JPH10336038A - Method for coding audio signal - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method by which digital audio signals can be coded with high-qualities and little delay time in the case of coding the signals. SOLUTION: Audio signals are coded at every frame having a prefixed length and each frame is constituted of a scale factor 21 and a plurality of small frames 22-25 which commonly use the scale factor 21. Therefore, the number of bits of the scale factor 21 which has been required at every frame in the conventional example is reduced and the delay time can be shortened by introducing the small frames 22-25.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for encoding an audio signal, and more particularly to a method for maintaining the quality of audio data even when encoding the audio signal by shortening the frame length.

[0002]

2. Description of the Related Art In recent years, digital audio signal encoding methods have been actively developed. Particularly, as a high quality audio encoding method, MPEG1 audio encoding (ISO / IEC 1117) has been proposed.
2-3) is known.

FIG. 5 is a diagram for explaining an MPEG1 audio encoding method. FIG. 5A is a block diagram showing a basic structure of an encoder for MPEG1 audio encoding, and FIG.
FIG. 1 is a block diagram showing a basic structure of a decoder for MPEG1 audio encoding.

In FIG. 5A, when a PCM audio sample is input to an encoder 51, a mapping unit 52 performs a filtering process on the input PCM audio sample and generates a thinned-out expression. The psychoacoustic model unit 53 generates a data set that controls quantization and encoding. Quantization and coding unit
54 performs processing according to any of the encoding methods of layer 1, layer 2 and layer 3.

[0005] The frame assembling section 55 includes a quantization and encoding section.
Assemble the actual encoded bitstream from the 54 output data and the ancillary data and provide other information (eg,
Error check) is added as necessary.

In FIG. 5B, when an encoded bit stream is input to a decoder 56, a frame decomposer 57 separates the encoded bit stream in order to restore various parts of information. The restoration unit 58 restores a quantized form of a series of mapping samples. Inverse mapping section 59 returns these mapping samples to PCM audio samples.

[0007] MPEG1 audio encoding is composed of three layers, Layer 1, Layer 2 and Layer 3. The higher the layer, the more complex the scheme, but higher quality coding is possible even at lower bit rates. FIG. 6, FIG. 7, and FIG. 8 show schematic diagrams of the bit streams of Layer 1, Layer 2, and Layer 3.

FIG. 6 shows layer 1 of MPEG1 audio encoding.
2 shows a bit stream format of the. A bit stream is composed of one or more consecutive frames. FIG. 6 shows a format of one frame. The frame includes a header 61, an error check 62, a bit allocation as first audio data 63, a scale factor and a sample, and an ancillary data 64.

FIG. 7 shows layer 2 of MPEG1 audio encoding.
2 shows a bit stream format of the. As in FIG. 6, the bit stream is composed of one or more consecutive frames. FIG. 7 shows a format of one frame. The frame is composed of a header 71, an error check 72, bit allocation as second audio data 73, scale factor selection information, scale factors and samples, and ancillary data 74.

FIG. 8 shows layer 3 of MPEG1 audio encoding.
2 shows a bit stream format of the. A bit stream is composed of one or more consecutive frames. FIG. 8 shows the format of one frame. The frame includes a header 81, an error check 82, and additional information and main information described as third audio data 83. The main information includes a scale factor, Huffman coded data, and ancillary bits. Usually, the main information starts from the previous position by the header and the additional information of the frame, and the position is specified by a value indicating the main information start position in the additional information. The number of bytes of the additional information is 17 bytes for a single channel and 32 bytes for a two channel.

[0011]

In the conventional audio signal encoding method, the length of a frame is automatically determined by the sampling frequency and the encoding layer of the audio signal. For example, when the sampling frequency is 48 KHz, In this case, the time is 8 ms for Layer 1 and 24 ms for Layer 2 and Layer 3. When the sampling frequency is 32 KHz, layer 1 has 12 m
s, 36 ms for layer 2 and layer 3. When encoding and decoding are performed, the delay time is twice as long.

This method of encoding an audio signal is a fatal defect for an application such as a wireless microphone that requires instantaneous sound pickup and loudspeaking. However, when trying to shorten the frame length unnecessarily, the overhead such as the header, the number of quantization bits, and the scale factor increases,
The number of bits allocated to the sample decreases, and the coding efficiency deteriorates. That is, there is a problem in that sound quality deteriorates when encoding is performed with the same number of bits as compared with the case where encoding is performed by a conventional audio signal encoding method.

The present invention has been made to solve the above-mentioned conventional problems, and it is an object of the present invention to code an audio signal having a short frame length, that is, a small delay time generated in encoding and decoding while maintaining the quality of audio data. The purpose of the present invention is to provide a conversion method.

[0014]

SUMMARY OF THE INVENTION In order to achieve the above object, the present invention provides a method in which a transmitting side prepares a signal before transmitting an audio signal, and a transmitting side determines a signal which has been negotiated between the transmitting side and the receiving side. , The synchronization signal in the header can be deleted.

Further, in order to achieve the above object, the present invention encodes an audio signal for each frame of a predetermined length, and the frame further includes a plurality of small frames and a scale factor. The number of bits required for the scale factor can be reduced by expressing the movement width of the bit position of the quantized audio sample by the scale factor in common with the frame.

Further, in order to achieve the above object, the present invention provides a method of dividing an audio signal into subbands or converting the frequency, and then encoding the audio signal with a fixed number of quantization bits for each frame. The number of quantization bits to be assigned to each frequency bandwidth is determined using a scale factor. When decoding, the number of bits required for the number of quantization bits is calculated by calculating the number of quantization bits of the audio sample from the scale factor. Can be eliminated.

Still further, in order to achieve the above object, the present invention is a method of encoding an audio signal for each frame after subband division or frequency conversion,
By quantizing audio samples using a variable length code, the number of bits required for the number of quantization bits can be eliminated.

[0018]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention according to claim 1 of the present invention provides an audio signal in which an audio signal is encoded and transmitted on a transmission side, and an encoded audio signal is received and decoded on a reception side. In a transmission system, when a transmitting side is in a preparatory state before transmitting an audio signal, an audio signal encoding method in which a signal determined in advance by the transmitting side and the receiving side is transmitted by the transmitting side to the receiving side. , The synchronization signal in the header can be deleted.

According to a second aspect of the present invention, an audio signal is encoded for each frame having a predetermined length, and the frame further includes a plurality of small frames and a scale factor. An audio signal encoding method comprising the steps of: sharing a scale factor between a plurality of small frames; and expressing a movement width of a bit position of the quantized audio sample by a scale factor. This has the effect that the number of bits required for the scale factor can be reduced.

The third aspect of the present invention is a method for dividing an audio signal into sub-bands or converting the frequency, and then encoding the audio signal with a fixed number of quantization bits for each frame. The number of quantization bits to be allocated to the bandwidth is determined by using a scale factor, and the quantized audio sample and the scale factor are encoded. When decoding, the number of quantization bits of the audio sample calculated from the scale factor is calculated. The method used to encode the audio signal in such a way that the audio samples were inversely quantized, and the number of quantization bits allocated to each subband or each frequency bandwidth was determined using a scale factor. The number of bits required for the number of quantization bits can be eliminated. Having an iodine.

According to a fourth aspect of the present invention, there is provided a method for encoding an audio signal for each frame after subband division or frequency conversion, wherein quantization of audio samples is performed using a variable length code. It is a method of encoding an audio signal characterized by the fact that quantization of audio samples is performed using a variable length code, so that the number of bits required for the number of quantization bits is unnecessary. Has the effect of being able to

Hereinafter, an embodiment of the present invention will be described with reference to FIG.
This will be described with reference to FIG.

(First Embodiment) FIGS. 1A and 1
FIGS. 1 (b) and 1 (c) are diagrams for explaining an audio signal encoding method according to the first embodiment, and a transmitting side that performs encoding in a preparation state for encoding performs decoding. FIG. 1A shows a sine wave, FIG. 1B shows a square wave, and FIG. 1C shows a sawtooth wave. The periods of these signal waveforms are determined in advance.

Therefore, on the receiving side for decoding, the transmitted signal waveform (one of FIG. 1 (a), FIG. 1 (b), and FIG. 1 (c)) is read for a period, and is read in advance. If the signal waveform is a predetermined signal waveform, the subsequently transmitted signal waveform is read again. If the signal waveform is not a predetermined signal waveform, decoding is started.

As described above, according to the first embodiment of the present invention, a predetermined signal waveform is transmitted from a transmitting side performing encoding to a receiving side performing decoding in a preparation state for encoding. Since synchronization is established between the transmitting side and the receiving side by transmitting a signal waveform in such a preparation state for encoding, the synchronization signal in the header required in the conventional example is deleted. be able to.

(Second Embodiment) FIG. 2 shows a second embodiment of the present invention.
14 illustrates the encoding method of an audio signal according to the embodiment, and shows a format of one frame of the encoded audio signal. One frame includes a scale factor 21 and a plurality of small frames 22 to 25. FIG. 2 shows a case where the number of small frames is four. For example, when the audio signal is divided into sub-bands and then encoded, there are scale factors equal to the number of sub-bands. To split into 32 bands
There will be 32 scale factors.

The samples in the small frames 22 to 25 are quantized with the number of quantization bits determined for each subband.

As described above, according to the second embodiment of the present invention, the scale factor is shared by a plurality of small frames, and the movement width of the bit position of the quantized audio sample is represented by the scale factor. By doing so, the number of bits required for the scale factor can be reduced.

(Third Embodiment) FIG. 3 shows a third embodiment of the present invention.
3A illustrates an encoding data flow in an encoder, and FIG. 3B illustrates a decoding data flow in a decoder. Is shown.

In FIG. 3A, the encoder of the audio signal is the same as the encoder of the conventional MPEG1 audio encoding.
When the CM audio sample is input to the encoder 31, the mapping unit 32 filters the input PCM audio sample,
Generate a thinned-out expression.

The scale factor extraction and quantization bit number calculation unit 33 divides the PCM audio sample into subbands or converts the frequency, and obtains the maximum value of the absolute value of the audio sample for each subband or frequency bandwidth. The shift width of the bit position required to normalize the sample having the largest absolute value is defined as a scale factor. Then, the number of quantization bits of the audio sample is determined based on the scale factor. For example, as a simple method, a method of allocating the number of quantization bits that can be used according to the scale factor ratio to each subband can be adopted.

The quantization and coding unit 34 performs quantization and coding processing according to the quantization bit number of the audio sample determined by the scale factor extraction and quantization bit number calculation unit 33.

The frame assembling unit 35 includes a quantization and encoding unit.
Assemble the actual coded bitstream from the 34 output data and provide other information (eg error checking, etc.)
Is added as needed.

In FIG. 3B, the decoder for the audio signal is the same as the decoder for the conventional MPEG1 audio encoding.
When the encoded bit stream is input to a decoder 36, a frame decomposer 37 reconstructs various parts of the information,
Separate the coded bitstream.

The quantization bit number calculation and inverse quantization unit 38
The number of quantization bits of each sub-band is calculated from the value of the scale factor in the separated encoded bit stream, and the value of the sample of each sub-band is read. Thereafter, the samples of each subband are dequantized using the scale factor.

The inverse mapping section 39 returns the inversely quantized mapping sample of each subband to a PCM audio sample.

As described above, according to the third embodiment of the present invention, the number of quantization bits allocated to each subband or each frequency bandwidth is determined by using the scale factor. The number of bits required for the number of bits can be made unnecessary.

(Fourth Embodiment) FIG. 4 shows a fourth embodiment of the present invention.
4 (a) shows a data flow of encoding in an encoder, and FIG. 4 (b) shows a data flow of decoding in a decoder. Is shown.

In FIG. 4 (a), the encoder of the audio signal is the same as the encoder of the conventional MPEG1 audio encoding.
When the CM audio sample is input to the encoder 41, the mapping unit 42 filters the input PCM audio sample,
Generate a thinned-out expression.

The quantization bit allocating section 43 examines the frequency characteristics of the audio samples by performing processing such as FFT on the PCM audio samples. The number of quantization bits for each subband or bandwidth is obtained from the frequency characteristics. Scale factor information may be added when determining the number of quantization bits.

The quantization and variable length coding section 44 divides the subbands or quantizes the samples for each bandwidth according to the number of quantization bits obtained by the quantization bit allocation section 43. At this time, encoding is performed using a variable length code such as a Huffman code.

The frame assembling unit 45 assembles an actual coded bit stream from the output data of the quantization and variable length coding unit 44, and adds other information (for example, an error check) as necessary.

In FIG. 4B, the decoder for the audio signal is the same as the decoder for the conventional MPEG1 audio encoding.
When the coded bit stream is input to the decoder 46, the frame decomposer 47 performs the following to recover various parts of the information:
Separate the coded bitstream.

The inverse quantizer 48 inversely quantizes the scale factor and the value of each subband sample in the separated encoded bit stream. Since the samples of each subband are variable-length coded, the samples can be dequantized without the number of quantization bits of the samples of each subband.

The inverse mapping unit 49 returns the inversely quantized mapping sample of each subband to a PCM audio sample.

As described above, according to the fourth embodiment of the present invention, since the quantization of the audio sample is performed using the variable length code, the number of bits required for the number of quantization bits is not required. can do.

[0047]

As described above, according to the present invention, when the transmitting side is in a preparation state before transmitting an audio signal, a signal determined in advance by the transmitting side and the receiving side is transmitted from the transmitting side to the receiving side, By decoding when a signal of a type different from the predetermined signal is transmitted, the synchronization signal in the header can be deleted.

Further, the audio signal is encoded for each frame of a predetermined length, and the frame further includes a plurality of small frames and a scale factor. By expressing the movement width of the bit position of the sample by the scale factor, the number of bits required for the scale factor can be reduced.

Furthermore, a method of dividing the audio signal into sub-bands or converting the frequency and then encoding the audio signal with a fixed number of quantization bits for each frame, wherein the number of quantization bits allocated to each sub-band or each frequency bandwidth is The number of bits required for the quantization bit number can be made unnecessary by determining using the scale factor and calculating the quantization bit number of the audio sample from the scale factor at the time of decoding.

[0050] Still further, a method of sub-band-dividing or frequency-converting an audio signal and encoding for each frame, wherein quantization of audio samples is performed using a variable-length code, The number of bits required for the number of bits can be made unnecessary.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining an audio signal encoding method according to a first embodiment of the present invention, and is a diagram showing a signal waveform transmitted from a transmission side to a reception side in a preparation state for encoding;

FIG. 2 is a diagram for explaining an audio signal encoding method according to a second embodiment of the present invention, showing a format of one frame of an encoded audio signal.

FIGS. 3A and 3B are diagrams for explaining an audio signal encoding method according to a third embodiment of the present invention, showing a data flow of encoding in an encoder, and FIGS. FIG. 14 is a diagram for explaining an audio signal encoding method according to the third embodiment, and is a diagram showing a data flow of decoding in a decoder.

FIG. 4A is a diagram for explaining an audio signal encoding method according to a fourth embodiment of the present invention, and is a diagram showing a data flow of encoding in an encoder, and FIG. FIG. 14 is a diagram for explaining an audio signal encoding method according to the fourth embodiment, showing a data flow of decoding in a decoder;

FIG. 5A is a block diagram illustrating a basic structure of an encoder for explaining an MPEG audio encoding method,
(B) A block diagram illustrating a basic structure of a decoder, for explaining an MPEG audio encoding method,

FIG. 6 is a diagram showing a conventional layer 1 bit stream format;

FIG. 7 is a diagram showing a conventional layer 2 bit stream format;

FIG. 8 is a diagram showing a conventional layer 3 bit stream format.

[Explanation of symbols]

21 Scale factor 22, 23, 24, 25 Quantized audio sample (small frame) 31, 41, 51 Encoder 32, 42, 52 Mapping unit 33 Scale factor extraction and quantization bit number calculation unit 34, 54 Quantization And coding units 35, 45, 55 frame assembling units 36, 46, 56 decoders 37, 47, 57 frame decomposing unit 38 quantization bit number calculation and inverse quantization units 39, 49, 59 inverse mapping unit 43 quantized bits Allocation unit 44 Quantization and variable length coding unit 48 Inverse quantization unit 53 Psychological psychology model unit 58 Restoration units 61, 71, 81 Headers 62, 72, 82 Error check 63 First audio data 64, 74 Ancillary data 73 Second audio data 83 Third audio data

Claims (4)

[Claims] An audio signal transmission system that encodes and transmits an audio signal on a transmission side, and receives and decodes an encoded audio signal on a reception side.
When the transmitting side is in the preparation state before sending the audio signal,
An encoding method of an audio signal in which a signal determined in advance by a transmitting side and a receiving side is transmitted by the transmitting side to the receiving side. 2. Encoding an audio signal for each frame of a predetermined length, wherein the frame further comprises a plurality of small frames and a scale factor, wherein the plurality of small frames comprise quantized audio samples, Wherein a plurality of small frames are used in common, and a movement width of a bit position of the quantized audio sample is represented by a scale factor. 3. A method of dividing an audio signal into sub-bands or performing frequency conversion, and encoding the audio signal with a fixed number of quantization bits for each frame, wherein the number of quantization bits allocated to each sub-band or each frequency bandwidth is The quantization is performed using the scale factor, the quantized audio sample and the scale factor are encoded, and when decoding, the audio sample is dequantized using the quantization bit number of the audio sample calculated from the scale factor. Encoding method for an audio signal. 4. A method of encoding an audio signal for each frame after subband division or frequency conversion, wherein quantization of audio samples is performed using a variable length code. Signal encoding method.