FIELD OF THE INVENTION

The present invention relates to data processing technology and particularly to a method for compressing data files of either text, image, sound or video format. The invention may be used in data transmission, in devices for storage of information, in audio and video control systems, as well as for the encryption of information. BACKGROUND OF THE INVENTION

It is known the method for data compression using the algorithm LZ characterised by replacement of most often occurred sequences of symbols or lines in a stream being transmitted with references to "samples" stored in a specially created table or dictionary (Method of dictionaries, 1977. Algorithm Lempel-Ziv (LZ)). The disadvantage of this method is a low degree of data compression.

The prototype of the present method is a method characterised by a logical re- multiplication of values of bits of the input data stream by values of coefficients of a reference function with a following converting of the obtained amplitudes into a smooth quasi-harmonic signal (Belarus patent No. BY8423U, 2012).

The disadvantage of the prototype is that the digital input data stream is compressed into an analogue signal and the criterion of compression is the frequency band of an analogue signal. As a result, the compressed signal has low interference protection, and is not widely used in telecommunication systems for the reason of being in the analogue format.

A device disclosed in BY8423U including the compression-decompression processor, digital-to-analogue and analogue-to-digital converters is selected Ffr the prototype of the present multiprocessor converter. The disadvantages of the device are:

a) limited application due to the usage of analogue data format; b) low noise and interference protection;

c) limited data compression degree.

SUMMARY OF THE INVENTION

These disadvantages are eliminated by the present invention so that in a method for multiple data compression, based on the superposition of values of the data bit stream and corresponding values of coefficients of the basic function followed by the summing of the values obtained, the approximation of the sequence of amplitudes of input data stream converted into a low frequency signal is used for the file compression, and the interpolation of the compressed signal is used for the file decompression.

This goal has been realised by additionally incorporating an approximation block and an interpolator into the multiprocessor converter comprising the compression- decompression processor.

DESCRIPTION OF DRAWINGS

Further the invention will be described in more detail with reference to drawings wherein:

Fig. 1 shows the structure of a multiprocessor converter;

Fig. 2 represents the basic function of type sinX/X used to compress the data;

Fig. 3 shows an uncompressed signal;

Fig. 4 shows the compressed data stream. DETAILED DESCRIPTION OF A PREFERRED EMBODIMENT

The multiprocessor converter 1 comprises a compression-decompression processor 2, approximator 3 and interpolator 4.

The method is being realised as follows. The data to be compressed is entered in a serial stream to the Input I of the compression-decompression processor 2 which in accordance with the compression algorithm converts it into a series of amplitude counts. The basis function Y(t) is calculated so that its duration T_ is significantly longer than the sampling interval T. As a result, the superposition of the data Xj being transmitted and amplitudes counts of the basis function of type Y (t) = sinX/X will represent the pulses overlapping in time. The resulting signal U _k, i is mathematically described as follows: where k is a serial number of a reference Y(t);

i = 1, 2, 3,... is a number of a group of a data binary code being transmitted for compression;

Q is a number of groups T_ can be divided by.

The sampling interval T is calculated in advance as follows:

T = T /Q , (2)

and is an internal parameter of efficiency of the multiprocessor converter.

According to Kotelnikov 's theorem, the initial data stream represented by a smooth quasi-harmonic signal shown in fig. 3 will be restored with sufficient degree of accuracy if values of coefficients U _k, i are taken at frequency of times intervals Tl, and additionally

Tl < l/2-F _max , (3) where F _max is maximum frequency in the spectrum of quasi-harmonic signal U _ki i .

In accordance with the compression algorithm

T « T1 , (4) therefore the signal U _ki i is being approximated in each section by a block 3 through its entire length. The degree of compression C is thus defined as: C = T1/D T , (5) where D is a dynamic range of a signal U^. Thus, a value of a coefficient of a compressed data stream will be detected at the Output II at each time point Tl.

The reverse data recovery, i.e. the decompression is realized through the Input II of the multiprocessor block 1 by the interpolator 4. Decimated values of counts access the input of decompressor (block 2) in series from the output of interpolator 4, where the decoding takes place in accordance with the algorithm. The recovered data goes to the source file through the Output II.

The present method makes possible the multi-compression of the compressed file resulting in the increase of the compression degree by several times.