KR100346733B1 - Audio coding/decoding method and apparatus capable of controlling scale of bit stream - Google Patents

Abstract

PURPOSE: An audio coding/decoding method and an apparatus capable of controlling scale of a bit stream are provided to improve the service quality by forming the bit stream in compatible with various transmitting paths and restoring devices. CONSTITUTION: A coding apparatus is comprised of a coder(21) for forming a bit stream and a bit stream forming unit(22). The bit stream forming unit(22) capable of controlling a scale thereof changes the formed bit stream to a bit stream capable of controlling the scale. A decoding apparatus is comprised of a bit stream decoding unit(23) capable of controlling the scale for restoring the bit stream and capable of controlling the coded scale and a decoder(24) for getting original data from the restored bit stream.

Description

Audio encoding / decoding method and apparatus capable of scaling a bitstream

The present invention relates to an audio encoding / decoding method and apparatus, and more particularly, to an audio encoding / decoding method and apparatus capable of scaling a bitstream for representing data processed by a general audio codec as a scalable bitstream. .

Signals in nature exist as analog signals, usually in the form of continuous signals. When the analog signals are used in the reproducing apparatus or the transmitting apparatus, deterioration occurs easily, compensation for possible error signals is difficult, and the size of the system increases when the system is configured. These problems can be improved by transforming analog signals into digital signals through sampling and quantization processing. At this time, since the information obtained changes with the passage of time, the data must be obtained at every hour interval. In consideration of this, when implementing any system, the sampling interval is determined only for real-time processing of the reference data. Data in the time domain are obtained by the determined sampling interval. The data information accepted by human beings is values in the time domain, but the information actually felt by human beings is the power of the signal. When the signal is processed in the digital method, the sound quality is improved compared to the analog method, but the data amount is increased in comparison with the analog method.

Compression is performed to reduce the influence of the increase in data volume, but if compression is not performed, transmission cannot be performed because the channel does not provide the bandwidth required for the transmission of video information or audio information. . Many studies on data compression have been conducted. For example, in a gray scale photographic image having a resolution of 1000 × 1000 lines, it is assumed that each pixel is composed of 8 bits and is represented with different density of 256 gray levels for each pixel. In this case, without compression, the number of bits used to transmit this photographic image is 8 million bits, and in the case of a general telephone line, 9600 bits can be transmitted per second. Takes As another example, for audio sampled at 44.1 kHz, assuming that the sampled information is sampled with 16 bits and has 65536 gradations, the number of bits for one channel is about 700 kbits, which takes about 73 seconds when using a regular telephone line. Can be said to be impossible.

For this reason, studies on data compression have been actively conducted, and various data compression schemes such as waveform coding, transform coding, subband coding, fractal coding, predictive coding, and motion compensation coding have been proposed. Recently, these data compression methods have been used in the commercialization of digital compact cassettes (DCC) using MPEG Audio 1 (ISO / IEC 11172) by Philips, Sony's mini-disc (MD), ISO / IEC 13818 for HDTV, etc. Research on MPEG-IV and the like for use in video telephones or multimedia devices is in progress.

In the case of transmitting the compressed data as described above, the bit rate used for transmission and the bit rate for real-time processing are determined according to the characteristics of the transmission channel, and the data is fixed at a state that can guarantee the maximum sound quality or image quality at the determined bit rate. Sent. The determination of the transport channel has led to the development of devices for optimally reconstructing the data to be transmitted and devices for encoding the most efficient use of the data rate.

However, in this case, since the transmission rate is fixed, it is possible to design a specific reproducing apparatus that is effective at the transmission rate.However, the method of forming the data to match the transmission rate used in various systems and the method of continuously sending various kinds of bitstreams Not presented. That is, when arbitrary information is transmitted, the user cannot arbitrarily select and restore the bit rate of the transmitted data. In the restoring system, for example, when video and audio information are stored on an optical disc, the data cannot be reproduced by adjusting the bit rate. That is, the process of adjusting the quality while adjusting the bit rate assigned to the video information and the bit rate assigned to the audio information was not possible.

A bitstream applicable to these various bit rates is called a bitstream capable of scaling. For such a scalable bitstream, K. Brandenburg has shown an example of processing to enable various bit rates in "First Ideas on Scalable Audio Coding" (AES 97th Conv. Preprint 3924, San Francisco, 1994). In this example shown in FIG. 1, a bitstream for a core portion is first obtained, then an error signal at that time is calculated, and a bitstream for compensating for the errors generated for the error signals is formed. As a result, in order to express bitstreams for various bit rates, the complexity of hardware and the complexity of operations increase according to the kind of bit rate of the corresponding bit stream, and thus, the processing time increases. For example, if the bitstream can be scaled in four steps, it has four times the complexity.

Accordingly, an object of the present invention is to select the stored data or the data being transmitted from a predetermined type of bit rate when changing the sound quality or image quality of the transmitted data by adjusting the bit rate by the user in order to solve the above problems. The present invention provides an audio encoding / decoding method capable of scaling a bitstream to form a bitstream of a bit rate.

Another object of the present invention is to provide an apparatus most suitable for realizing an audio encoding / decoding method capable of scaling the bitstream.

In order to achieve the above object, an audio encoding / decoding method capable of scaling a bitstream according to the present invention is provided.

An encoding process for compressing and encoding the input audio signal to form a bitstream;

A scaleable bitstream forming step of changing the bitstream formed by the encoding process into a bitstream capable of scaling;

A bitstream reconstruction process capable of scaling to reconstruct the encoded bitstream into a form suitable for decoding; And

And a decoding process for recovering the original bitstream to obtain the original data.

In order to achieve the above another object, an audio encoding / decoding apparatus capable of scaling a bitstream according to the present invention is provided.

Audio encoder

An encoder for compressing and encoding the input audio signal to form a bitstream; And

A scaleable bitstream forming unit for converting the bitstream formed by the encoder into a scaleable bitstream,

Audio decoder

A bitstream reconstruction unit capable of scaling to reconstruct the encoded bitstream into a form suitable for decoding; And

And a decoder for reconstructing the recovered bitstream to obtain original data.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

2 is a block diagram illustrating an audio encoding / decoding apparatus capable of scaling a bitstream according to an embodiment of the present invention, in which an encoding part includes an encoder 21 for forming a bitstream and a bitstream for scaling a bitstream. Composed of a bitstream forming unit 22 capable of scaling, the bitstream forming unit 22 capable of adjusting the scale inputs a control signal for determining the length of the bitstream according to the bit rate for bit rate adjustment (not shown). At the same time, a portion (not shown) for adjusting the length of the bitstream according to the control signal and a portion (not shown) for adjusting the length of the entire frame are provided. On the other hand, the decoding part is composed of a scalable bitstream recovery unit 23 for recovering the coded bit-adjustable bitstream and a decoder 24 for obtaining the original data from the recovered bitstream.

The operation of each part based on the configuration of FIG. 2 is as follows.

In FIG. 2, the encoder 21 and the decoder 24 are for compressing or reconstructing data, and make the best use of the channel transmission capacity or storage capacity. This part can use a variety of existing compression schemes, but in the present invention, for convenience of description, ISO / IEC 11172 audio will be taken as an example.

The bitstream encoded by the ISO / IEC 11172 audio encoder 21 is as shown in the third soil. 4 shows an example of decoding such a bitstream, and it can be seen that many parts of the bitstream are always used in the representation of information with the same size by a predetermined form. The important values here are allocation [ch] [sb], which stores the values assigned to the bits by their varying importance according to the signal, and assigns bits to the samples by this value.

If there are various bit rates constituting the scalable bitstream determined using this information, that is, the first bit rate, the second bit rate, ..., the n-th bit rate, the bit stream is scalable with the encoded bit stream In the forming unit 22, a bitstream capable of adjusting a scale including all of the nth bit rate from the first bit rate may be formed. The scaleable bitstream forming method is performed by the following series of steps.

1. Allocate bits for sideinfo.

2. From the side information, the number of bits BB [ch] [sb] = allocation [ch] [sb] allocated for each processing band is read.

3. Read the number of samples sample [ch] [sb] [s] for each processing band according to each BB [ch] [sb]. This step is performed by the following steps a to i.

a. Initialize to CC [.] = 0 and M = 1.

b. Find the maximum BB [ch] [sb] for all BB [ch] [sb] in the high frequency band from the low frequency band.

c. The number of bits allocated to the samples of the band having the maximum BB [.] Value is adjusted as follows.

CC [ch] [sb] = CC [ch] [sb] + 1

d. Adjust the BB [ch] [sb] value as follows:

BB [ch] [sb] = BB [ch] [sb] = 1

e. The number of bits _used (Bits _used ) is estimated as follows.

Here, SPL [.] Means a fixed number of samples for each processing band.

f. If the Bits _used value does not exceed the value of the bit rate M, the process returns to step b.

g. Read the number of samples using CC [ch] [sb]. At this time, samples of each processing band are read by the most significant bit (MSB) by CC [ch] [sb]. Thereafter, each sample [ch] [sb] [s] value is moved left by CC [ch] [sb].

h. Set CC [.] = 0 and return to step b if M is not the last bit rate.

i. End

In this way, when bit allocation is performed, as shown in FIG. 5A, bits are allocated to each band as in FIG. 5A. The bitstream obtained by the processing result is as shown in FIG.

Here, since the number of bits forming the bitstream is determined, and the bit information according to each bit rate is from the meaningful maximum value (MSB) of the original data to the kth bit, the original coded bitstream when the final bit rate is N. Is changed to a scalable bitstream while the total number of bits does not change. However, until the number of bits for the bit rate k is exceeded, the bits are allocated according to the number of samples of each band in the band where the bits are allocated a lot and the low frequency band. The case occurs. For example,

k step; Find the band and allocate bits. Bits _used is m.

K + 1 step: find the band and allocate bits. Bits _used is called m ', and m' is larger than Bits (L). Where Bits (L) is the number of bits at the bit rate L.

In this case, since the number of bits read when the bit rate is L is Bits (L), the bits increased in the step k + 1 of (Bits (L) -m) are regarded as meaningless data when the bit rate is L, and processed during decoding. In many cases, a bit rate higher than L is used as a valid data. 7 shows a processing method for the case where the bits are located at the boundary.

Meanwhile, the method of forming the bitstream used by the decoder 24 from the scalable bitstream in the scalable bitstream recovery unit 23 is performed by the following series of steps.

1. Allocate bits for sideinfo.

2. From the side information, the number of bits BB [ch] [sb] = allocation [ch] [sb] allocated for each processing band is read.

3. Read the number of samples sample [ch] [sb] [s] for each processing band according to each BB [ch] [sb]. This step is performed by the following steps a to i.

a. Initialize to CC [.] = 0 and M = 1.

b. Find the maximum BB [ch] [sb] for all BB [ch] [sb] in the high frequency band from the low frequency band.

c. The number of bits allocated to the samples of the band having the maximum BB [.] Value is adjusted as follows.

CC [ch] [sb] = CC [ch] [sb] + 1

d. Adjust the BB [ch] [sb] value as follows:

BB [ch] [sb] = BB [ch] [sb] + 1

e. The number of bits _used (Bits _used ) is estimated as follows.

Here, SPL [.] Means a fixed number of samples for each processing band.

f. If the Bits _used value does not exceed the value of the bit rate M, the process returns to step b.

g. Read the number of samples using CC [ch] [sb].

h. The signal is adjusted using the CC [ch] [sb] value and allocation [ch] [sb].

i. Decryptor input and exit

Here, the role played in step h is to minimize an error that occurs when only a part of all allocated bits are used in any decoding process. For example, if K bits are allocated to each sample in a band, KB to K bits are used to minimize the error signal generated when K + 4 bits are allocated to each band in that band at the highest bit rate. Must be the same, and this can be scaled by moving allocation [ch] [sb] -CC [ch] [sb] to the left with data using CC [ch] [sb] values.

As described above, in the audio encoding / decoding method and apparatus capable of adjusting the scale of a bitstream according to the present invention, by forming a bitstream to be compatible with various transmission lines and various reconstruction devices, a service provided according to a user's desire can be obtained. The quality can be improved, and the service can be differentiated according to the cost paid, and the effective channel transmission and storage number of bits can be controlled in transmission and storage. In addition, it is possible to provide a variety of services to the user by enabling the use of a variety of existing devices of different types of different bit rates used.

1 is a block diagram showing a conventional audio encoder capable of scaling a bitstream.

2 is a block diagram showing an audio encoding / decoding apparatus capable of scaling a bitstream according to the present invention.

3 illustrates an audio bitstream of ISO / IEC 11172.

4 is a diagram illustrating an example of decoding a bitstream shown in FIG. 3;

5 is an example of bit number allocation.

6 is an example of a bitstream capable of scaling according to the present invention.

7 is an example of representing a bitstream to enable scaling if bits are located at a boundary.

Claims (8)

(a) encoding the received audio signal and converting the received audio signal into a bitstream; (b) classifying the audio data of the coded bitstream according to a predetermined importance level and then reconstructing the audio data according to the order of importance to convert the coded bitstream into a scalable bitstream; (c) restoring the scaleable bitstream into a bitstream suitable for decoding according to predetermined selected scale information; And and (d) decoding the reconstructed bitstream. The method of claim 1, wherein step (b) A first step of allocating bits for sideinfo; A second step of reading the number of bits BB [ch] [sb] = allocation [ch] [sb] allocated to each processing band from the side information; And And a third step of reading the number of samples sample [ch] [sb] [s] for each processing band according to each BB [ch] [sb]. Way. The method of claim 2, wherein the third step Initializing CC [.] = 0 and M = 1; Finding a maximum BB [ch] [sb] for all BB [ch] [sb] in the high frequency band from the low frequency band; Increasing the number of bits allocated to samples of the band having the maximum BB [.] By one; Increasing the BB [ch] [sb] value by 1; Estimating the number of bits used from the side information, CC [ch] [sb], and a fixed number of samples for each processing band; Returning to the step of finding the maximum BB [ch] [sb] if the used number of bits does not exceed the value of bit rate M; Reading samples of each of the processing bands by the most significant bit (MSB) by CC [ch] [sb]; And If the CC [.] = 0 is set and M is not the final bit rate, the method returns to the step of finding the maximum BB [ch] [sb]. . The method of claim 1, wherein step (c) A first step of allocating bits for sideinfo; A second step of reading the number of bits BB [ch] [sb] = allocation [ch] [sb] allocated to each processing band from the side information; And And a third step of reading the number of samples sample [ch] [sb] [s] for each processing band according to each BB [ch] [sb]. Way. The method of claim 4, wherein the third step Initializing CC [.] =: 0, M = 1; Finding a maximum BB [ch] [sb] for all BB [ch] [sb] in the high frequency band from the low frequency band; Increasing the number of bits allocated to samples of the band having the maximum BB [.] By one; Increasing the BB [ch] [sb] value by 1; Estimating the number of bits used from the side information, CC [ch] [sb], and a fixed number of samples for each processing band; Returning to the step of finding the maximum BB [ch] [sb] if the number of used bits exceeds the value of the bit rate M; Reading samples of each of the processing bands by the most significant bit (MSB) by CC [ch] [sb]; And And adjusting the magnitude of the signal by using the CC [ch] [sb] value and allocation [ch] [sb]. An encoder for encoding an input audio signal and converting the received audio signal into a bitstream; And An audio encoding including a bitstream forming unit which classifies the audio data of the encoded bitstream according to a predetermined importance and then reconstructs the audio data according to an important order and converts the encoded bitstream into a scalable bitstream Device and A bitstream recovery unit for restoring the scaleable bitstream into a bitstream suitable for decoding according to predetermined selected scale information; And And an audio decoding device including a decoder for decoding the reconstructed bitstream. The apparatus of claim 6, wherein the bitstream forming unit comprises: a control signal input unit configured to input a control signal for determining a length of the bitstream according to the bit rate for bit rate adjustment; A first adjusting unit for adjusting the length of the bitstream according to the control signal; And An audio encoding / decoding apparatus capable of adjusting a scale of a bitstream, characterized in that it comprises a second adjusting unit for adjusting the length of an entire frame. The apparatus of claim 6, wherein the bitstream recovery unit comprises: a control signal input unit configured to input a control signal for determining a length of a bitstream according to a bit rate for bit rate adjustment; A first adjusting unit for adjusting the length of the bitstream according to the control signal; And An audio encoding / decoding apparatus capable of adjusting the scale of a bitstream, characterized in that it comprises a second adjusting unit for adjusting the length of the entire frame.