JPH06252773A - High efficient coder - Google Patents

Abstract

PURPOSE:To compress an audio signal with various sampling frequencies effectively in response to a human audible sense characteristic and to effectively utilize a transmission line by variable-control-bit allocation algorithm in the case of compression coding of an input signal depending on a sampling frequency of the input signal. CONSTITUTION:An FS discrimination circuit 11 of a DSP (encoder) 10 measures a sampling frequency of an input digital audio signal and a compression table selection circuit 12 selects a fixed table in response to sampling frequency information of a measured input signal from plural fixed tables including a prepared different allocation bit components in advance. That is, an optimum fixed table is selected for use in response to a sampling frequency when an information quantity is shared. Then a compression rate selection circuit 13 and a compression coding circuit analyze a spectrum of the input signal to share bits with respect to each block. Thus, the optimum compression in response to the content of the input signal is attained.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a high-efficiency coding apparatus for compressing and coding a digital signal obtained by sampling an analog signal.

[0002]

2. Description of the Related Art In recent years, various signal compression coding techniques have been proposed and put to practical use in order to record a long-duration digital audio signal on, for example, a small-sized optical disk or magnetic tape suitable for carrying. ing.

As an example of this signal compression coding technique,
There is a technique of expanding a signal on the time axis into a spectrum on the frequency axis to perform band compression, and for example, there is a technique for compressing and recording a digital audio signal of 74 minutes on an optical disc having a diameter of 64 mm. More specifically, in this signal compression coding algorithm for audio signals, the entire band of the input audio signal sampled at, for example, 44.1 kHz is divided into 52 blocks on the frequency axis, and each block is divided into 52 blocks. First, bit allocation is performed according to a predetermined fixed table, and then the number of blocks to be recorded on the optical disc and the bit allocation for each block are changed according to the content of the signal. It is possible to compress 1/4 to 1/5 of the digital signal of.

Here, in the above compression coding algorithm, when changing the number of blocks or bit allocation to each block, processing is performed in consideration of human auditory characteristics. In other words, since the high frequency components of the audio signal are almost inaudible to the human ear, the bit allocation to this high frequency component (block) is reduced, or the block itself is deleted (not recorded. Do). For example, as described above, when the sampling frequency is 44.1 kHz, the human ear hardly hears a component (frequency component) of 15 kHz or higher, and thus the high frequency block (15 kHz or higher) of the above 52 blocks is not heard. For the component, the bit allocation is lowered or the block data itself is not used (not recorded).

[0005]

By the way, if the signal compression coding technique as described above is used in the field of digital signal communication, more information (voice and acoustic information) can be obtained in the transmission path of the conventional transmission rate. Can be transmitted.

However, in the field of digital signal communication, the sampling frequency is generally 32.
kHz and 48 kHz are used. Since the sampling frequency is different from the sampling frequency of 44.1 kHz in the signal compression coding technique described above, the following adverse effects occur.

That is, in the signal compression coding technique,
As described above, in consideration of human auditory characteristics, the number of allocated bits for the high frequency block is reduced or the high frequency block itself is not transmitted, but the sampling frequency is 32 kHz. The high-frequency block at that time corresponds to 11 kHz to 16 kHz. Since the band of 11 kHz to 16 kHz falls within a band that is sufficiently audible to humans, if the number of allocated bits for quantization in the block in this band is reduced or the high frequency block itself is not transmitted, Will be recognized as a deterioration of sound quality.

The sampling frequency is 48 kHz.
The above high frequency range at z corresponds to 16 kHz to 24 kHz. Therefore, considering that the human ear hardly hears 15 kHz or more as described above, the audio signal having the sampling frequency of 48 kHz is transmitted more than the audio signal having the sampling frequency of 44.1 kHz. It is possible to reduce the number of bits to be used. This means that more information can be transmitted at the conventional transmission rate, and thus the above-mentioned sampling frequency is 44.1.
If the signal compression coding technique for a kHz audio signal is directly applied to an audio signal having a sampling frequency of 48 kHz, the transmission line cannot be effectively used.

Therefore, the present invention has been proposed in view of the above situation, and audio signals of various sampling frequencies can be effectively compressed according to human auditory characteristics, and the transmission line can be effectively used. It is an object of the present invention to provide a high-efficiency coding device that can be used.

[0010]

The present invention has been proposed in order to achieve the above-mentioned object, and is measured (or manually switched) by sampling frequency measuring means for measuring the sampling frequency of an input signal. It has an algorithm control means for variably controlling the bit allocation algorithm according to the sampling frequency of the input signal, and a compression encoding means for performing compression encoding of the input signal based on the bit allocation algorithm by the algorithm control means.

Here, the algorithm control means selects a fixed table corresponding to the sampling frequency of the input signal from a plurality of fixed tables prepared in advance, each of which contains different allocated bit components. Variable control is performed. In particular, for an input signal having a sampling frequency lower than the reference frequency, a bit allocation algorithm in which the number of allocated bits on the high frequency side is increased is used.

[0012]

According to the present invention, the bit allocation algorithm at the time of compression coding of the input signal is variably controlled according to the sampling frequency of the input signal, so that the optimum compression according to the content of the input signal is possible. Becomes

[0013]

Embodiments of the present invention will be described below with reference to the drawings.

As shown in FIG. 1, the high-efficiency coding apparatus of the embodiment of the present invention comprises an FS judging circuit 11 which is a sampling frequency measuring means for measuring the sampling frequency of an input digital audio signal, and the FS judging circuit 11 described above. The compression table selecting circuit 12 and the compression ratio selecting circuit 13 as the algorithm controlling means for variably controlling the bit allocating algorithm according to the sampling frequency of the input digital audio signal measured in 1., and the input based on the bit allocating algorithm by the algorithm controlling means. A DSP (Digitial Signal Proces) including a compression encoding circuit 14 that performs compression encoding of a digital audio signal.
sor) 10.

In FIG. 1, an analog audio signal is supplied to the terminal 1. The analog audio signal is converted into a digital signal (digital audio signal) by the A / D converter 2. here,
The sampling frequency in this A / D converter 2 is, for example, 32 kHz, 44.1 kHz, 48 kHz.
It is assumed to be either.

In the signal compression coding algorithm described above, the spectrum expanded on the frequency axis
Each block is divided into blocks (quantization unit, Quantize-Unit: QU), but in the case of this embodiment, the frequency assigned to each block varies depending on the sampling frequency as shown in Table (1) of Table 1.

[0017]

[Table 1]

Next, the audio signal converted into a digital signal by the A / D converter 2 is supplied to the input terminal 3 of the apparatus of this embodiment. The digital audio signal via the input terminal 3 is the DSP 10 of the apparatus of this embodiment.
That is, it is sent to the encoder. Note that the connection between the A / D converter 2 and the DSP 10 is normally performed serially. Therefore, in the case of the present embodiment, the A / D converter
The transmission rate changes depending on the sampling frequency of the D converter 2.

The input digital audio signal from the input terminal 3 is supplied to the FS judging circuit 1 of the DSP 10 concerned.
Sent to 1. The FS determination circuit 11 measures the sampling frequency of the input digital audio signal as described above. Here, since the DSP normally receives an input signal by an interrupt, in the present embodiment, by measuring the interval of this interrupt using a timer or a counter inside the DSP 10,
The sampling frequency of the currently input digital audio signal is determined. That is, in this embodiment, the timer or counter inside the DSP 10 is used as the FS determination circuit 11.

The FS determination circuit 11 outputs the measurement result of the sampling frequency measurement (sampling frequency information) and the input digital audio signal,
Compression table selection circuit 12 of the algorithm control means
Sent to. The measurement of the sampling frequency is
In addition to the automatic operation as described above, for example, by providing a changeover switch that outputs a signal corresponding to each sampling frequency, and by manually switching this switch,
It is also possible to obtain the sampling frequency information.

The compression table selection circuit 12 is provided with a plurality of fixed tables (which will be described later in detail) including different allocation bit components prepared in advance. The compression table selection circuit 12 has a plurality of fixed tables. From the fixed table, a fixed table corresponding to the sampling frequency information of the input digital audio signal obtained by the FS determination circuit 11 is selected.

That is, in the signal compression coding algorithm, the amount of information (the amount of information to be transmitted) according to a fixed table considering human auditory perception characteristics is allocated to the spectrum included in each block. However, when allocating this amount of information, the apparatus of the present embodiment does not use one fixed table, but uses a fixed table optimal for each sampling frequency prepared in advance by switching according to the above-mentioned sampling frequency. I am trying to do it. In this embodiment, as the fixed table corresponding to each of the sampling frequencies, a table having the allocated bit amount as shown in FIG. 2 is prepared for each block in the entire band. Further, in the present embodiment, as shown in FIG. 2, among the fixed tables corresponding to the respective sampling frequencies, a sampling frequency lower than a reference frequency (for example, 44.1 kHz in the present embodiment) is used. As the fixed table (32 kHz in the present embodiment), the number of allocated bits on the high frequency side is increased in consideration of human auditory characteristics.

Next, the compression rate selection circuit 13 analyzes the spectrum of the input digital audio signal, and when the sampling frequency of the input digital audio signal is 44.1 kHz or 48 kHz, the components are concentrated particularly in the high frequency range. If not, the 44th block (QU) and above (15 kHz and above) shown in Table 1 are omitted (not transmitted), and the bits allocated in the fixed table to the omitted block. Is allocated to each of the other blocks, and the amount of transmission information for each of the other blocks is increased to improve the sound quality of the signal portion corresponding to the other block.

On the other hand, the sampling frequency is 32k.
The 44th block (QU) in the case of Hz
The spectrum included in each of the above blocks is well within the audible band, and if these are omitted, the sound quality will have an extremely narrow dynamic range. Therefore, the compression rate selection circuit 13 (and compression encoding) of the present embodiment. In the circuit 14), at the sampling frequency of 32 kHz, bits are allocated so that all blocks are transmitted without fail. As a result, good sound quality can be ensured even in the high frequency range.

That is, in the compression rate selection circuit 13 (and the compression encoding circuit 14), specifically, the spectrum analysis (quantization sub information calculation) of the input digital audio signal is performed by the following procedure. , Bits are allocated to each block.

The procedure of spectrum analysis (calculation of quantization sub-information) in the compression rate selection circuit 13 (and compression encoding circuit 14) is as follows: scale factor calculation, maximum scale factor detection, tonality calculation, bandwidth It is in the order of determining, assigning bits, and adjusting fractions of assigned bits.

First, in the calculation of the scale factor,
The scale factor ID in each block (QU) is obtained by using the formulas (1) and (2) shown in the formulas 1 and 2.
For example, the scale factor ID (IDSF (k)) of the k-th block (QU _k ) is the minimum scale factor ID (((SFK (k))) that is greater than or equal to the Max value (PK (k)) of the absolute value of the spectrum in that block. IDOFSF).

[0028]

[Equation 1]

[0029]

[Equation 2]

Next, in the detection of the maximum scale factor,
The maximum scale factor ID is calculated using Equation (3) of Equation 3.
The ID (k ₀ ) of the block (QU) taking However, when there are some that take the maximum value, the above k ₀
Is the smallest of them.

[0031]

[Equation 3]

In the calculation of the tonality, whether the input signal is tone-like by using the equations (4), (5), (6) and (7) of the equations (4), (5), (6) and (7) Or is it noise-like,
The tonality T indicating the degree is calculated from the scale factor distribution in the vicinity of the peak band.

[0033]

[Equation 4]

[0034]

[Equation 5]

[0035]

[Equation 6]

[0036]

[Equation 7]

In determining the band, the numbers (8) and (9)
As in (9), the block (Q
U _k0 ), and the tonality T, the number L of quantized sub-information is determined.

[0038]

[Equation 8]

[0039]

[Equation 9]

The bit allocation is as follows. First, in the range of 0 ≦ k ≦ L, the numerical formulas (10), (11), and (12) are used.
1), B ₀ (k) is calculated using the equation (12). In these equations, DIVMODE (k) is the orthogonal transform size of the modified discrete cosine transform (MDCT), and IDVAR (k, D
IVMODE (k)), IDFIX (k, DIVMODE
The values of (k)) are shown in Tables (2) and (3) of Tables 2 and 3.

[0041]

[Equation 10]

[0042]

[Equation 11]

[0043]

[Equation 12]

[0044]

[Table 2]

[0045]

[Table 3]

Here, the tables (2) and (3) are tables as shown in table (4) of table 4 on software.
Further, Table (5) in Table 5 is a table for the sampling frequency of 32 kHz. That is, Table (4) above is the fixed table for sampling frequencies 44.1 kHz and 48 kHz, and Table (5) is the fixed table for sampling frequency 32 kHz.

[0047]

[Table 4]

[0048]

[Table 5]

Next, X is calculated using the following formulas (13), (14), (15), and (15), and B
Let the sum of ₀ (k) and X be the word length represented by a real value. NSPS (k) represents the number of spectra in each block (QU).

[0050]

[Equation 13]

[0051]

[Equation 14]

[0052]

[Equation 15]

Finally, using the mathematical expression (16) of equation 16, B
(k) is converted to IDWL (k) representing the word length.

[0054]

[Equation 16]

In adjusting the fraction of the allocated bits, the total number of bits allocated by the above method is, for example, 1032.
If it exceeds 952 (when L = 44) or 952 (when L = 52), I is set from the high frequency side so that the value becomes less than this value.
Decrease DWL (k) one by one. At this time, IDWL
If (k) = 0, the number of bits cannot be reduced.

If IDWL (k) = 1, then setting it to 0 reduces 2 × NSPS (k) bits.
At other times, by reducing IDML (k) by 1,
NSPS (k) bits are reduced.

The total number of bits assigned by the above method is 1032 (when L = 44) or 952 (L).
= 52), the IDWL (k) is increased by 1 from the low frequency side so as to approach these numerical values. At this time, if IDWL (k) = 15, the number of bits cannot be increased. If IDWL (k) = 0, setting it to 1 increases 2 × NSPS (k) bits.

At other times, by increasing IDML (k) by 1, NSPS (k) bits are increased.

An example of bit allocation in this embodiment described above will be described with reference to FIGS.

First, in FIG. 3, the sampling frequency is 32.
The state of bit allocation at the time of kHz is shown. Figure 3
A indicates the bit allocation amount for each block based on the fixed table selected by the compression table selection circuit 12 when the sampling frequency is 32 kHz, and b in FIG. 3 indicates the bit allocation by the variable table (IDVAR).
Is shown. Note that in FIG. 3B, since the calculation is actually performed at one time like B ₀ (k) of bit allocation in the above-described quantization sub-information calculation procedure, the allocation graph is not a step-like curve but a curved line. Get closer. This is because it takes real values, not integer values. Also, in FIG.
As shown in FIG.
As shown by the dotted line in the figure, reduce the bit allocation amount from the high range that is hard to hear, and stop the reduction if sufficient.

4 and 5, the sampling frequency is 4
The figure shows the state of bit allocation in the case of 4.1 kHz and 48 kHz. 4a and 5a show the bit allocation amount for each block based on the fixed table selected by the compression table selection circuit 12 when the sampling frequencies are 44.1 kHz and 48 kHz, and b in FIG. And b in FIG. 5 show sampling frequencies 44.1 kHz and 48 k.
Bit allocation by variable table in case of Hz (ID
VAR) is shown. In addition, in FIG. 4b and FIG. 5b, unlike the case where the sampling frequency is 32 kHz, when the peak is larger than a certain amount and it is not located in the upper part of the block (QU), The bit allocation of is zero (0). In reality, since a negative assignment occurs during the calculation, it is almost never 0. In addition, in c of FIG. 4 and c of FIG. 5, as in the case where the sampling frequency is 32 kHz, if the bit allocation is over, the band is reduced from the high band as shown by the dotted line in the figure. At the end, the negative allocation is naturally 0.

Returning to FIG. 1, thereafter, from the compression rate selection circuit 13, for example, the input digital audio signal, the sampling frequency information obtained by the FS determination circuit 11, and the compression table selection circuit 12 are transmitted. The selected fixed table information and the compression rate information in the compression rate selection circuit 13 are output and sent to the compression encoding circuit 14.

The compression encoding circuit 14 performs compression encoding of the input digital audio signal based on the fixed table selected by the compression table selection circuit 12 and the compression rate information in the compression rate selection circuit 13. .

The signal from the DSP 10 is output via the terminal 4. The block (QU) of the signal output from the terminal 4 is as shown in a and b of FIG. 6 according to each sampling frequency. FIG. 6a shows blocks with sampling frequencies of 48 kHz and 44.1 kHz. It should be noted that the sampling frequency 4 shown in FIG.
In the case of blocks of 8 kHz and 44.1 kHz, whether to transmit the 44th to 51st blocks depends on the signal component. The sampling frequency 32k is shown in FIG.
The block of Hz is shown. In the case of the block having the sampling frequency of 32 kHz shown in FIG. 6B, the number of blocks is fixed at 52. As described above, in this embodiment, the total amount of information to be transmitted is changed so that data of different sampling frequencies can be transmitted at a fixed transmission rate.

Next, the configuration of the decoding device corresponding to the high efficiency coding device of this embodiment will be described.

Returning to FIG. 1 again, D on the encoder side
The signal from the SP 10 is sent to the DSP 20 on the decoder side via the terminals 4 and 31. The signal supplied to the DSP 20 is sent to the FS information detection circuit 22. The F
The S information detection circuit 22 separates the sampling frequency information and the compression-coded audio signal, and the decompression circuit 21 separates the compression-coded audio signal.
The sampling frequency information is sent to the timing generation circuit 36 via the terminal 32.

The expansion circuit 21 expands and decodes the compression-encoded audio signal by using the expansion ratio and expansion table corresponding to the compression ratio and fixed table (compression table) selected in the configuration on the encoder side. To do. The audio signal (digital audio signal) expanded and decoded by the expansion circuit 21 is sent to the D / A converter 34 via the terminal 33.

The D / A converter 34 uses the clock based on the sampling frequency information formed by the timing generation circuit 36 to convert the digital audio signal into an analog audio signal. This analog audio signal is sent to a sound reproducing device such as an amplifier or a speaker via the terminal 35.

As described above, according to the high-efficiency coding apparatus of this embodiment, the sampling frequency is automatically determined by measuring the interval of the input data, and the compression coding for the optical disk described above is performed. By preparing three types of tables and three types of compression rates for the same algorithm as the method, the sampling frequency is 44.1 kHz.
Not only z but also 32 kHz and 4 used in the field of communication
Even with a sampling frequency of 8 kHz, optimum compression that does not cause a hearing problem can be performed, and a fixed transmission rate can be used by making the compression rate variable.

FIG. 7 shows a schematic configuration of, for example, a signal transmission system to which the high efficiency coding apparatus according to the above-described embodiment of the present invention is applied.

In FIG. 7, in the transmission signal source 110, for example, a video signal reproducing device 111 such as a so-called VTR (video tape recorder) or video disc player.
And an audio signal reproducing device 112 such as a so-called CD (Compact Disc) player or tape recorder, and T
A computer data output device 113 such as V game software is provided. Also, a so-called PA control device 11 having a function for in-flight broadcasting using a microphone, etc.
6 is provided.

Among the output signals from these devices, the audio or audio signals from the video signal reproducing device 111 and the audio signal reproducing device 112 and the computer data from the computer data output device 113 are sent to the audio signal transmitting device 120. Has been. The video signal from the video signal reproducing device 111 is sent to another video signal transmitting device (not shown), and the audio signal from the PA control device 116 is directly sent to a speaker (not shown) provided on the inside wall or ceiling of the aircraft. Audio signal transmitting device 12
It has also been sent to 0.

The audio signal transmitting device 120 includes, for example, up to four audio signal transmitting units 121A and 12A.
1B, 121C, 121D, and one audio signal transmission unit 121, for example, considers the audio characteristics of 32 channels and the data signals of 8 channels (or audio signals of 36 channels) in consideration of human auditory characteristics. The signals are compression-encoded to 1/4 to 1/5 (the high-efficiency encoding apparatus of this embodiment is applied) and multiplexed, and the so-called RF modulator 122 modulates and outputs an RF signal. An RF mixing circuit for inputting and mixing RF output signals from other audio signal transmitting units is provided in the RF modulator 122. Specifically, each of the audio signal transmitting units 121A, 121B, 121C, The RF signals of different carrier frequencies are output for each 121D after being frequency-multiplexed. In this case, the video signals of a plurality of channels from the video signal reproducing device 111 are also modulated by RF modulators (not shown) at different carrier frequencies, frequency-multiplexed together with the RF modulated signal of the audio signal, and a single coaxial cable is used. It may be transmitted.

The frequency-multiplexed RF signal is sent to the RF branching device 131 of the zone management unit 130 provided for each zone that divides the inside of the aircraft into zones, and the RF branching device 13 is provided.
1 through the RF amplifier 132 to the zone management unit of the next zone. The RF signal input to the RF branching device 131 is branched and sent for each column which is a divided area for dividing the zone. A plurality of seat units 140 are connected in series in one column, and the R input to the seat unit 140 is input.
The F signal is sent to the next seat unit 140 via the RF branching device 141 and the RF amplifier 142.

In one seat unit 140, for example, 3
Circuits for seats are collectively provided, and the RF signal branched by the RF branching device 141 is divided into three RF signals by the RF distributor 144, and three RF receiving units 150 are provided.
A, 150B, 150C respectively. The RF receiving unit 150 extracts the RF signal of one band of the compression-encoded and frequency-multiplexed signal and outputs the RF signal.
After demodulating and decompressing the demodulated signal, the signal of the desired channel is extracted and output. The signals output from the RF receivers 150A, 150B, and 150C are output to the passenger-side control units 146A, 146B, and 146.
Sent to C respectively. Passengers in each seat can attach headphones 14 to each control unit 146A, 146B, 146C.
7A, 147B, and 147C are connected to each other, and an audio signal of a desired channel can be arbitrarily selected and heard.

[0076]

According to the high efficiency coding apparatus of the present invention,
The sampling frequency of the input signal is measured, the bit allocation algorithm is variably controlled according to the sampling frequency of the input signal, and the compression coding of the input signal is performed based on the bit allocation algorithm. The audio signal can be effectively dealt with, and the transmission line can be effectively used.

[Brief description of drawings]

FIG. 1 is a block circuit diagram showing a schematic configuration of a high efficiency encoding device (encoder) and a decoding device (decoder) according to an embodiment of the present invention.

FIG. 2 is a diagram for explaining a fixed bit allocation amount according to a sampling frequency of the encoder of the present embodiment.

FIG. 3 is a diagram for explaining bit allocation when the sampling frequency is 32 kHz.

FIG. 4 is a diagram for explaining bit allocation when the sampling frequency is 44.1 kHz.

FIG. 5 is a diagram for explaining bit allocation when the sampling frequency is 48 kHz.

FIG. 6 is a diagram for explaining blocks to be transmitted.

FIG. 7 is a block circuit diagram showing a schematic configuration of a signal transmission system as a specific example to which the device of this embodiment is applied.

[Explanation of symbols]

 2 ... A / D converter 10 ... DSP (encoder) 11 ... FS determination circuit 12 ... Compression table selection circuit 13 ... ...... Compression ratio selection circuit 14 ...... Compression coding circuit 20 ・ ・ ・ ・ ・ DSP (decoder) 21 ・ ・ ・ ・ ・ Expansion circuit 22 ・ ・ ・ ・ ・..FS information detection circuit 34 ..... D / A converter 36 ........ Timing generation circuit

Claims (3)

[Claims] 1. An algorithm control means for variably controlling a bit allocation algorithm according to a sampling frequency of an input signal, and a compression coding means for compressing and coding an input signal based on the bit allocation algorithm by the algorithm control means. A high-efficiency encoding device having. 2. The variable of the bit allocation algorithm is selected by the algorithm control means by selecting a fixed table according to the sampling frequency of the input signal from a plurality of fixed tables prepared in advance that include different allocation bit components. The high efficiency encoding device according to claim 1, wherein the high efficiency encoding device performs control. 3. The algorithm control means uses a bit allocation algorithm in which the number of allocated bits on the high frequency side is increased for an input signal having a sampling frequency lower than a reference frequency. Alternatively, the high-efficiency encoding device according to item 2.