer 61, and the latter receives two input values at least, and an input value is a S territory sampled value, and another input value is an interpolation.Multiplexer carries out gating to two input values under the control of interpolation counter.The interpolation counter generally is in reset mode, and gating S territory sampled value is as input.When receiving the underflow signal almost_empty that output buffer sends here, interpolation counter gating interpolation is as input, and begins counting simultaneously; When count value equaled the filter width value, the interpolation counter resetted again, thus multiplexer again gating S territory sampled value as input.
The interpolation that interpolating unit is given filter for example is general constant value or null value.In a kind of execution mode preferably, receive when decoding unit under the situation of underflow signal almost_empty that the interpolation that interpolating unit is given filter is a null value.According to the characteristics of sub-filter, once insert the null value of counting equal number with filter channel, be 32 null values at this.Correspondingly, the interpolation counter constitutes (not shown) by one 5 register and one 5 adder, can carry out 32 countings.
In another embodiment of the present invention, subband synthesis filter is 1024 grades for example, comprises 32 passages, and as a set of samples, each subband comprises 32 points with 32 subband sampling points.Receiving at decoding unit under the situation of underflow signal almost_empty, equally also is to insert the null value of counting same widths with filter channel, i.e. 32 null values.The null value number of inserting and the counting figure place of interpolation counter are consistent with the filter channel number, under the situation that changes filter order and width, all can change.
Fig. 7 is receiving the oscillogram of inserting interpolation under the underflow signal almost_empty situation for audio decoder of the present invention.When the response speed of resource such as bus, processor goes wrong, may cause that underflow appears in output buffer.Decoding unit send interpolation to give filter according to the almost_empty underflow signal of output buffer.Part 71 among waveform that interpolation produces such as the figure between two dotted lines.Waveform behind the filter filtering after the interpolation is able to continuously signal, and has eliminated the waveform transition that the high frequency boundary produces.
The filter here is the subband synthesis filter of carrying out idct transform, is the element that necessarily requires in the audio decoder.When the response speed of resources such as bus, processor goes wrong, before idct transform, insert null value, filter carries out idct transform and windowing process together to S territory sampled value and null value that decoding unit is sent into, makes it to be transformed into time domain from frequency domain.Like this, the null value of insertion can not produce the transition of waveform, does not just need extra low pass filter to carry out smothing filtering yet.
In other embodiments of the invention, when the response speed of resources such as bus, processor went wrong, decoding unit for example obtained interpolation according to the signal estimation of front, and gave filter the interpolation that prediction obtains.
In other embodiments of the invention, if previous frame is correctly decoded, decoding unit can repeat output and give filter with the decoded result of previous frame as interpolation.
Fig. 8 is the flow chart of a kind of audio-frequency decoding method according to an embodiment of the invention.Referring to Fig. 8 and in conjunction with the description of front, step S81 decodes to the outside elementary stream ES that receives, and obtains S territory sampled value.Step S82 carries out inverse discrete cosine transform and windowing process to described S territory sampled value or interpolation, output PCM sampled value (do not providing under the situation of interpolation, this step is only carried out inverse discrete cosine transform and windowing process to S territory sampled value).Step S83 detects the data volume of also judging the PCM sampled value of output and whether reaches underflow condition, and as not, program is returned step S82 and continued S territory sampled value is carried out inverse discrete cosine transform and windowing process, output PCM sampled value; In this way, then provide interpolation, program is returned step S83, and carry out inverse discrete cosine transform and windowing process and export PCM sampled value described interpolation this moment.
Fig. 9 is the flow chart of the smooth filtering method in a kind of according to an embodiment of the invention audio decoder.In conjunction with Fig. 4 and Fig. 9, described audio decoder comprises the outside elementary stream ES that receives decoded and obtains the decoding unit 43 of S territory sampled value, described S territory sampled value is carried out inverse discrete cosine transform and windowing process, obtain the filter 45 of PCM sampled value, and the output buffer 47 in order to export described PCM sampled value.Described smooth filtering method comprises: step S91, when the response speed of resources such as bus, processor goes wrong, when the data volume that detects output buffer reduces to underflow condition, send underflow signal almost_empty and give decoding unit.Step S93, decoding unit send interpolation to give filter under the triggering of underflow signal almost_empty.Step S95, filter carries out inverse discrete cosine transform and windowing process to S territory sampled value and interpolation that decoding unit is sent into.
Present embodiment is just in order further more clearly to describe the present invention, but not limitation of the present invention.Be to be understood that the present invention is not limited to the elaboration that embodiment does, anyly all should be encompassed within the spirit and scope of claim of the present invention based on modification of the present invention and equivalent of the present invention.
Claims (10)
1. audio decoder comprises:
In order to receive the input buffer of outside elementary stream ES;
Described elementary stream ES is decoded, obtain the decoding unit of S territory sampled value;
Described S territory sampled value is carried out inverse discrete cosine transform and windowing process, obtain the filter of PCM sampled value; And
In order to exporting the output buffer of described PCM sampled value,
Wherein, described decoding unit also receives the underflow signal of output buffer feedback to described filter input interpolation, and described filter carries out inverse discrete cosine transform and windowing process to finish the conversion of frequency domain to time domain to described interpolation and described S territory sampled value.
2. audio decoder as claimed in claim 1 is characterized in that, described decoding unit comprises the parsing module of resolving in order to described elementary stream ES, and decodes in order to the data that parsing is obtained and to obtain the reconstructed module of S territory sampled value.
3. audio decoder as claimed in claim 2, it is characterized in that, described decoding unit comprises in order to the interpolating unit to described filter input interpolation, and controls the interpolation counter that described interpolating unit is carried out interpolation in order to the underflow signal that feeds back according to output buffer.
4. audio decoder as claimed in claim 3 is characterized in that, described interpolating unit comprises the multiplexer that the S territory sampled value and the interpolation of input is carried out gating according to the control of interpolation counter.
5. as the arbitrary described audio decoder of claim 1 to 4, it is characterized in that described filter is a subband synthesis filter.
6. audio-frequency decoding method comprises:
The outside elementary stream ES that receives is decoded, obtain S territory sampled value;
Described S territory sampled value is carried out inverse discrete cosine transform and windowing process, output PCM sampled value;
When the data volume of the PCM sampled value that detects output reaches underflow condition, provide interpolation;
Described interpolation is carried out inverse discrete cosine transform and windowing process to finish the conversion of frequency domain to time domain with described S territory sampled value.
7. method as claimed in claim 6 is characterized in that, the described step of interpolation that provides is by an interpolation counter controls.
8. method as claimed in claim 7, it is characterized in that, the step that described interpolation is carried out inverse discrete cosine transform and windowing process with described S territory sampled value comprises: according to the control of described interpolation counter, and the step that the described S of gating territory sampled value or the described interpolation of gating are carried out described processing.
9. the smooth filtering method in the audio decoder, described audio decoder comprises the outside elementary stream ES that receives decoded and obtains the decoding unit of S territory sampled value, described S territory sampled value is carried out inverse discrete cosine transform and windowing process, obtain the filter of PCM sampled value, and in order to export the output buffer of described PCM sampled value, described method comprises:
When the data volume of output buffer reduces to underflow condition, send underflow signal and give decoding unit;
Decoding unit send interpolation to give filter under the triggering of underflow signal;
S territory sampled value that filter is sent into decoding unit and interpolation are carried out inverse discrete cosine transform and windowing process to finish the conversion of frequency domain to time domain.
10. as claim 6 or 9 described methods, it is characterized in that described interpolation is a null value, or value, or repeat decoded result with previous frame under the situation about being correctly decoded for previous frame as interpolation for obtaining according to the signal estimation of front in the decode procedure.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005101110847A CN100527849C (en) | 2005-12-02 | 2005-12-02 | Audio-frequency decoding method, audio-frequency decoder and mooth filtering-wave method therein |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005101110847A CN100527849C (en) | 2005-12-02 | 2005-12-02 | Audio-frequency decoding method, audio-frequency decoder and mooth filtering-wave method therein |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1980397A CN1980397A (en) | 2007-06-13 |
| CN100527849C true CN100527849C (en) | 2009-08-12 |
Family
ID=38131362
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2005101110847A Expired - Fee Related CN100527849C (en) | 2005-12-02 | 2005-12-02 | Audio-frequency decoding method, audio-frequency decoder and mooth filtering-wave method therein |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100527849C (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103065641B (en) * | 2013-02-01 | 2014-12-10 | 飞天诚信科技股份有限公司 | Method for analyzing audio data |
-
2005
- 2005-12-02 CN CNB2005101110847A patent/CN100527849C/en not_active Expired - Fee Related
Non-Patent Citations (1)
| Title |
|---|
| MPEG音频解码中子带合成滤波器的快速算法及定点DSP实现. 王剑虹，吴海华，陈健.上海交通大学学报，第34卷第6期. 2000 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1980397A (en) | 2007-06-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US5619197A (en) | Signal encoding and decoding system allowing adding of signals in a form of frequency sample sequence upon decoding | |
| TWI463790B (en) | Adaptive hybrid transform for signal analysis and synthesis | |
| US20090055005A1 (en) | Audio Processor | |
| US8977557B2 (en) | Method, medium, and apparatus encoding and/or decoding multichannel audio signals | |
| JP2008539462A (en) | Method and system for operating audio encoders in parallel | |
| KR100804640B1 (en) | Subband synthesis filtering method and apparatus | |
| EP1719118A1 (en) | Decoding scheme for variable block length signals | |
| CN100561582C (en) | Audio decoder and audio-frequency decoding method | |
| US20230108008A1 (en) | Method and apparatus for decoding a bitstream including encoded higher order ambisonics representations | |
| US6487528B1 (en) | Method and apparatus for encoding or decoding audio or video frame data | |
| US6173024B1 (en) | Bit stream reproducing apparatus | |
| CN100527849C (en) | Audio-frequency decoding method, audio-frequency decoder and mooth filtering-wave method therein | |
| US6542550B1 (en) | Method for processing and apparatus for encoding or decoding audio or video frame data | |
| US20120102538A1 (en) | Methods and apparatus for decoding multiple independent audio streams using a single audio decoder | |
| KR100346734B1 (en) | Audio coder and decoder having high speed analyzing filter and composite filter | |
| US6161088A (en) | Method and system for encoding a digital audio signal | |
| EP1020998B1 (en) | Method and apparatus for encoding audio frame data | |
| TWI328358B (en) | An audio decoder and an audio decoding method | |
| JP2004520757A (en) | Inverse filtering method, synthesis filtering method, inverse filter device, synthesis filter device, and device having such a filter device | |
| Grueger et al. | MPEG-1 low-cost encoder solution | |
| KR100322703B1 (en) | Audio encoder/decoder | |
| EP1020997A2 (en) | Method for processing and apparatus for encoding audio or video frame data | |
| KR100370412B1 (en) | Audio decoding method for controlling complexity and audio decoder using the same | |
| WO2006106356A1 (en) | Encoding and decoding a signal | |
| KR960012476B1 (en) | Frame bit apparatus |
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 | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20090812 Termination date: 20111202 |