SU1624702A1 - Device for coding and decobing broadcasting signals - Google Patents

Description

control pulse driver outputs, the first output of the synchronization unit is connected to the synchronization inputs of the second code converter, the first and second buffer memory blocks,

the inverse Fourier transform unit and the control pulse driver, and the second output of the synchronization unit is connected to the inputs of the input frequency of the first and second code converters.

The invention relates to radio engineering and communications and can be used in digital television stereo systems, digital broadcasting systems and multi-functional communications, in terrestrial and satellite digital long-distance systems of sound broadcasting programs, in digital systems for recording, storing and reproducing sound signals.

The aim of the invention is to increase the bandwidth while maintaining high quality playback of audio broadcast signals during decoding.

In this device, in each of the 24 critical bands of hearing, the spectral component with the largest absolute value of the real part and the spectral component with the largest absolute value of the imaginary part are identified. For these spectral components, values of the order of the code with a block-wise floating point are selected, and then the spectral components of this critical band are encoded. At the same time, the Km mantissa width is chosen so as to ensure the necessary value of the coding error of the largest spectral component. The discharge factor Kp must provide the ability to cover the entire range of the spectral component, which is equal to 16 bits. The coding errors of the rest of the spectral components may exceed a predetermined value, since they will still not be perceived by the hearing due to the effect of masking inside the critical bands of hearing.

Fig. 1a shows a structural electrical circuit of the encoder; in fig. 16 is a decoder circuit of a device for encoding and decoding audio broadcast signals; 2 and 3, the code converters of the encoder and the decoder, respectively; 4 shows timing diagrams for the operation of the device.

The device for encoding and decoding audio broadcast signals contains an encoder, which includes a low-pass filter (LPF) 1, an analog-to-digital signal

Creator (ADC) 2, buffer storage unit 3, block 4 of the direct Fourier transform, first 5 and second 6 code converters, driver 7

control pulses (UI) and a synchronization unit 8, each of the code 5 converters (6) contains a buffer storage unit 9, a floating-point code encoder 10, a detector of the 11th order of the maximum element and a code order register 12, and a decoder, which includes the first 13 and second 14 code converters, the first 15 and second 16 buffer storage blocks, the driver 17 MD,

an inverse Fourier transform unit 18, a digital-to-analog converter (D / A converter) 19, a low-pass filter 20, and a synchronization unit 21, each of the code converters 13 (14) containing a first key 22, 4-bit

order code register 23, comparator 24, second key 25, first 26 and second 27 registers of the mantissa code, key block 28, output 16-bit register 29, third key 30, and counter 31.

The device for encoding and decoding audio broadcast signals operates as follows.

The operation of the ADC 2, buffer storage unit 3, unit 4 of the direct Fourier transform, the first 5 and second 6 code converters and the imaging unit 7 UI is synchronized with a clock signal of 16 Rd, arriving at the corresponding inputs of these blocks from the first output

synchronization unit 8, from the second output of which the output clock frequency Pout is fed to the inputs of the first 5 and second b code converters.

Analog signal of the individual

the audio broadcasting channel is fed to the input of the low-pass filter 1, which limits the frequency band of the input signal to Pb 15 kHz, from the output of which the signal is fed to the input of the ADC 2 where a sample is taken

a signal with a sampling frequency of Rd 32 kHz, quantization and coding of the sampled samples by a 16-bit linear code. Formed in the ADC 2 code words counts with a clock frequency

Pd is fed to the input of the buffer storage unit 3 with a capacity of 2N 16-bit code words, which delays the sequence of code words by 2IM / FA in order to match the output of the A / D converter 2 to the input of block 4 of the forward Fourier transform, processing blocks of N code words . After accumulating in the buffer storage unit 3 2N code words, the code words corresponding to the first N samples of the broadcast signal are successively read and fed to the input of block 4 of the forward Fourier transform, in which the spectrum of a sequence of N samples is calculated based on the fast Fourier transform algorithm i.e. in the time interval AT M / RD, a current spectral analysis of the broadcast signal is performed.

From the output of block 4, the direct Fourier transform with a clock frequency Rd, the 16-bit codewords of the real and imaginary parts of the complex discrete decomposition coefficients in the Fourier series are successively removed. The signal length of N samples corresponds to N / 2 + 1 complex decomposition coefficients.

Хт - Хт + 1Хт (t 0N / 2) located

on the frequency axis evenly with a step of D f Ffl / N Hz. The first coefficient X0 corresponds to the frequency f0 0, and the last Xy / 2 corresponds to the frequency fw / 2 Rd / 2. The first and last decomposition coefficients are always valid, i.e. Ho 0 and Hm / 2 0, and the remaining coefficients - Xi, ..., XN / 2-1 - are complex. Thus, the spectrum of a broadcast signal with a length of N samples is determined by N numbers represented in a 16-bit linear code, where the bits of the code word of real parts are marked with right-slope shading, and the bits of imaginary parts are left-sloped.

In the first 5 and second 6 converters of the encoder code, the 1-bit linear code, respectively, of the real and imaginary parts of the spectral components is converted into a block-floating-code code, i.e. code used in pulse code modulation with almost instant companding. In this case, the difference is that, firstly, instead of a sequence of samples of audio broadcast signals taken in time, a sequence of real or imaginary parts of its spectral components comes in, and secondly, the length of the block is encoded with a constant value (scale factor) is variable and is set by the control signals to the four control inputs of the first 5 and

the second 6 code converters from the corresponding outputs of the imaging unit 7 (FIGS. 4 b, c, d, e, f, g, h, u), respectively.

Positive pulses taken from the outputs of the imaging unit 7 (Fig. 4b, e) open the inputs of the buffer storage units 9 of each code 5 converter (6) at those times when the code words of the real and imaginary parts of the spectral components arrive. Signals arriving at the other inputs of each converter code 5 (6) from the outputs of the imaging unit 7 (FIG. 4, g) control the reading of information from the 11 order detectors in the order code registers 12 and in the floating-code encoders 10. The position of the control pulses in time is determined as follows. The read pulses must occur at those times tm when the code words of the real and imaginary parts of the spectral components belonging to the same frequency group are accumulated in the buffer storage unit 9. The m-th frequency group includes Im real and Im imaginary parts of the spectral components, and in the first and last 24 frequency groups the number of real parts is one more than the number of imaginary parts of the spectral components, i.e. Ii - Ii + 1, 24 124 + 1, and in the other frequency groups Im Gt

(t 2.3 23). Therefore, the beginning

From the th read pulse during the processing of the -th block

tm (§ I, + Г /) / РД + 1ЦШ. J 1hfl

The read pulses for the first 5 and second 6 encoder converters begin at the same times tm. From the other outputs of the imaging unit 7, pulses go in order to read the bits of the order from code registers 12 of the order (FIG. 4 d), and pulses to control the reading of code words from the buffer storage units 9 and to the floating code coders 10 comma (Fig.4 d, p) of the first 5 and second 6 code converters.

The floating point encoders 10 select from the 16-bit code words of the mantiss bits in accordance with the code words entered into them (figure 4k), where the bits of the mantiss code word of the real parts of the spectral components are marked with hatching. with right slope, and bits of mantis of imaginary parts - with left-slope shading, control words are about

are marked with double hatching, and first the code word follows the order of the real parts of the spectral components, and then the imaginary parts.

Pulse frequency at the output of code converters 5 and 6

MOUTH (2-24-Kp + Km / N) Rd / N.

where Kp is the order of the order; CT is the mantissa size. The operation of the first 13 and second 14 converters, the first 15 and second 16 buffer storage blocks, the 17 IC generator, the 18 Fourier transform block and the 19 DAC decoder 19 is synchronized with a 16 Rd clock signal fed to the corresponding inputs of the 3iwx blocks from the first synchronization block output, with the second output of which the clock frequency Pwx Pvh of the following code bits with a floating point unit is fed to the corresponding inputs of the first 13 and second 14 code converters.

A digital stream of code bits with a floating point block is fed to the information inputs of the first 13 and second 14 code converters, which convert the code words of a code with a floating point block of real and imaginary parts of the spectral components into code words of a 16-bit linear code. The control inputs of the first 13 and second 14 code converters receive pulses from the corresponding outputs of the imaging unit (Fig. 4 l, m, n, u, t, y, f) controlling the first 22, second 25 and third 30 keys of the corresponding code 13 transducers (14). The pulses (Fig. 4 l, m, n) open the corresponding keys 22,25 and 30 of the first transducer of code 13 in order to separate the order, the mantissa bits (without the sign bit) and the sign bits from the received digital stream, respectively. d) mantissa code of the real parts of the spectral components. The pulses (Fig. 4 t, y, f) of the former 17 open the corresponding keys 22.25 and 30 of the second transducer of code 14, which is processing the code words of the imaginary parts of the spectral components.

The order code bits are written to a 4-bit code register 23, the order from which the parallel code is fed to the inputs of the comparator 24, to the other inputs of which the signal is output from the outputs of the counter 31, to the first input of which pulses are received from the first output of block 21 synchronization with a clock frequency of 16 Rd. At the output of the comparator 24, a signal is generated that controls the key block 28 and simultaneously sets the counter to the initial state.

The bits of the mantis code with a clock frequency Pout are input to the KT-bit first register 26 of the mantissa code, from where they are copied to the places KT the lower bits of the 15-bit second register 27 mantissa code in the parallel code,

0 where they are shifted with a frequency of 16 Ps by the required number of ticks, determined by an order code. At a certain point in time, the control signal opens the key block 28 and the contents

5 of the second register of the mantissa code in the parallel code is rewritten into the output 16-bit register 29. During the rewriting of the bits of the mantissa from the first register 26 of the mantissa code in the second register 27

The 0 code of the mantissa, its sign bit from the output of the third key 30, is written into the output 16-bit register 29, in which the code word of the 16-bit code of samples is formed.

5C of the outputs of the first 13 and second 14 n code converters, the recovered 16-bit code words of the samples are fed to the inputs of the first 15 and second 16 buffer storage blocks, which are pre-assigned to match the record to the inverse Fourier transform unit 18. To the control inputs of the first 15 and second 16 buffer storage units from the corresponding driver outputs

5 and 17, pulses are received that control the writing of code blocks to the corresponding buffer storage units 15 (16) and their reading (FIG. 4, p, p, c). The capacity of the first buffer storage unit

0 15 is 2x16 (N / 2 + 1) bits, the capacity of the second buffer storage unit is 16-2 x 16 (N / 2-1) bits.

The Fourier inverse transform unit 18 reverses the Fourier transform, as a result of which a block of N samples, represented in a 16-bit linear code, is reconstructed from the block of spectral components. From the output of block 18, the inverse Fourier transform

0 16-bit code words of the samples are fed to the input of the DAC 19, which restores the quantized samples of the original signal, the output of which is connected to the low-pass filter 20, smoothing the quantized samples of the audio broadcast signal and, thus, recovering the original analog signal.

The proposed audio encoding and decoding device provides sound quality

Audio broadcasting signals are not worse than the quality code, with the digital signal speed of the original digital signal, the current samples at the output of the device are almost 2 times of which are encoded by a 16-bit line lower than in the known device.

one

Phage /

f -i

t-th 8klock (/ for is-bit codebeds

IMS ... MSS

. lt p gi-i-„. ...

(m-r) Јlock / п-йц, & ТОК

a ojir ... jnjTjarmjr., ... JT

 ........ and .., ... lies

(m-i) nd 6 / th "t-ush lok (

KSHDY ... WJ M AW -% w, .®sh ... 1tshzhshsh ..

n ... JTJIJIAJIJT ..,

n JLJl ... ...., .... l Ji J iLL JL l JU IL

((u S / tOKtn-blU $ / IOK

about Tjnj., njnjnjnJ4n .., UT ..., JTn

lm-tl-th SAOK) 1st 5klok

, rijnJTrTri ..,. runj LJ :: iЈr

lm-0th & goof

Block

...- U- .r. lm-fr block

TJTJl..irbf ... JTTiriJ-l ..

1L th, & ioK

...... ..: TJTjn rLr ... LTLrLn

PP

Rj. .TJ1TLTLTIП .... GG1 JT

Jl. ...... jlj.

Fia.Z

fm.jjl th Bad

...

(shch) th flea

G ... JU1. ...

jn

Block

Claims (1)

AUDIO BROADCASTING SIGNAL CODING AND DECODING DEVICE, comprising an encoder, which includes a series-connected low-pass filter and an analog-digital converter, as well as a synchronization unit, the first output of which is connected to the synchronizing inputs of the analog-to-digital converter and the first code converter, and a decoder, which includes the first code converter, a digital-to-analog converter and a low-pass filter connected in series, as well as a synchronization unit, the first output which is connected to the synchronizing inputs of the first code converter and digital-to-analog converter, characterized in that, in order to increase throughput while maintaining high quality playback of audio broadcasting signals during decoding, sequentially connected buffer storage unit and direct Fourier transform unit are introduced into the encoder , as well as a second code converter and control pulse generator, while the outputs of the analog-to-digital converter are bitwise connected to the the corresponding inputs of the buffer storage unit, and the outputs of the direct Fourier transform unit are bitwise connected to the information inputs of the first and second code converters, to the control inputs of which the corresponding outputs of the control pulse generator are connected, the first output of the synchronization unit is connected to the synchronizing inputs of the buffer storage unit, the direct Fourier transform unit , a second code converter and a control pulse generator, and the second output of the synchronization unit is connected to Input frequency of the first and second code converters, the outputs of which are combined and are the output of the encoder, and the second code converter, two buffer memory blocks, the inverse Fourier transform block and the control pulse generator are introduced into the decoder, while the information inputs of the first and second code converters are combined and are the input of the decoder, and the outputs of the first and second code converters are bitwise connected respectively to the corresponding information inputs of the first and second buffer apominayuschih blocks whose outputs are connected to respective inputs of the inverse Fourier transform unit, the outputs of which are connected to respective inputs of a digital to analog converter to control inputs of first and second code converters and the first and second buffer memories corresponding to the blocks connected 1624702 A1 are the output outputs of the control pulse generator, the first output of the synchronization block is connected to the synchronizing inputs of the second code converter, the first and second buffer storage blocks, the inverse Fourier transform block and the control pulse generator, and the second output of the synchronization block is connected to the input frequencies of the first and second converters code.