JPH09115251A - Method and device for processing audio data - Google Patents

Abstract

PROBLEM TO BE SOLVED: To add precise encoding time information to audio encoding data with a simple constitution. SOLUTION: The encoding time information D5 related to respective unit encoding units are formed whenever the encoding related to respective unit encoding units by an encoding part 5 is started, and the relevant encoding time information D5 is registered successively in order of formation. Together with that, when the encoding data D2' of respective corresponding unit encoding unit are formed, the registered encoding time information D5 is read out by an FIFO system to be added to the relevant answering encoding data D2'.

Description

Detailed Description of the Invention

[0001]

[Table of Contents] The present invention will be described in the following order. TECHNICAL FIELD OF THE INVENTION Conventional Technology (FIGS. 4 and 5) Problems to be Solved by the Invention Means for Solving the Problems Embodiments of the Invention (FIGS. 1 to 3)

[0002]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an audio data processing method and an audio data processing device, for example, M
It is suitable for application when audio-encoded data encoded according to the PEG (Moving Picture Experts Group) audio standard is packetized.

[0003]

2. Description of the Related Art Conventionally, there is an MPEG audio standard as a standard for highly efficient coding of audio signals and for providing commonality among devices. This MPEG audio standard is
ISO / IEC standard method for stereo audio coding with high sound quality and high efficiency, which is an international standard.
IS 11172-3. The MPEG audio standard has three types of modes, layer 1, layer 2 and layer 3, depending on the encoding algorithm, but the data structure and bit stream configuration are common.

As shown in FIG. 4, one frame of the MPEG audio-bit stream is AAU (Audio Access U).
nit: audio decoding unit), which is the minimum unit that can be independently decoded into the original audio signal. The size of the audio access unit depends on the sampling frequency and transmission rate, but it is structurally designed to always contain 384 samples of data.

By the way, MPEG audio data and M
In order to apply an encoded bit stream such as PEG video data to an actual AV (audio / video) device, it is necessary to synchronize and multiplex these bit streams. Generally, a system that plays this role is called an MPEG system.

The MPEG system has two types of standards, that is, an MPEG1 system and an MPEG2 system. M
The PEG1 system is ISO / IEC IS 11172-1, mainly CD
-A standard for storage media such as ROM (Compact Disc-Read Only Memory). In addition to the storage media, the MPEG2 system is intended for a wider range of devices such as broadcasting and communication.

The basic data structure of the MPEG system is common to both MPEG1 and MPEG2.
A packet multiplexing method as shown in FIG. When multiplexing, divide each bit stream of video data and audio data into packets of appropriate length, add packet headers with attributes and time stamp information of each packet, and then time-share by packet switching. To transmit. The packet length may be variable or fixed within a certain length range. MPEG
In 2, the packet is called PES (Program Elementary St
ream) called a packet.

Here, the time stamp is information indicating when the MPEG audio data and the MPEG video data are decoded and reproduced, and is added to each reproduction unit of one frame for video data and one audio frame for audio data. Has been done. With this time stamp, the audio and video can be synchronized on the decoding and reproducing side. The time stamp indicating the decoding time is called DTS (Decoding Time Stamp), and the time stamp indicating the reproduction time is PTS (Presentation Time Stamp).
mp) and the actual time is SCR (System Cloc
It is obtained by referring to a time reference reference value called k Reference).

The SCR is provided in a header called a pack header, which is added for each unit (pack) composed of several packets, and sets and calibrates the time reference (STC: System Clock) of the system decoder. The decoder performs a decoding process when STC matches DTS, and reproduces and outputs the access unit when STC matches PTS.

The PTS and DTS are added to the packet header when the head of the access unit is present in a certain packet, and are not added when they are not present. Further, even if there are two or more access units at the head in the packet, only the time stamp corresponding to the first access unit is added.
Also, like MPEG video data, both PTS and DTS are added when there is a frame exchange at the time of encoding and the decoding order and the playback order are different, but when there is no frame exchange such as MPEG audio data, only PTS is added. It is designed to do.

[0011]

By the way, considering audio data, in order to realize a reproduction state at the same decoding time as encoding, the PTS added at the time of multiplexing is set to the target audio access unit (AA).
It must be the SCR value at the moment when U) encoding is started. However, when the audio access unit output from each encoder arrives at the packet processing unit, various time delays such as encoding processing time, system matching delay, buffer delay and transmission delay are seen from the moment of encoding. Because it has passed,
The SCR value at that time is not an accurate PTS.

As a result, the error in the time information makes it impossible to accurately synchronize the encoding and the decoding, which makes it difficult to perform good decoding. Therefore, it is necessary to add an accurate PTS to the encoded data. However, since the delay time described above differs from system to system, the correct PT
The circuit for generating S must be designed for each system, and the circuit configuration is very complicated.

In addition to this, in the conventional audio encoder and decoder, there is a method in which the same reproduction state as that at the time of encoding is obtained at the time of decoding by implicit synchronization without adding PTS. According to the implicit synchronization method, it is not necessary to generate and transmit time information such as PTS, so that the circuit configuration can be simplified. However,
In the implicit synchronization method, compared with the case where a time stamp is added and transmitted, the accuracy is considerably reduced in that the reproduction state at the time of decoding is the same as that at the time of encoding.

The present invention has been made in consideration of the above points, and has a simple structure and is capable of adding accurate coding time information to audio coded data, and an audio data processing method and an audio data processing apparatus. Is to propose.

[0015]

In order to solve such a problem, according to the present invention, the coding time information about each unit coding unit is started every time the coding unit starts coding for each unit coding unit. And sequentially register the coding time information in the order in which they are formed, and when the corresponding encoded data of each unit coding unit is formed, the registered coding time information is read by the first-in first-out method. It should be added to the corresponding encoded data.

As a result, since the coding time information is formed at the time when the coding for each unit coding unit is started, it may be a value that does not include the delay time such as the coding delay and the transmission delay. it can. Also, the coding time information formed earlier in time than the coded data of the corresponding unit coding unit is the unit code that is read by the first-in first-out method when the corresponding coded data is formed. It is added in synchronization with the encoded data of the encoding unit.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below in detail with reference to the drawings.

In FIG. 1, reference numeral 1 denotes an audio data processing apparatus as a whole, which is roughly divided into an encoder 2 and a packet processing unit 3. The audio data processing device 1 converts the input audio signal S1 into audio data D1 via an analog digital (A / D) converter 4, and supplies the audio data D1 to an audio encoder 5. The audio encoding unit 5 forms audio encoded data D2 by performing fixed-length encoding on the audio data D1 using an encoding method such as subband encoding.

The audio coded data D2 passes through the system delay section 6 and is delayed by a certain time according to the system and then input to the multiplexing section 7 of the packet processing section 3. Here, the system delay unit 6 does not actually have such a part, and is, for example, a time required for the encoding process of the audio encoding unit 5, a buffer delay, or a delay necessary for system matching. These days are represented by blocks.

Here, the audio encoding unit 5 includes an audio access unit (AA) composed of 384 samples.
U) is encoded as an encoding unit, and a data stream in audio access unit units is output as audio encoded data D2. At this time, the audio encoder 5 generates a data pulse signal D3 at the head position of each audio access unit and outputs it. The data pulse signal D3 is delayed by the same time as the audio encoded data D2 via the system delay unit 6 and then input to the packet processing unit 3.

The frame pulse generator 8 receives the audio data D1 from the analog digital converter 4 and inputs 1
A frame pulse signal D4 is generated every time the first data constituting one audio access unit is input, and this is generated by the SCR (System Clock Referenc) of the packet processing unit 3.
e) Send to the counter 9. The SCR counter 9 has a time base reference value that counts up over time,
When the frame pulse signal D4 is input, a PTS (Presentation Time Stamp) signal D5 is formed based on the time reference reference value at that time, and this is sent to the memory 10.

The memory 10 is a first-in first-out memory, that is, a FIFO memory, and is supplied with the PTS signal D5.
Are sequentially stored in that order. And the data pulse signal D
3'is used as a read control signal for the data pulse signal D
When 3'is input, PTS signals D5 are output one by one in the order in which they are stored.

The PES (Program Elementary Stream) header generator 11 outputs the PTS signal D output from the memory 10.
5'is used to form a PES header as shown in FIG. 5 (in the embodiment, since it is audio, DTS in FIG. 5 may be omitted), and data pulse signal D3 '
The PES header data D6 according to the arrival time of
To the multiplexing unit 7. The multiplexing unit 7 adds at the head of each audio access unit of the audio coded data D2 ',
An audio PES stream D7 is formed by adding the corresponding PES header data D6. The audio PES stream D7 is supplied to a multiplexing unit (not shown) in the subsequent stage and time-division multiplexed with the video packet data.

Here, the packet processing unit 3 executes the packetizing process according to the procedure shown in FIG. The packet processing unit 3 enters the packet processing procedure at step SP0,
When the frame pulse signal D4 is received in the following step SP1, the SCR counter 9 is received in step SP2.
The PTS signal D5 is formed based on the time reference reference value of, the PTS signal D5 is written in the memory 10 in step SP3, and the process returns to step SP1. The packet processing unit 3 repeats the loop of steps SP1-SP2-SP3-SP1 to sequentially store the PTS signal D5 in the memory 10.
I will write.

When the packet processing unit 3 receives the data pulse signal D3 'in step SP4, it reads the PTS signal D5 stored in the memory 10 in the following step SP5, and the PES header generating unit 11 reads it in the following step SP6. A PES header is generated using the PTS signal D5 ', and the multiplexing unit 7 multiplexes the PES header at a predetermined position of the audio access unit data stream (in the embodiment, the head position of each audio access unit) in step SP7. As a result, packet data is formed, and the process returns to step SP4. The packet processing unit 3 uses this step SP4-SP5-SP6.
By repeating the loop of -SP7-SP4, the PTS signal D5 'stored in the memory 10 is sequentially added to the corresponding audio access unit to form the packet data.

In the above structure, each audio access unit data D2 'input to the packet processing section 3 is inputted.
Is delayed by 6 system delays. Similarly, the data pulse signal D3 is also delayed by the amount of the system delay 6, and each audio access unit data D
It is input to the packet processing unit 3 in synchronization with 2 '. On the other hand, the frame pulse signal D4 is input to the SCR counter 9 of the packet processing unit 3 at almost the same time as the time when the corresponding audio access unit data D2 is coded by the audio coding unit 3.

As a result, the SCR counter 9 provides the PTS signal D5 which accurately represents the coding time. Then, the PTS signal D5 is stored in the FIFO memory 10. The PTS signal D5 'is read from the FIFO memory 10 at the timing when the data pulse signal D3' arrives. As a result, the corresponding audio access unit is synchronized with the PTS signal D5 ', and the PTS corresponding to each audio access unit is multiplexed in the multiplexing unit 7.
The signals will be multiplexed. Therefore, the packet unit is 1AAU = 1PES, and the PTS header is PTS.
And AES, which is PES data, always has a one-to-one correspondence.

According to the above construction, the frame pulse signal D4 generated at the start of encoding of each audio access unit is directly input to the SCR counter 9 separately from the audio encoded data D2 and D2 ', and each audio access is performed. By forming the PTS signal D5 for the unit, it is possible to form an accurate PTS signal D5 that is not adversely affected by the delay time required for audio encoding or transmission.

The PTS signal D5 thus formed is temporarily stored in the FIFO memory 10 and is synchronized with the data pulse signal D3 'having a delay time equal to that of the audio coded data D2' to store the PTS signal D5 '. Since the data is read, the audio coded data D
2'and the PTS signal D6 can be synchronized,
The signals can be multiplexed into the corresponding audio access unit.

Thus, it is possible to realize the audio data processing device 1 which can add an accurate PTS signal to the corresponding audio access unit with a simple structure.

In the above embodiment, the audio encoder 5 forms the data pulse signal D3 representing the beginning of each audio access unit, and the data pulse signal D3 is passed through the same path as each audio access unit. As a result, each audio access unit has a delay time equal to that of the corresponding PT.
The case where the S signal D5 'and the PES header signal D6 are used as the synchronization control signal for reading has been described, but the present invention is not limited to this, and the packet processing unit 2 may be configured, for example, as shown in FIG. You may make it perform a synchronous control.

That is, an audio access unit header detection unit 21 is provided in the packet processing unit 20, and a header (see FIG. 4) added to the head of each audio access unit is detected by the audio access unit header detection unit 21. The data pulse signal D10 is generated each time it is detected. And this data pulse signal D
10 is sent to the memory 10 and the PES header generator 11, the PTS signal D
5'and PES header signal D6 can be synchronized with the corresponding audio access unit. Further, in this way, the encoder unit 2 is changed to the packet processing unit 2
It is not necessary to transmit the data pulse to 0, and the interface between them and the processing of the encoder unit 2 are simplified.

In the above embodiment, the case where the SCR counter 9 and the FIFO memory 10 are provided in the packet processing section 3 has been described, but the present invention is not limited to this, and they are provided in the encoding section 2. Even if the PTS signal D5 'output from the FIFO memory 10 is sent to the packet processing unit 3, the same effect as that of the above embodiment can be obtained.

In the above embodiment, the present invention is
Encoding time information (PTS signal D5) is added to audio encoded data D2 'encoded according to the MPEG audio standard.
However, the present invention is not limited to the MPEG audio standard and can be widely applied to the case where the coding time information is added to the audio coded data.

In the above embodiment, the case where PTS information is used as the time information to be added to the audio coded data D2 'has been described, but the coded time information to be added is not limited to the PTS information, and the point is that Any time information can be used as long as it can represent the coding time of the unit coding unit.

Further, in the above-mentioned embodiment, the case where the FIFO memory 10 is used as the means for carrying out the first-in first-out of the coding time information (PTS signal D5) has been described.
The present invention is not limited to this.
IFO registration may be performed.

Further, in the above-mentioned embodiment, the coding time information (PTS) is added to the head of each audio access unit.
The case where the header information D6 including the signal D5 ') is added has been described, but the position where the header information D6 is added is not limited to this.

[0038]

As described above, according to the present invention, the coding time information representing the time when the coding for each unit coding unit is started is formed, and this is sequentially registered in the order in which it is formed. When the corresponding coded data is formed, the registered time information is read by the first-in first-out method and added to the corresponding coded data, so that the audio coded data can be coded with high accuracy with a simple configuration. Activation time information can be added.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an audio data processing device according to an embodiment.

FIG. 2 is a flow chart for explaining the operation of the packet processing unit.

FIG. 3 is a block diagram showing a modified example of the packet processing unit.

FIG. 4 is a schematic diagram showing a frame structure of an MPEG audio bitstream.

FIG. 5 is a schematic diagram showing a packet data structure of an MPEG system stream.

[Explanation of symbols]

1 ... Audio data processing device, 2 ... Encoder,
3, 20 ... Packet processing unit, 5 ... Audio encoding unit, 7 ... Multiplexing unit, 8 ... Frame pulse generating unit, 9 ...
... SCR counter, 10 ... memory, 11 ... PES header generation section, 21 ... audio access unit header detection section, S1 ... input audio signal, D1 ... audio data, D2, D2 '... audio encoding Data, D3, D3 '... Data pulse, D4 ... Frame pulse, D5, D5' ... PTS signal, D6 ... PES
Hezda signal, D7 ... Audio PES stream.

Claims (3)

[Claims] Claims: 1. An input audio signal is encoded for each predetermined data unit, and then header information, which is information about the unit encoding unit, is added to each unit encoding unit of the formed encoded data. In the audio data processing method for forming audio packet data according to the above, the input audio signal is encoded in a predetermined data unit to form the encoded data divided for each unit encoding unit. Encoding time information representing the time when the encoding of the encoding unit is started is formed, and the encoding time information is sequentially registered in the order in which it is formed, and the corresponding encoded data of the unit encoding unit is formed. At the same time, the registered coding time information is read by the first-in first-out method, and the corresponding coding data is read. Audio data processing method characterized by adding to. 2. An encoding section for encoding an input audio signal for each predetermined data unit, and information about the unit encoding unit for each unit encoding unit of encoded data formed by the encoding section. In an audio data processing device having a packet processing unit for forming audio packet data by adding header information, a unit encoding unit provided in the encoding unit and encoding audio data in a predetermined data unit. Encoding means for forming the encoded data, which is divided for each unit, and a pulse signal, which is provided in the encoding section, each time the encoding means starts encoding for each unit encoding unit. Pulse signal generating means, and each unit for each input of the pulse signal The time information forming means for forming the encoding time information of the encoding unit and the respective encoding time information are sequentially registered in the order of input, and the encoding time information is input each time the head of the unit encoding unit arrives. First-in first-out registration means for outputting in the order of, the header processing unit provided in the packet processing unit, and forming the header information containing at least the encoding time information output from the first-in first-out registration unit, and the packet processing unit. , An audio data processing device comprising: multiplexing means for multiplexing the header information output from the header information forming means into the corresponding unit encoding units. 3. The encoding section further comprises a second pulse signal generating means for generating a second pulse signal synchronized with the head of each unit encoding unit, the first-in first-out registration means and the header information formation. 3. The audio data processing device according to claim 2, wherein the means outputs the encoding time information and the header information each time the second pulse signal arrives.