JPH11259978A - Signal processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To record/reproduce a sound faithful to the original sound by amplifying/level compressing a low level signal among digital audio signals divided into plural frequency bands and compression encoding a spectrum signal generated from the level compressed signal. SOLUTION: An A/D convertor 5 inputs the signal S1, and samples/quantizes it to output the signal S2 to a level compression part 10. The level compression part 10 amplifies selectively the low level signal. An ATRAC encoder 30 inputs the signal S3, and compresses she signal to record the signal S4 on a disk 40. At a reproducing time, an ATRAC decoder 50 reads out the signal S4 from the disk 40, and performs processing opposite to the ATRAC encoder 30, and outputs the signal S5 having a spectrum that the low level signal is selectively amplified to a level expansion part 60. The level expansion part 60 performs the processing opposite to the level compression part 10, and attenuates the selectively amplified signal to return it to an original spectrum.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention
Disc, hereinafter referred to as “MD”. ), And more particularly to signal processing during recording and playback of audio signals on the MD.

[0002]

2. Description of the Related Art In an MD, an ATRAC (Adaptive) is used in order to compress information to about 1/5 of that of a compact disc (Compact Disc, hereinafter referred to as "CD").
It uses a high-efficiency audio data encoding method called TRansform Acoustic Coding. This ATRAC
The idea is to take into account the human auditory characteristics and remove the redundant parts of the signal that cannot be perceived by humans before performing compression. In other words, on the assumption that a difference occurs between the original signal and the compressed signal (that is, quantization noise is generated), the compression is performed while controlling the noise so that a human cannot perceive it.

[0003] Specifically, ATRAC performs high-efficiency coding by utilizing an equal loudness characteristic (minimum audible characteristic), a masking effect, and the like related to human auditory characteristics. This will be described with an example. In FIG. 5A, a curve 100 is a curve showing the minimum audible characteristic of a human, and it is assumed that a sound at a level lower than this curve cannot be heard by a normal human ear.

Assuming that the input signal includes sound A and sound B in the figure, sound B is audible to human ears, but sound A is located below curve 100, so that it is heard by human ears. Is considered inaudible. Therefore, this sound B
Is recorded as an effective sound signal, but sound A is not recorded as an unnecessary signal (thinned out). Alternatively, as another method, encoding is performed by extremely reducing the allocation of quantization bits in the total number of quantization bits per unit time.

FIG. 5B shows a masking characteristic 110 of the sound C in addition to the curve 100. The sound B is located above the curve 100 showing the minimum audible limit characteristic, but is masked by the loud sound (C) in the vicinity thereof (hatched portion in the figure) and cannot be heard by ordinary human ears. In this case, ATR
In the AC, the sound B is not recorded as a redundant portion (thinned out).

[0006]

As described above, there is usually no problem even if the sound is thinned out due to human auditory characteristics, but from the viewpoint of recording and reproducing the sound as faithfully as possible to the original sound as much as possible. It is preferable that the sound signal is as small as possible. Further, there is a concept that even a sound having a sound pressure level lower than the curve 100 showing the minimum audible limit characteristic has an effect on the reproduced sound field. desirable.

The present invention has been made in view of the above points, and it has been made possible to record and reproduce a sound more faithful to the original sound by compressing a sound signal, for example, minimizing a signal decimated by ATRAC as much as possible. It is an object to provide a signal processing circuit that performs

[0008]

According to a first aspect of the present invention, there is provided a signal processing apparatus, comprising: a band dividing means for dividing an input digital audio signal into a plurality of frequency bands; Level compression means for amplifying a low-level signal among the signals included in the
Coding means for generating a spectrum signal from the level-compressed signal and compression-coding the generated spectrum signal.

According to the information processing apparatus configured as described above, an input audio signal is divided into a plurality of frequency bands, and a low-level signal among signals included in the band-divided signals is amplified. . As a result, a signal in which the level of the low-level signal is raised is generated and supplied to the encoding means. The encoding means generates a spectrum signal from the signal, and performs compression encoding according to the spectrum.

According to a second aspect of the present invention, in the signal processing apparatus according to the first aspect, the level compression unit detects a peak level of a signal included in the band-divided signal, and detects the detected peak level. A gain adjustment circuit that determines an amplification degree for each level signal included in the band-divided signal based on a level, and an amplifier that amplifies the signal according to the determined amplification degree. I do.

According to the signal processing device configured as described above, the peak level in the band-divided signal is detected, and the amplification degree for amplifying the signal of each level is determined accordingly. Therefore, only a low-level signal can be selectively amplified with an appropriate amplification degree, and the level-compressed signal spectrum is prevented from significantly changing from the original signal spectrum.

According to a third aspect of the present invention, in the signal processing device according to the first or second aspect, the encoding means comprises an AT.
It is configured to be a RAC encoder.

[0013]

Preferred embodiments of the present invention will be described below with reference to the accompanying drawings.

[0014] Rationale First, a description will be given of the basic principle of the present invention with reference to FIG. FIG. 1 is a diagram showing a flow of recording and reproduction of an audio signal in an MD according to the present invention in comparison with a flow of recording and reproduction of an audio signal in a normal MD.

First, normal processing will be described. In the figure, the spectrum indicated by the symbol D indicates that the audio signal to be recorded on the MD is first divided into a predetermined band (one of a low band, a middle band, and a high band), and the frequency-divided signal is about 10 mS. It shows the frequency characteristics of one of the short blocks. A
In the TRAC, the spectrum data obtained in this way is further divided into coding units for each two-dimensional minute area of frequency and time, and the redundant signal is thinned out using the above-described minimum audible characteristics and masking effect. . In addition, E1 in the spectrum indicated by the code E indicates the maximum level in a predetermined coding block, and normalization is performed for each coding unit at this level, so that the level above the line E1 in the spectrum indicated by the code E is obtained. The redundant part of is removed. The portion below the line E2 in the spectrum is a redundant portion of the signal that is removed in consideration of human auditory characteristics. Therefore, as shown in the figure, a signal having a level lower than the line E2 is decimated. Only the signal thus compressed, that is, the signal between the lines E1 and E2 in the spectrum indicated by the symbol E is recorded on the MD.

In the spectrum of the reproduced signal, as shown by a symbol F, a signal having a small level (a level that is inaudible to human ears) is thinned out (a signal marked with "x" in the figure). Spectrum), many signals are missing as compared with the spectrum indicated by reference symbol D.

Next, an audio signal recording process according to the present invention will be described. In the present invention, as shown in a spectrum indicated by a code G, a normal AT is used in each coding unit.
The RAC selectively amplifies a signal of a level which will be decimated. As a result, as in the spectrum indicated by the symbol H, the signal which is at a level close to the audible level but is lower than the audible level in the minimum audible level and the masking characteristic, but exceeds the audible level, (I.e., goes above line H2 of the spectrum indicated by the symbol H due to amplification). Then, after amplifying the low-level signal in this way, the signal is supplied to an ATRAC encoder, encoded, and recorded on the MD.

At the time of reproduction, a process of attenuating the signal amplified at the time of recording by the reverse characteristic is performed for each of the same encoding units (spectrum indicated by J) of the recorded signal (spectrum indicated by J). With the above processing, as can be seen by comparing spectra F and J, ATRAC
In this case, the number of signals thinned out as inaudible to human ears is reduced, and a reproduced signal having a spectrum closer to the original sound can be obtained.

MD Recording / Reproducing Apparatus Next, the configuration and operation of the MD recording / reproducing apparatus according to the present invention will be described with reference to FIGS. Figure 2 shows the MD
FIG. 2 is a block diagram showing a configuration relating to signal recording and reproduction processing of the recording / reproduction device 1.

An analog input signal S 1, which is an audio signal to be recorded on an MD disk 40 (hereinafter simply referred to as a “disk”), is first input to the A / D converter 5. The input signal S1 is a signal from an audio source (not shown), and has a sampling frequency of 44.1 kHz, 16
The data is sampled and quantized by bits and supplied to the level compression unit 10 as a digital signal S2.

The level compression section 10 is a block for selectively amplifying a low-level signal described in the above description of the principle. That is, in the ATRAC encoder 30 at the subsequent stage, a sound of a level that would be thinned out so far as to be inaudible to the human ear due to the minimum audible characteristics and masking characteristics is amplified to some extent in advance, and the ATRA
A process for reducing the number of sounds to be decimated in the C encoder 30 is performed.

FIG. 3 shows the configuration of the level compression section 10. The digital input signal S2 is supplied to a low-pass filter 11 and band-pass filters 12, 13, respectively, and is divided into a low-frequency signal and a plurality of signal bands. Each signal band is preferably matched for each frequency of the coding unit in the subsequent ATRAC encoder. Alternatively, three bands that are first roughly divided into bands in the ATRAC encoder (that is, low band: 0 to about 5.5 kHz, middle band: about 5.5 to 5.5)
11 kHz, high frequency: about 11 to 22 kHz). In addition, as each filter, QMF (Quadrature M
It is preferable to use an irror filter. This is because an aliasing component generated by band division can be canceled at the time of band synthesis.

The signals in each band are supplied to level detection circuits 15, 1
6, 17 and the amplifiers 21, 22, 23
Respectively. Also, the input signal S which is not band-divided
2 is supplied to the level detection circuit 14 as it is. The outputs of the level detectors 15, 16, and 17 are output from the gain adjustment circuits 18, 1
9 and 20, respectively. Gain adjustment circuit 18, 1
The outputs of 9 and 20 are input to the amplifiers 21, 22 and 23 as amplification adjustment input signals.

The level detection circuit, the gain adjustment circuit, and the amplifier perform a process of adjusting (amplifying) the gain of the band-divided signal according to its level. Here, since the processing performed for each band is the same, the level detection circuit 1
5. The operation of the gain adjustment circuit 18 and the amplifier 21 will be described as a representative. The level detection circuit 15 continuously monitors the level of the input low-frequency signal, detects a peak (maximum) level therein at predetermined time intervals, and supplies the value to the gain adjustment circuit 18. The gain adjustment circuit 18 determines a gain adjustment characteristic of the amplifier 21 according to the given peak level. This gain adjustment characteristic is shown in FIG. In FIG. 4A, the horizontal axis indicates the level detected by the level detection circuit, and the vertical axis indicates the degree of amplification for each level signal. In the figure, a characteristic 25 in which the amplification degree is a positive value
Is a gain adjustment characteristic in the level compression unit 10 (a characteristic 26 in which the amplification degree is a negative value corresponds to the level expansion unit 6 described later).
The gain adjustment characteristic at 0 is described later.

The gain adjustment circuit 18 includes the level detection circuit 1
The characteristic 25 is applied so that the peak level detected by 5 coincides with the peak level point 27 in FIG. 4A, and the amplification degree for each level (digital value) of the input signal is determined. Specifically, in the gain adjustment characteristic of FIG. 4A, the detected peak level is set to 30%
The signal of the level up to the level is amplified at the maximum amplification degree _αmax , and as for the signal of the level of about 30% to 50% of the peak level, the amplification degree decreases as the level increases. Also, a signal having a level of about 50% or more of the peak level is not amplified. The characteristic 25 is determined in consideration of the minimum audible characteristic described above.

By using such characteristics, a relatively high level signal among signals in each band is originally AT
Since it is not thinned out in RAC, the level is kept as it is without performing amplification. On the other hand, a signal that would normally be decimated by ATRAC due to its relatively low level is amplified to increase its level, thereby reducing the decimated signal. That is, as shown in the spectrum indicated by the symbol H in FIG.
The number of signals to be decimated can be reduced by placing it in the upper part of 2.

Next, the operation of the level detection circuit 14 will be described. The level detection circuit 14 detects a peak level in the entire band signal of the input signal S2, and according to the peak level, determines a gain adjustment characteristic 2 of each amplifier shown in FIG.
5 is changed. Specifically, the value of the maximum amplification degree α _max shown in FIG. 4A is changed according to the detected peak level in the entire band signal. FIG. 4B shows a characteristic 28 for changing the maximum amplification degree α _max . The horizontal axis indicates the peak level of the entire band signal detected by the level detection circuit 14, and the vertical axis indicates the maximum amplification _αmax . As can be understood from the figure, when the peak level of the entire band signal is close to zero, the maximum amplification α _{max is set} to zero (that is, no amplification), and as the peak level increases, the maximum amplification α
_The value of _max is increased, and when the value exceeds a certain value, the maximum amplification degree α _max is set to a certain value.

Such characteristics are obtained for the following reasons. If the peak level of the full band signal is close to zero, it is considered that the audio data itself is small, so if this is amplified, most of the noise will be amplified,
Also, the signal spectrum after amplification will be significantly different from that of the original signal. Further, even when the peak level of the entire band signal is large, if too much amplification is performed only on the low-level signal, the signal is different from the original signal. Therefore, the characteristic as shown in FIG. 4B is adopted, and only a signal which has a level close to the audible range but would normally be thinned out is amplified to some extent. Thus, each of the amplifiers 21, 2
The signals amplified by 2 and 23 are added in an adder 24, and a signal S3 of the entire band is output.

The signal S3 is input to the ATRAC encoder 30. The ATRAC encoder 30 divides the input signal S3 into three band signals of a low band, a middle band, and a high band. Furthermore, after clipping into short blocks of about 10 mS (maximum 11.6 mS), it is divided into coding units for each two-dimensional minute area of frequency and time, and each coding unit is modified discrete cosine transform (MDCT) or the like. To the spectral data, and performs signal compression according to the minimum audible characteristics while considering the energy distribution and the like in the signal by signal analysis.
Record at 0.

On the other hand, at the time of reproduction, the recorded signal S4 is read out from the disk 40, and the ATRAC decoder 50
Is input to ATRAC decoder 50, ATRAC
It performs a process reverse to that of the encoder 40 and outputs a reproduction signal S5. The reproduction signal S5 has a spectrum in which a low-level signal is selectively amplified, such as a spectrum indicated by reference numeral I (and a spectrum indicated by reference numeral H) in FIG.

Next, the level extending section 60 outputs the reproduced signal S
5 is received, and a process reverse to that of the level compression unit 10, that is, a process for attenuating the selectively amplified signal and returning it to the original spectrum is performed. The configuration of the level extending section 60 is shown in FIG.
Is basically the same as the configuration of the level compression unit 10 shown in FIG.
That is, the reproduction signal S4 is divided into three bands, and the peak level is detected for each band. However, unlike the level compression section 10, in the level expansion section 60, each of the amplifiers 21, 22, 23
Attenuates the input signal for each band. The attenuation characteristic at this time is a characteristic 26 shown in FIG. 4A, which is steeper than the amplification characteristic of each of the amplifiers 21, 22, and 23 of the level compression unit 10.

Further, similarly to the level compression section 10, the level detection circuit 14 detects the peak level of the entire band signal, and changes the value of the maximum amplification (attenuation) α _max of the attenuation characteristic 26 according to the detected level. Thus, an output signal S6 having the spectrum J of FIG. 1 is generated. In the output signal S6, as described in the section on the principle, the number of signals decimated by the ATRAC encoder is reduced. (A spectrum indicated by reference symbol J in FIG. 1) is obtained.

In the above description, the level compression unit 10
Performs the level compression by dividing the input signal into bands corresponding to the encoding units of the ATRAC encoder 30, but the application of the present invention is not limited to this. That is, level compression can also be performed by dividing into three different bands or three or more bands. However, in this case, it is necessary to configure the level expansion unit 60 that performs the reverse process to perform the same band division.

As described above, according to the present invention, A
Since signals to be thinned out in the TRAC processing can be reduced, recording and reproduction more faithfully to the original signal can be performed. Also, by reducing the number of signals to be decimated, quantization noise caused by encoding is reduced.

[0035]

As described above, according to the first aspect of the present invention, a signal which is originally at a low level and which is thinned out at the time of encoding is amplified before encoding. It is possible to reduce a signal to be decimated at the time of encoding, and to encode a signal having a more faithful spectrum than the original signal.

According to the second aspect of the present invention, it is possible to selectively amplify only a low-level signal with an appropriate amplification factor, and the signal spectrum after the level compression greatly changes from the original signal spectrum. Is prevented.

According to the third aspect of the present invention, ATRAC
The encoder can reduce a signal that is thinned out as inaudible to human hearing, thereby enabling more accurate recording of the original sound. In the ATRAC encoder, the bit allocation is performed in the coding unit based on the signal of the maximum level in the unit. According to this, the bit allocation ratio to the low-level signal is reduced. However, in the invention described in the claims, since the bit allocation amount is determined after amplifying the level of the low-level signal, the bit allocation amount is expected to increase compared to when the signal is kept at the low level. Thus, the quality of the signal itself is improved.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating the basic principle of signal processing according to the present invention.

FIG. 2 is a block diagram showing a configuration of a signal processing unit of the MD recording / reproducing apparatus according to the present invention.

FIG. 3 is a block diagram illustrating a configuration of a level compression unit and a level expansion unit illustrated in FIG. 2;

FIG. 4 is a diagram showing level adjustment characteristics in a level compression section and a level expansion section shown in FIG. 2;

FIG. 5 is a diagram illustrating the concept of compression processing in ATRAC.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... MD recording / reproducing apparatus 5 ... A / D converter 10 ... Level compression part 11 ... Low-pass filter 12, 13 ... Band-pass filter 14,15,16,17 ... Level detection circuit 18,19,20 ... Gain adjustment circuit 21 , 22, 23 ... Amplifier 24 ... Adder 30 ... ATRAC encoder 40 ... Disk 50 ... ATRAC decoder 60 ... Level expansion unit 70 ... D / A converter

Claims (3)

[Claims] 1. A band dividing means for dividing an input digital audio signal into a plurality of frequency bands; a level compressing means for amplifying a low-level signal among signals included in the band-divided signal; An encoding unit that generates a spectrum signal from the signal and compression-codes the generated spectrum signal. 2. The level compression means comprises: a detection circuit for detecting a peak level of a signal included in a band-divided signal; and a level included in the band-divided signal based on the detected peak level. The signal processing device according to claim 1, further comprising: a gain adjustment circuit that determines an amplification degree for the signal; and an amplifier that amplifies the signal according to the determined amplification degree. 3. The signal processing device according to claim 1, wherein the encoding unit is an ATRAC encoder.