WO2023200349A1

WO2023200349A1 - Linguistic multiplexing

Info

Publication number: WO2023200349A1
Application number: PCT/RU2022/000113
Authority: WO
Inventors: Александр Валерьевич ИВАНОВ
Original assignee: Александр Валерьевич ИВАНОВ
Priority date: 2022-04-14
Filing date: 2022-04-14
Publication date: 2023-10-19

Abstract

The proposed invention is a method for primary multiplexing of digital data transmission channels and consists in that a multiplexible sequence is transformed in a specific way, as set out in the claims and disclosed in the description of the proposed invention, which makes it possible to multiplex transmission channels that use low-redundancy encoding schemes.

Description

LINGUISTIC CONDENSATION

DESCRIPTION OF THE INVENTION

Field of technology. The present invention relates to methods for primary multiplexing of digital data transmission channels.

Note - Hereinafter, the term “primary compaction” refers to the compaction (English, multiplexing), which is achieved by the mechanisms inherent in the transmission channel initially, that is, at the planning stage of the design of such a channel, in contrast to the “secondary compaction”, which, as a rule, it is considered in relation to communication lines and implies densification of an existing line.

State of the art. There is a method for compressing the main binary data transmission channel with an additional binary data transmission channel, known (method) from the following sources:

[1] Jerchen Kuo, Gerry Pesavento, Multiplexing an additional bit stream with a primary bit stream. — US Patent No. 6,624,736 B2 dated September 23, 2003, application No. 10/245,854 dated September 17, 2002.

[2] Jerchen Kuo, Clock distribution in a communications network. — Continuation of the patent [1]. — US Patent No. 6,775,300 B2 dated 08/10/2004, application No. 10/435,431 dated 05/09/2003.

If we consider the data stream transmitted over such a channel statically in time as a conditionally endless text, that is, as a kind of continuous vocabulary sequence, in each position of which there is one word, then we can express:

- the number of different words that can be placed in any position of the compacted (main, initial) sequence, conventionally called “clear in meaning” - an integer N> 2;

— the number of different words that can be placed in any position of the condensed (resulting) sequence, conventionally called “controversial in meaning,” — an integer N** > N.

Using such metrics, you can see that the well-known method according to [1] and [2], based on redundancy of coding of the form xB/yB, where y > x, has the following limitation (disadvantage):

N** > 2 ^m x N, t > 1 . (1)

The present invention is aimed at overcoming the limitations inherent in the above method and is applicable when:

Also, there are ways to organize a combined binary data transmission channel, which can be considered as the primary compaction of the user data transmission channel by one or more channels for transmitting various service data, including those necessary for the operation of the combined channel itself, for example, such as those involved in the 10GBASE-T, 100GBASE-R, 1000BASE-T1 protocols, known from the following sources, respectively:

[3] IEEE Std 802.3an-2006 // DOI: 10.1109/IEEESTD.2006.231802.

[4] IEEE Std 802.3bj-2014 // DOI: 10.1109/IEEESTD.2014.6891095.

[5] IEEE Std 802.3bp-2016 // DOI: 10.1109/IEEESTD.2016.7564011.

As follows from their specifications, these methods have the same limitation (1) as the method according to [1] and [2].

Note - It is worth noting that for methods according to [1], [2], [3], [4], [5], restriction (1) is only a necessary, but by no means sufficient condition for their applicability, while for of the proposed invention, limitation (2) fully establishes both the necessary and sufficient conditions for its applicability.

Based on the foregoing, the author believes that the proposed invention does not have an analogue of the degree of similarity at which such an analogue can be considered a prototype of the invention.

Disclosure of the invention. To ensure simplicity and clarity of further presentation, it is convenient to express:

- the number of different words different from words that are clear in meaning, conventionally called (words that are different from clear in meaning) “figurative” within the meaning” — an integer N* >1;

- the set of all words that are clear in meaning and allowable for writing the sequence being compressed - by a discrete (finite) set T with capacity (number of elements) |T| = N

- the set of all figurative words allowed for writing a compressed sequence - a discrete set T* of cardinality |T*| = N*, defined such that T P T* = 0;

- the set of all words of controversial meaning that are acceptable for writing a condensed sequence - a discrete set T** of cardinality |T**| = N**, defined such that T + T* = T**;

— constraint (2), taking into account the above, first reduced to the inequality N** > N + N* and then, again taking into account the above, to an equality that defines its (the restrictions) optimal, that is, the minimum case according to 7V**, — in the form of a ratio:

#** = + #* . (3)

This approach allows us to further disclose the proposed invention in the most compact and natural way for it.

§ 1. To solve the problem of overcoming the limitation (1) inherent in the methods according to [1], [2], [3], [4], [5], in the proposed invention -

■ a compacted vocabulary sequence,

■ ■ composed exclusively of words that are clear in meaning,

■ ■ ■ whose location in it is uniquely determined by their positions, ■ modified in such a way that

■ (firstly, the so-called “click” phase)

■ ■ one vocabulary group, called replaceable,

■ ■ ■ associated with the occurrence of a compaction event,

■ ■ ■ composed of a selected number of initial words,

■ ■ ■ ■ located at positions selected relative to the event,

■ ■ replaced by a vocabulary group,

■ ■ ■ corresponding to the combination of source words and the event itself,

■ ■ ■ placing the same number of words in the same positions,

■ ■ ■ ■ of which at least one is figurative in meaning,

■ (secondly, the so-called “echo” phase)

■ ■ another vocabulary group, called recountable,

■ ■ ■ associated with compensation of the compaction event,

■ ■ ■ composed of a selected number of initial words,

■ ■ ■ ■ located at positions selected relative to the event,

■ ■ ■ ■ ■ different from the positions of the words of the group being replaced,

■ ■ recalculated by bundle

■ ■ ■ source words located in the selected positions,

■ ■ ■ ■ which belong to the recalculated group,

■ ■ ■ source words located in the selected positions,

■ ■ ■ ■ which belong to the group being replaced,

■ ■ ■ the event itself,

■ ■ provided that

■ ■ ■ at the word positions of the recalculated group

■ ■ ■ ■ after recalculation it is permissible to place

■ ■ ■ ■ ■ both clear and figurative words.

The main technical result of this solution is the ability to compress digital data transmission channels, using coding schemes that are less redundant than those corresponding to constraint (1), such as those corresponding to constraint (2) and, above all, its optimal case (3).

Let us give examples of the implementation of such an invention.

Example 1A. As described above, the proposed invention can be implemented using two - one for each phase - basic operations, conventionally called “replacement” and “recalculation” (substitution and representation, respectively).

Let's assume that:

T = {a, b, c, d}, A= |T| = 4,

T* = {e*} , #* = |T*| = 1,

T** = {A, B, C, D, E} , A** = |T**| = 5 = 7V + V*.

Elements of sets are encoded according to the following rule:

That's why:

T A T* = 0 T + T* = T**

Let us assume that there is a single compacting event that leads to an effect, conventionally called a “click” (previously English, snap), on the sequence by replacing one of its words like this:

Such a click generates a certain disturbance, conventionally called an “echo” (previously English, echo), whose numerical value is limited by the value 1/E, where E is the echo modulus, which in the case under consideration is equal to:

E = E^ = N/ N* = / 1 = 4.

The echo module indicates how much data - expressed numerically as a data module - is not encoded in the vocabulary group being replaced due to the replacement and therefore should be additionally encoded in the recalculated dictionary group when recalculating. Therefore, the number of words in the vocabulary group being counted, n, is selected according to the following rule:

[#**]"> £ X [TV]" .

What does this inequality mean? The expression on its left side corresponds to a unit of data that can be encoded in n words of a compressed sequence. The expression on its right side corresponds to the module of data that is CUMULATIVELY encoded in the echo and the same number of words of the compressed sequence. A comparison between these sides shows the condition for sufficient redundancy of data modules - for the selected number of words n n = 1 5' = 5 < 16 = 4*4' // small and = 2: 5 ² = 25 < 64 = 4*4 ² // small n = 3 5 ³ = 125 < 256 = 4*4 ³ // small n = 4: 5 ⁴ = 625 < 1,024 = 4*4 ⁴ // small n = 5: 5 ⁵ = 3,125 < 4,096 = 4*4 ⁵ // small n = 6: 5 ⁶ = 15,625 < 16,384 = 4*4 ⁶ // small n = T. 5 ⁷ = 78,125 > 65,536 = 4*4 ⁷ // redundant

// redundant...

Let us further assume that n = 7 is chosen.

Let us assume that in the compacted sequence, at the positions of the words of the replaced and recalculated vocabulary groups, the following one + seven = eight (1 + 7 = 8) words are located:

(original text)

replacement is the only word

(source text) b - > ? ~ recount is the first word

(source text) with -> ? ~ recount is the second word

(source text) d — > ? ~ recount is the third word (source text) a - > ? recount - word four

(source text) b - > ? recount - word five

(source text) with -> ? recount - word six

(source text) d — > ? recount is the seventh word

Let's move from symbolic to numeric notation and represent the data as an eight-bit number in positional form in base N = 4: a,b,c,d,a,b,c,d —> 0123 0123( ₄₎ .

Let's recalculate the resulting eight-digit number in base 4 into a seven-digit number in base N** = 5:

0123 0123(4, ^ 0210 224(5) .

Let's move back from numeric to symbolic notation and get the changed words of the recalculated vocabulary group:

0210 224 ₍ 5) -► A, C, B, A, C, C, E.

Thus, the implementation of replacement (indicated by “|”) and recalculation (indicated by “~”) translates the original words into the modified ones like this:

Taking into account the previously specified encoding rule, the following entries of the words of the event-compressed sequence are identical: e*, A, C, B, A, C, C, E = E, A, C, B, A, C, C, E = e*, a, c, b, a, c, c, e* .

The right entry above reveals that, as agreed, the words of the recalculated group become both clear and figurative in meaning (prev. English, clear and noted, respectively), and the only word of the replaced group becomes only figurative in meaning.

As a result, in the compacted sequence, at the positions of the words of the replaced and recalculated vocabulary groups, the following, again one + seven = eight words, will be located:

[condensed text] E = e* (<— a) “click” - one word

[condensed text] A = a (<— b) “echo” is the first word

[condensed text] С = с (<— с) “echo” is the second word [condensed text] B = b (<— d) “echo” is the third word

[condensed text] A = a (<— a) “echo” is the fourth word

[condensed text] С = с (<— b) “echo” is the fifth word

[condensed text] С = с (<— с) “echo” is the sixth word

[condensed text] E = e* (<— d) “echo” is the seventh word

This also shows that basic] 2 operations, applied in conjunction with the corresponding rules, keep the number of words in the sequence unchanged, and at the same time, allow both the data of the compressed sequence and the compressing event to be transported in a compressed sequence, thereby compressing such a transmission channel.

Summarizing the above, we express the compaction process characteristic of the case considered in the following brief form as follows:

Example 1B. Let's take example 1 A as a basis, but assume that now there are five different (distinguishable) compacting events, so the replacement is carried out according to the following rule: at event "1": x y - * e * a, where x, y belong to T; at event “2”: x -> e*b, where x, y belong to T; at event “3”: x -> e*c, where x, y belong to T; at event “4”: x_y - ► e*d, where x, y belong to T; at event “5”: xy -> e*e*, where x, y belong to T.

Among other things, this reveals that, as agreed, among the words of the group being replaced, in at least one position after the replacement there is always a word that is figurative in meaning.

The replaced group consists of two words, each of which has a data module equal to N = 4, which is completely excluded from coding in these positions during replacement, since after the replacement one of the five events that caused the compaction will be encoded in these positions. As earlier, To prevent the loss of such data, they must be encoded during recalculation. The echo module in the case under consideration will be equal to:

According to rule (4), derived earlier, we determine the minimum number of words in the recalculated group, p p < 11: >< u = 12: 5 ¹² = 244,140,625 < 268,435,456 = 16*4 ¹² p = 13: 5 ¹³ = 1 220 703 125 > 1 073 741 824 = 16*4 ¹³ p > 14: >>

Further actions are similar to example 1A, taking into account the fact that now the replaced group consists of two (pairs) of words, the recalculated group consists of (assuming that) thirteen (u = 13) words, and the recalculation itself is carried out by converting the corresponding 15-bit number by base N = 4 into a 13-bit number in base N** = 5, compensating for the increased modulus of the echo, equal to NOT = 1/16.

The compaction process characteristic of the present example is modified accordingly and, in the same form, now looks like this:

Example 1B. Let's continue example 1B, but now reduce the number of different compacting events to two (below, z belongs to T): for event “1”: z a -> e*a, zb ->e*a;

at event "2":

The echo module in the case under consideration is reduced to a value that collectively (multiplicatively) corresponds to one whole and one half of the word data module in the compressed sequence:

E = E1b = A*(A/2) = 4-(4/2) = 8.

Further actions are similar to examples 1A and 1B, taking into account the fact that now the number of words in the recalculated group can also be reduced according to rule (4): w < 8 : >< n = 9 : 5 ⁹ = 1,953,125 < 2,097,152 = 8-4 ⁹ // small n = 10: 5 ¹⁰ = 9,765,625 > 8,388,608 = 8-4 ¹⁰ // redundant w > 11: >> . . . .

The compaction process in this example changes accordingly and looks like this:

1/1 — >• replacement{2} — 1/8 recalculation{10} 1/1 .

§ 2. If it is necessary to reduce the delay of encoding (decoding) data when transmitting (receiving), then in the proposed invention -

■ recalculation is carried out in several approaches.

The main technical result (see § 1 earlier) remains the same with this solution. An additional technical result with this solution is to reduce the size of the buffer required to accumulate words of the compressed (and, accordingly, compressed) sequence, which reduces the delay in encoding (decoding) data on the transmitting (receiving) side.

Let us give examples of the implementation of such an invention.

Example 2A. As shown above, the proposed invention can be implemented by recalculating not a single operation on the number of words w _D , but several operations conventionally called “approaches” (echo multiplexing] round), each of which is performed on a smaller number of words l _E < n _D .

Conversion into several approaches is based on the ability to cover the corresponding data module with the product two or more integers greater than one. If such a possibility is absent, recalculation in several approaches is not feasible.

Let's take example 1A as a basis and note that z is required to convert into one unified approach? _D = n according to rule (4) = 7 words per approach, that is, the same as for the entire recount - since there is only one approach.

Now let us note that the data module of the word of the compressed sequence N = |T| = 4 we represent the product of two integers greater than one (in the case under consideration, also the second power of the smallest of all prime numbers - the number two):

The number of words in the vocabulary group being recalculated, n, for the recalculation approach with the echo module £ _B x and £ out BEFORE and after recalculation, respectively, is selected according to the following rule:

[#**Г > ^ _ВХ x [А]^ £out . (5)

Rule (5) is arithmetically equivalent to rule (4), if we assume that E = E _ъх / /-out, but is semantically different in that it explicitly takes into account both the amount of data that must be additionally encoded for the approach under consideration, and the amount of data which is excluded from encoding in this approach and therefore must be additionally encoded in the next approach when performing recalculation.

Let us assume that for the first (/ = 1) approach: in,1 ⁼ E\ = N= 4,

Then for the following (z = 2) approach:

^in,2 ⁼ out,1 = N / [d] = 2,

Eve ,1 ⁼ output / d = N / [<7*<7] = 2 / 2 = 1 .

Thus, recalculation can be done with only two (k = 2) approaches, limited accordingly by rule (5), which gives the same calculations for both:

exactly n > 5: . >

Next, let's assume that 7 = 4 is chosen.

From here we get that only n _E = n is required according to the rule (5) = 4 < /? _D words for each approach H fc x « _E = 2 >< 4 = 8 > « _D words for the entire recount.

Reducing n s /7 _D TO n _E correspondingly reduces the size of the encoding (decoding) buffer. An increase in the total number of words in the recalculated vocabulary group does not affect the reduced size of such a buffer, and is therefore allowed as insignificant.

Suppose that in the compacted sequence, at the positions of the words of the replaced and recalculated vocabulary groups, the following one + two by four = nine (1 + 2x4 = 1 + 8 = 9) words are located:

(original text)

replacement is the only word

(source text) a - > ? ~ recount is the first word

(source text) b - > ? ~ recount is the second word

(source text) with -> ? ~ recount is the third word

(source text) d — » ? ~ recount - word four

(source text) a - > ? ~ recount - word five

(source text) b - > ? ~ recount - word six

(source text) with -> ? ~ recount is the seventh word

(source text) d — > ? ~ recount - word eight

Let us carry out the first recalculation approach, taking the first four words of the group being recalculated and the only word of the group being replaced as a conditional “incoming balance”, acting similarly to what was considered in example 1A case, where applicable, to obtain the first half of the recalculation result and the data moving to the next approach, as a conditional “outgoing balance”: d 3 ₍₄₎ // incoming balance a, b, c, d - > 0123(4 ) // recalculated subgroup d, a, b, с, d - 3 0123(4) // incoming balance and subgroup

£out,1 = L ^g / I = 4 / [2] = 2 // module of excluded data (since -£out,1 ^> 1, then there is data that goes to the next approach)

3 0123(4) div out,1 = 1 2031(4) // integer division

3 0123(4) mod £out,1 ⁼ 1(4) // remainder of division above

1 2031(4) —> 3042(5) // recalculation - first approach

3042(5) D, A, E, B // result of recalculation for approach

1(4) b // leaving balance

Let's carry out the second recalculation approach, taking the last four words of the recalculated group and the “outgoing remainder” of the first approach as a conditional “incoming remainder” for the second, and so on: b - > 1(4) // incoming remainder a, b, c, d - > 0123(4) // recalculated subgroup b, a, b, c, d - > 1 0123(4) // incoming remainder and subgroup out, 2 - N / \d*d\ = 4 / [2*2] = 1 // module of excluded data

(since _In х,2 = 1, then there is no data moving to the next approach) 1 0123(4) div £out,2 ⁼ 1 2031 ₍₄ ) // equal to the original one, since £out, 2 ⁼ 1 1 0123(4) mod £out,2 ⁼ null // does not carry data, because £out,r ⁼ 1 1 0123 (4) 2113(5) // recalculation - second approach 2113(5) -> C, B, B, D // result of recalculation for approach null -> missing // leaving balance

Thus, the implementation of replacement and recalculation in two approaches translates the original words into modified ones like this: d|, a~, b~, c~, d~, a~, b~, c~,d - ► e*, D, A , E, B, C, B, B, D. Given the encoding rule used, the following word entries of the event-compressed sequence are identical: e*,D,A,E,B,C,B,B,D = E,D,A,E,B,C,B,B,D = e*,d,a,e*,b,c,b,b,d .

As a result, the following words with the same (1 + 2x4 - 1 + 8 - 9) number will be located in the compacted sequence at the positions of the words of the replaced and recalculated vocabulary groups:

[sl. text] E = e* (<— d) “click” - one word

[sl. text] D = d (<— a) “echo” - first approach, first word

[sl. text] A = a (<— b) “echo” - first approach, second word

[sl. text] E = e* (<— s) “echo” - first approach, third word

[sl. text] B = b (<— d) “echo” - first approach, fourth word

[sl. text] C = d (<— a) “echo” - second approach, first word

[sl. text] B = a (<— b) “echo” - second approach, second word

[sl. text] B = e (<— s) “echo” - second approach, third word

[sl. text] D = b (<— d) “echo” - second first, word fourth

The compaction process in this example changes accordingly and now looks like this:

For convenience, we rewrite the above entry into a more concise form, hiding the intermediate echo (1/2), we get:

The theoretical encoding (decoding) delay of data on the transmitting (receiving) side, directly proportional to the size of the encoding (decoding) buffer, for this example will be five (and _E + 1) words, which is three words less than the similar delay characteristic of example 1A , equal to eight (n _D + 1) words.

Note - Conversion into several approaches is universal, flexible, and easy to apply. For example, for the case considered in example 1B, The following options for the compaction process are possible with conversion into one single, two and four identical approaches, respectively:

1/1 — > replacement {2} — * 1/16 — > recalculation {13} — > 1/1, E _in / E _out - 16 ;

The theoretical encoding delay for the listed processes will be 2 + 13 = 15, 2 + 7 = 9, 2+ 4 = 6 words, respectively.

For the case considered in example 1B, different options for the corresponding compaction process are also possible, for example:

1/1 -> replacement{2} -> 1/8 -> recalculation{10} -> 1/1, E _in / E _out ⁼ 16 ;

1/1 -> replacement {2} -> 1/8 -> recalculation {4x3} -> 1/1, E _BX / E _out , _r ⁼ 2.

The theoretical encoding delay for the listed processes will be 2 + 10 = 12, 2 + 4 = 6 words, respectively.

Let's act more exotically and set the compaction process in such a way that the first and second approaches to its recalculation are not the same:

1/1 — > replacement{2} — > 1/8 — * recalculation {4} — > 1/4 — > recalculation{7} - 1/1 .

The theoretical encoding delay for such a process would be 2 + 4 = 6 and 1 + 7 = 8 words for the first and second approaches, respectively.

Example 2B. Let's continue example 2A and note that among 5 ⁴ = 625 four-digit base numbers N** = |T**| = 5 there are exactly 512 such that contain the number 4 ₍₅₎ = e* = E in no more than one of their digits. The remaining 625 - 512 = 113 numbers contain such a digit in two, three, or all four of their digits.

The number of five-bit base N = |T|= 4 numbers that need to be uniquely reversibly recalculated in a single approach is also 2x4 ⁴ = 512, which allows each of them to match one unique of the 512 four-bit base 5 numbers containing the digit 4( 5) in no more than one of its digits, using nonlinear table function (English, look-up table), for example, the following form:

(001) 00000(4) - 0000(5) ; . . . ; (004) 00003 ₍₄₎ 0003 ₍₅₎ ;

(005) 00010(4) 0004(5) ; . . . ; (008) 00013(4) ^ 0012(5) ;

(009) 00020(4) 0013(5) ; . . . ; (012) 00023(4) - 0021(5) ;

(501) 13310 ₍ 4) 4310 ₍ 5); (504) 13313 ₍₄₎ <-> 4313 ₍₅₎ ;

(505) 13320(4) ^> 4320( ₅ ) ; . . . ; (508) 13323 ₍₄₎ 4323 ₍₅₎ ;

(509) 13330(4) 4330(5) ; ■ ■ • ; (512) 13333(4) 4333(5) ■

This nonlinear comparison reduces the probability (frequency of occurrence) of figurative words in the compacted sequence, which, in turn, makes it less statistically different from the compacted sequence.

§ 3. If it is necessary to set a direct order in the relative positions of the replaced and recalculated words in the sequence, then in the proposed invention - in the context of a single event -

■ select positions for recalculation,

■ ■ coming after the items selected for replacement.

The main technical result (see § 1 earlier) remains the same with this solution. An additional technical result with this solution is to arrange the replaced and recalculated words in such a way that when transmitting a compressed sequence word-by-word, the words of the recalculated group related to a specific event will follow strictly after the words of the replaced group related to the same event.

Let us give an example of the implementation of such an invention. Example 3. As shown above, the present invention can be implemented by choosing the positions of words relative to each other so that the positions of the words of the word group being replaced precede the positions of the words of the word group being recalculated.

Let's take example 2A as a basis and agree that the positions of all consecutive words in the sequence are specified monotonically and without gaps by an increasing integer m, -co < m < +oo, while the word in the smaller position is considered to precede the word in the larger position.

Then the case considered in Example 2A can be realized by choosing the positions of words relative to each other as follows:

(text] replacement, unit @ t 1 d -> E “click”, one unit.

(text] recalculation, unit 1 @ t + 1 ~ a - > D “echo”, approach 1, unit 1

(text] recalculation, unit 2 @ t + 2 ~ b -> A “echo”, approach 1, unit 2

(text] recalculation, unit 3 @ t + 3 ~ s - > E “echo”, approach 1, unit 3

(text] recalculation, unit 4 @ t + 4 ~ d — > In “echo”, approach 1, unit 4

(text] recalculation, unit 5 @ t + 5 ~ a - > C “echo”, approach 2, line 1

(text] recalculation, line 6 @ t + 6 ~ b - In “echo”, approach 2, line 2

(text] recalculation, line 7 @ t + 7 ~ s - > In “echo”, approach 2, line 3

(text] recalculation, line 8 @ t + 8 ~ d — -> D “echo”, approach 2, line 4

Summarizing the above, we will express the process of compaction, characteristic of the case considered, in the conditional form adopted below so as to visually link the meanings, quantities and positions of the words involved in the compacted sequence, as well as the echo that arises:

The above entry can be somewhat simplified if it is not necessary to disclose in detail some aspects of the process in question:

1/1 -» [T*] ¹ 1/4 [T**] ⁸ -► 1/1. In summary, the considered process of compaction, conventionally called “[echo] cancellation” (pronounced English, echo cancellation), is characterized by the fact that it first allows the appearance of an echo and then suppresses it.

§ 4. If it is necessary to set the reverse order in the relative positions of the replaced and recalculated words in the sequence, then in the proposed invention - in the context of a single event -

■ select positions for recalculation,

■ ■ coming before the positions selected for replacement.

The main technical result (see § 1 earlier) remains the same with this solution. An additional technical result with this solution is to arrange the replaced and recalculated words in such a way that when transmitting a condensed sequence word-by-word, the words of the recalculated group related to a specific event will follow strictly before the words of the replaced group related to the same event.

Let us give examples of the implementation of such an invention.

Example 4A. As shown above, the present invention can be implemented by choosing the positions of words relative to each other so that the positions of the words of the word group being replaced follow the positions of the words of the word group being counted.

Let’s take example 2A and the condition for m from example 3 as a basis and immediately present the result, the receipt of which will be revealed further:

(text] recalculation, next 1 @ "/ 8 ~ d -> D “echo", approach 2, unit 1

(text] recalculation, line 2 @ m - 7 ~ a -> A “echo”, approach 2, line 2

(text] recalculation, verse 3 @ t - 6 b - "E "echo", approach 2, verse 3 (text] recalculation, line 4 @ t - 5 ~ s —> In “echo”, approach 2, line 4

(text] recalculation, line 5 @ t - 4 ~ d — ► With “echo”, approach 1, line 1

(text] recalculation, line 6 @ t - 3 ~ a -> In “echo”, approach 1, line 2

(text] recalculation, line 7 @ t - 2 ~ b — > In “echo”, approach 1, line 3

(text] recalculation, line 8 @ t - \ ~ s - > D “echo”, approach 1, line 4

(text] replacement, units @ t

E “click”, one syllable.

This implementation is a “mirror image” of what is given in example 3, due to which both one and the other inherently repeat each other up to a certain point: d, a, b, c, d, a, b, c, d — > 3012 3 0123 ₍₄ );

301230123 ₍ 4) 12031 ₍ 4) , 10123(4) ;

12031(4) - 3042( ₅ ) -> D, A, E, B;

10123(4) 2113( ₅ ) C, B, B, D.

The difference appears at the last step of obtaining a condensed sequence, which can be seen when comparing both cases: d|, a~, b~, c~, d~, a~, b~, c~, d - > e*, D, A, E, B, C, B, B, D;

The expected difference is also observed when comparing records of the corresponding compaction processes:

With a brief note, the “mirror” nature of these compared processes appears even more clearly:

1/1 — replacement {1} — > 1/4 — > recalculation {4x2} — > 1/1 ;

1/1 — > recalculation {4x2} 4/1 replacement {1 } — > 1/1 .

Let's pay attention to the echo, which in the entries above takes on new, previously unappearing values. Let us denote: BX,I+ = in,1 from example 2A = 4 = 4/1 > 1, out,1+ = out,1 from example 2A = 2 = 2/1 >1; BX,2+ = ^in,2 from example 2 A = 2 = 2/1 > 1 , OUT,2+ = OUT,2 FROM example 2 A = 1 = 1/1 = 1 .

By analogy with the above, we define the corresponding modules characteristic of the recalculation approaches in the case under consideration:

Rule (5) and the conversion into two approaches introduced in Example 2A remain valid for the case under consideration, since:

Thus, echo, introduced in example 1A as the inverse value of the “migrating” (excluded and added) data module, now “mirror” rightly takes values greater than one.

The compaction process typical for the case in this example, with different degrees of detail, looks like this:

In summary, the considered compaction process, conventionally called “[echo] anticipation” (echo obviation), is characterized by the fact that it first prevents the appearance of an echo and then implements it.

Example 4B. Let's continue example 4A, combining it with example 3 as follows, gluing them along the boundary of the same echo:

Considering that the original words of the compacted sequence occupying positions different from the selected positions of the words being replaced and recalculated vocabulary groups are not subject to such operations and are therefore present unchanged in the condensed sequence, that is, they do not generate an echo, new or additional, in the middle of the above process - between the positions of words relating to the initial (left) and final (right) events , — you can place any integer number L > 0 of such initial words:

In other words, this makes it possible to transmit data in frames of arbitrary size, starting from the minimum, which in this case is 9 + 9 = 18 words; The maximum size of such frames is theoretically unlimited.

In addition, this compaction process allows you to transmit data frames without losing bandwidth for framing purposes, understanding here bandwidth as the number of words of the compressed sequence transmitted per unit of time.

Thus, the considered compaction process uses compaction events to transmit data in frames at the rate and volume of occurrence of the words of the sequence being compressed.

Example 4B. Let's continue example 4B, changing the compaction process discussed in it as follows:

The resulting process is correct, since, as previously explained in Example 4B, the original words occupying positions different from the selected positions of the words of the replaced and recalculated vocabulary groups do not generate a new or additional echo, so they can be placed in positions such as shown above.

However, the result of such a process will be slightly different from the result corresponding to the process from example 4B. Let's pretend that Between the “echo” and “click” phases of the final event, four new words are pasted (L = 4), we get the following picture:

(text] p-t, words.

~ d — > D “echo”, approach 2, next. 1 (text] p-t, sl.

“echo”, approach 2, next. 2

(text] p-t, syllable 3 @ t - L - 6 ~ b -> E “echo”, approach 2, lyric 3

(text] p-t, words.

“echo”, approach 2, next. 4

(text] p-t, words.

“echo”, approach 1, next. 1

(text] n-t, next 6 @ m -L - 3 ~ a - > In “echo”, approach 1, next 2 (text] n-t, next 1 @ m -L - 2 ~ b - > B “echo”, approach 1, line 3 (text] p-t, line 8 @ t - L - 1 ~ s — > D “echo”, approach 1, line 4

(text] transfer @ m - L + 0 d - > a taken in the next position below

(text] transfer @ t - L + 1 a -> b taken in the next position below

(text] transfer @ tp - L + 2 b - > with taken next position below

(text] transfer @ t - L + 3 s -> d taken in next position below

(text] z-a, singular @ t

"click", one word

Comparing the positions of the original words for which replacement and recalculation are carried out, and words without such manipulations (indicated by “=”), for the process from example 4B and this example, we obtain: d=, a=, b=, c=, d~, a~, b~, c~, d~, a~, b~, c~, dj. ;

From the differences above follows the expected difference in the compacted sequence obtained in the compared processes, while other characteristics of the processes are completely the same:

Note - For cases where the condition for m from example 3 applies, such as the cases in example 3 itself, examples 4A, 4B, 4C, the following interpretation (purpose, meaning) of the echo is valid. An echo greater than one indicates that the data of a word located in the position of the condensed (original) sequence affected by such an echo is subject to transfer to one or more words of the condensed (resulting) sequence located in a position smaller than the original position, in proportion to the echo .

An echo less than one indicates that the data of a word located in the position of the condensed (original) sequence affected by such an echo is subject to transfer to one or more words of the condensed (resulting) sequence located in a position greater than the original position, in proportion to the echo .

An echo equal to one indicates that there is no echo effect, so no data transfer is required, hence the positions of the word not affected by the echo are the same in both sequences.

In other words, the echo is equivalent to the current state (status, mode) of the compaction process and is numerically proportional to the relative position of the data in the words of the sequences, generally characterized as leading, lagging, or coincidence, respectively.

§ 5. If it is necessary to respond to only one of the emerging series of compacting events that differ in some way from each other, then in the proposed invention - in the context of a single event -

■ an event for which

■ ■ at least one word position

■ ■ ■ from any vocabulary group corresponding to this event,

■ ■ matches the position of the word

■ ■ ■ from a vocabulary group corresponding to another event,

■ ignored. The main technical result (see § 1 earlier) remains the same with this solution. An additional technical result with this solution is the ability to isolate one single event from the emerging series of compacting events.

Let us give an example of the implementation of such an invention.

Example 5 As discussed above, the present invention can be implemented by discarding all events from the sequence except one, which condenses the original sequence.

Let's take example 3 as a basis and note that the compacting event may differ in the position of which word in the compacted (initial) sequence it falls on. The sequence events cannot differ in anything else here, since the corresponding compaction process can only indicate the position(s) of the word at which the compacting event that caused it occurred, and do this only for one event.

Thus, any other events of the sequence that fall at the positions of the words of the vocabulary groups being replaced and recalculated within the framework of such a process, in addition to the event (as a cause) that caused this process (as a consequence), are discarded.

In total, in the limit - with a continuous sequence of events - the compaction process will take into account no more than one event for every nine words of the original sequence and discard all others.

§ 6. If it is necessary to respond to more than one of the emerging series of compacting events that differ in some way from each other, then in the proposed invention - in the context of a single event -

■ an event for which

■ ■ at least one word position ■ ■ ■ from any vocabulary group corresponding to this event,

■ ■ matches the position of the word

■ ■ ■ from a vocabulary group corresponding to another event,

■ taken into account.

The main technical result (see § 1 earlier) remains the same with this solution. An additional technical result with this solution is the ability to transmit several events from the emerging series of compacting events.

Let us give examples of the implementation of such an invention.

Example 6A. As discussed above, the present invention can be implemented without discarding other events from the sequence other than the one with which the original sequence is compressed.

Let's take example 2B as a basis and further consider each of the possible [№=*]" = 5 ⁴ = 625 four-digit numbers in base L'** as a unique instance of echo (sentence English, [echo] sample).

Note that

= 2»4 ⁴ = 512 of the 625 unique echo instances are matched to the same number of all possible results of the recalculation approach, while the remaining 625 - 512 = 113 instances are free from such matching and therefore can be used in some other way.

Based on this, we divide all unique echo instances into those that are compared to the results of the recalculation approach, conditionally calling such an echo “natural”, and the remaining or free ones, conditionally calling such an echo “forced” (sentence English, native and forced, respectively).

Therefore, when a compacting event hits the word position of the recalculated group of another event, the natural echo (an instance of a natural echo) is “replaced by force” with a forced one (an instance forced echo), which increases the current echo module, similar to replacing the words of the replaced vocabulary group, and requires a proportionally larger number of approaches to compensate for it (the echo).

We will use the maximum possible of the available instances of forced echo so that the increase in the echo module when replacing with “force” is minimal, for which we will find their number as

= d*.. *d < 113, we get that

= [ ] ⁶ = 2 ⁶ = 64. It follows that replacing the echo with a force increases (remember, multiplicatively) the modulus of the echo by the amount:

1024 = d ^w .

One recalculation approach reduces (remember, multiplicatively) the modulus of the echo by the amount b'in,/ / E'out,/ ⁼ d, therefore, each replacement of the echo by force leads to a lengthening (here additive) of the recalculation by an additional ten

approaches beyond those already available.

Suppose that the first of a series of events causes the following compaction process, leading to the replacement (indicated by "|") and recalculation (indicated by "~") of the words of the compacted sequence:

If the second of the series of events falls on a position related to the word of the first recalculation approach, the compaction process above is proportionally lengthened, changing as follows:

If the second of the series of events falls on a position related to the word of the second recalculation approach, the compaction process above changes in a similar way, but without additional lengthening:

If the second and third of the series of events fall into positions related to the words of the first and second recalculation approaches, the compaction process given above is extended even further:

Thus, the considered compaction process can convey a series of compaction events with the assumption that the events occurring at word positions in one recalculation approach are not different.

Example 6B. Let us assume that it is required to organize a channel for transmitting digital data with frames of two types, preemptable and urgent (English, preemtable and express, respectively), supporting the displacement (English, preemption) of frames of the first type (P) by frames of the second type (X).

Let us assume that the channel in question is capable of transmitting one of the twenty-one (21) distinguishable words at each position of the compressed vocabulary sequence, of which sixteen and one (16 + 1 = 17) are also intended to represent sixteen data values and data errors in the compressed sequence:

T = {a; b; With; d; e; f; g; h; i; j; k; 1; m; n; o; p; q } , #= |T| = 17 ;

T* = { r*, s*, t*, u* } , TV* = |T*| = 4 ;

We will assume that the compaction process - with respect to each position (m) of a word in the sequence - is implemented either by an interruptible process, which generates a preemptible (interruptible) data frame, or a nested process, which generates an urgent (interruptible) frame, using four different (distinguishable) data frames. compacting events, replacement for which is carried out according to the following rule: at the event “Start of frame P”: z -> r*, where z belongs to T; at the event “End of frame P”: z -> s*, where z belongs to T; at the event “Start of frame X”: z -> t*, where z belongs to T; at the event “End of frame X”: z -> u*, where z belongs to T.

Thus, the smallest integer covering the modulus of the echo arising from the substitution above is E _mH1 = N= 17: 1/1 substitution{1} 1/17 ... . The echo modulus is a prime number, unattractive compared to other prime numbers smaller than it, in comparison according to rules (4) or (5) with E _in , _; / £out, _; = E = d for these numbers: d ₂ = 2 : [TV**] ³ <d ₂ -[TV] ³ , [#**] ⁴ > d ₂ -[TV] ⁴ => n ₂ = 4 d = 3 : [TV**] ⁵ <<7 ₃ *[TV] ⁵ , [TV**] ⁶ > d ₂ “[TV] ⁶ => n = 6 d ₅ = 5 : [TV**] ⁷ <d ₅ *[N] ⁷ , [TV**] ⁸ > d ₅ “[TV] ⁸ => n ₅ = 8

d»> 17 : >=>and»>« ₁₇

In addition, such an integer is inferior in comparison to ordinary integers covering the echo module, but allowing recalculation by several approaches, which, in turn, reduces the encoding-decoding delay when transmitting data:

17= 17= 17 ¹ => /li ₇ = l

17 <18= 2 ¹ • Z ³ =>& ₁₈ = 4, E and _z = 22, max {i _; } = 6

17 <19= 19 ¹ => £ ₁₉ = 1

17 < 20 = 2 ² • 5 ¹ => k _2(i = 3, £i _r = 12, max{i _r } = 8

17 <21= З ¹ • 7 ¹ => к ₂ \ = 2 , Ей, = 16, max{n _z } = 10

17 <22= 2 ¹ * !! ¹ => to ₂₂ = 2, Ei _; = 16, max{« _; } = 12

17 <23 = 23 ¹ => £ ₃₁ = 1

17 <28= 2 ² • 7 ¹ => k ₂ % = 3, E and _z = 18, max{i _; } = 10

17 <29= 29 ¹ => ^9= 1

17 <30= 2 ¹ • З ¹ • 5 ¹ => Azo = 3 , Sw _; =18, max {u,} = 8 17 < 31 = 31 ¹ =>&3i = 1

Based on the above, we will obviously make the optimal choice of parameters for the compaction process under consideration: d = d ₂ ⁼ 2 j p = p ₂ = 4 ; k = k ₂₂ = 5 , i = !...£ = 1, 2, 3, 4, 5 .

Based on this, the implementation of echo cancellation that occurs when the original sequence is compressed by the “Start of frame P” event is expressed in the case under consideration by writing: 5} -> 1/1, or

Similarly, the implementation of echo anticipation that occurs when the original sequence is compressed by the “End of frame P” event is expressed in the case under consideration by writing:

Consequently, the implementation of echo cancellation that occurs when the original sequence is compressed by the “Start of Frame X” event, which falls at a position different from the positions of the words of the recalculated vocabulary group in the interrupted process, is expressed as follows:

When the “Start of Frame X” event hits the position of any of the words of the vocabulary group being recalculated in the interrupted process, the echo is replaced by force, for the implementation of which instances of forced echo are used, selected from among those not matched to the result of the recalculation approach, in the following quantity:

19652 < 27439 = [A**]" - "[7V]" .

This allows you to specify, with accuracy down to the word position, one of n = 4 word positions related to the approach being replaced by such a “force”, introducing additional echo with familiar module:

Hence, the implementation of echo cancellation when replacing by force is expressed by the entry below, where _P is the current echo module in the interrupted process:

The number R in the entry above determines the number of approaches that must be made to bring the echo to the same value as the echo cancellation ends when the event does not fall into the positions of the recalculated group in the interrupted process:

1/ _P = 1/32 1/16 1/8 1/4 1/2 1/1 2/1 4/1 8/1 16/1 ;

R = 5+10 5+9 5+8 5+7 5+6 5+5 5+4 5+3 5+2 5+1 .

Accordingly, the implementation of echo anticipation that occurs when the original sequence is compressed by the “End of frame X” event is expressed in the case under consideration by writing:

Thus, the interruptible compaction process that forms the preemptible (interruptible, P) data frame must follow the following structural rule:

Start of frame P:

End of frame P:

In turn, the nested compaction process that forms an urgent (interrupting, X) data frame must be carried out according to a more complex - in relation to the part that is responsible for the beginning of the frame - structural rule given below:

The beginning of frame X is in positions located between the echo, 1/2G _R ,

TO

It is worth noting, however, that the sequence of compaction events should be correctly limited so that the considered process can process them without losing their differences.

Example 6B. Let's continue Example 6B and add the compaction process under consideration so that the data transmission channel that implements it supports, in addition to the existing one, duplication of data in preempted and urgent frames.

Let's increase the number of instances of forced echo in order to indicate the situation when the events “Start of Frame X” and “Start of Frame P” occur at the same word position in the sequence:

- 17 ³ = 24565 < 27439 = [N* *] ^and - d>[N] ⁿ .

Let's supplement the process with a rule to implement the simultaneous start of urgent and preempted frames with identical content:

Let's supplement the process with a rule to implement simultaneous, advanced and delayed completion of duplication of the contents of an urgent frame in relation to the contents of the preempted frame:

As a result, thanks to more diverse rules, the augmented process is able to consistently process a more diverse series of compaction events without losing their differences.

Note - By analogy with the known and used communication mechanisms with packet switching and line switching (English, packet-switched and circuit-switched, respectively), that discussed in examples 6A, 6B, 6B, it is appropriate to call it a communication mechanism with event switching (event-switched), since in such a mechanism it is the compacting events that influence the decision on the corresponding switching, determining to whom and which ones are transmitted in parallel with such events data are intended.

The examples discussed above show that event switching generalizes both types of switching and therefore replaces them both.

§ 7. If it is necessary to carry out multiple multiplexing of a digital data transmission channel, then in the proposed invention - of course, with the available variety of compaction events -

■ events are divided into categories,

■ ■ at the same time,

■ each category is compared with its own

■ ■ many positions of words in a sequence,

■ ■ ■ which (set)

■ ■ ■ ■ element-wise does not intersect with any of the sets,

■ ■ ■ ■ ■ mapped to categories other than this one,

■ ■ ■ and within which

■ ■ ■ ■ it is allowed to select word positions for vocabulary groups,

■ ■ ■ ■ ■ corresponding to events in this category.

The main technical result (see § 1 earlier) remains the same with this solution. An additional technical result with this solution is to obtain several sealing channels that are available simultaneously, arbitrarily, independently of each other.

Let us give examples of the implementation of such an invention. Example 7A. As discussed above, the present invention can be implemented by associating each category of compacting events with its own set of word positions in the sequence, element-wise not intersecting with the sets of positions associated with other categories of compacting events.

Let's take example 3 as a basis and assume that there is not one, but five distinguishable (different) compacting events, each representing its own separate category, that is, event “1” belongs to category “1”, event “2” - to category “2” ", and so on.

In order to keep the process of compaction, considered in example 3, unchanged, we compare each category with its own set of word positions in the sequence so that the positions of words belonging to different categories will belong to only and only one of the sets: for category “1”: m\ = 5t + 0 ; for category “2”: t ₂ = 5t + 1; for category “3”: t ₂ = 5t + 2; for category “4”: t ₄ = 5t + 3; for category “5”: t ₅ = 5t + 4.

The accepted comparison ensures the parallel operation of five separate (isolated) compaction processes - one for each category of compaction events - available simultaneously, arbitrarily, independently of each other, which expresses the possibility of multiple compaction of the original sequence.

Example 7B. Let us again take example 3 as a basis and assume that the minimum unit of data transfer is one byte (English, byte), consisting of eight bits, that is, its data module is equal to 2 ⁸ = 256, while the data bytes themselves must be transmitted in parcels (arrays ) of arbitrary length, that is, in frames. Hence, to transmit one such byte, log _tA r|256 = 4 words of the original sequence with modulus N= |T| = 4 each.

Let's introduce two categories, each including one event, and compare each with its own set of word positions in the sequence: for the category (event) “Beginning of frame”: m = 4m - 1; for category (event) “End of frame”: w _K = 4t + 2.

Thus, it is possible to organize the transmission of data bytes in frames using the compaction process discussed in Example 3.

Example 7B. Let's take examples 3, 4B, 4B as a basis, and assume that it is required to transmit synchronization pulses (English, pulse per second, PPS) and nibble frames (English, nibble) of data - in parallel and independently of each other - over one transmission channel.

So, a nibble consists of four bits, that is, its data module is 2 ⁴ = 16, therefore, to transmit one such nibble, only log v]16 = 2 words of the original sequence are needed. On the contrary, the synchronization pulse does not transfer data and is a compacting event in its pure form, isolated and independent.

Let us introduce two categories of compaction events, called “S” and “P”, for synchronization and data transmission events, respectively, and three different (distinguishable) compaction events with the purpose below: event “1C” = “synchronization pulse”; event "2P" = "beginning of frame"; event "ZP" = "end of frame" .

Let's divide the entered events into the entered categories, according to the purpose of the first, as follows: category “C” = { event “1C” } ; category "P" = { event "2P" , event "ZP" } .

Now let’s compare a set of positions to each of these categories words in condensed and, as a result, condensed sequences: for category “C”: t _c = 2/w; for category “P”: t _p = 2t + 1.

From here, it is possible to organize the transmission of synchronization pulses using the compaction process discussed in example 3, and the transmission of nibble data in frames using the compaction process discussed in examples 4B or 4B, in parallel and independently of each other.

§ 8. If it is necessary to use dictionary sequences built on the basis of a regular (fixed) structure, then in the proposed invention - having sufficient redundancy -

■ in all positions of words in the sequence

■ ■ use words from a single set of corresponding words.

The main technical result (see § 1 earlier) remains the same with this solution. An additional technical result with this solution is the ability to compress digital channels with a constant bit rate (CBR).

Let us give an example of the implementation of such an invention.

Example 8. As disclosed above, the proposed invention can be implemented using a single, unchanging set of words, which is expressed by taking the condition for m from example 3, in the general case like this for all m:

T _from = { . . . } = T = const => N _m = |T| = N = const ; m* = { . . . } = T* = const => N _m * = |T* | = N* = const ;

T„,** = { . . . } = T** = const => N _m ** = |T**| = N** = const .

Constraint (3) determines the condition for optimal redundancy, fair (condition) for the case under consideration.

Constraint (2) determines the condition for sufficient redundancy, which is valid (condition) for the case under consideration.

§ 9. If it is necessary to use dictionary sequences built on the basis of an irregular (variable) structure, then in the proposed invention - having sufficient redundancy -

■ in different positions of words in the sequence

■ ■ use words from different sets of corresponding words.

The main technical result (see § 1 earlier) remains the same with this solution. An additional technical result with this solution is the ability to compress digital channels with variable bit rate (VBR).

Let us give an example of the implementation of such an invention.

Example 9 As disclosed above, the present invention can be implemented using several sets of words, which is expressed by taking the condition on m from Example 3, in general as the occurrence of any one of the following conditions - on all m:

const => N _m = |T _m | const ;

T _m * = { . . . } const => N _m * = |T _m *| const ;

T _m ** = { . . . } const => A^** = |T,„**| const.

Constraint (2) determines the condition for sufficient redundancy, which remains valid (condition) for the case under consideration, when from all possible (w) positions of words in the sequence, only those (/) are selected for which the above constraint is met under the assumption that A = N _} and N** = A,**. Hence, rules (4) and (5) are applicable in the case under consideration in the following form, modified according to the purpose - for finite (finite size) samples in /, respectively:

G H > E x G ], (6)

П[^/*] > £вх >< П[Л)П £вь1х . (7)

§ 10. If it is necessary to increase the degree of stability (English, robustness) of recognition of compaction events during reception, then in the proposed invention - the focus of a single replacement -

■ when replacing at least one selected word

■ ■ use out-of-dictionary signaling.

The main technical result (see § 1 earlier) remains the same with this solution. An additional technical result with this solution is the ability to achieve the required degree of stability when recognizing compaction events on the receiving side.

Note - Out-of-band signaling is the use, as a rule, of out-of-band signals, or signal symbols, the transmission of which requires a change in the linear code, the symbols of which are used to transmit ordinary ones on the line (T, T*, T** ) words of the sequence, to another - often simpler and therefore more noise-resistant - linear code, the symbols of which ordinary words are never transmitted, for example, changing the type MLT3 / PAM2, RAM5 / RAMZ, RAMZ / RAM2, DME / PAM2, or, what is less preferably, the use of other, non-multiple or non-stream (aHi Ji.out-of-sct, out-of-stream) signal symbols, the transmission of which occurs without changing the linear code, the symbols of which are used to transmit ordinary words.

Thus, in the context of the disclosure of the present invention, the use of extra-dictionary signaling implies the presence of some additional set of such special signal symbols, or signals, available to the compaction process with the caveat that any of their occurrence in the sequence is an exceptional case, the number of which should be kept to a minimum.

Let us give examples of the implementation of such an invention.

Example 10A. As disclosed above, the present invention can be implemented using special signal symbols, or signals, providing a higher degree of recognition stability by the receiving side compared to ordinary words.

Let's take example 4B as a basis and once again present the compaction process characteristic of this example, which allows data to be transmitted over the channel in frames of (almost) arbitrary size:

Let us imagine this process by carrying out replacement (indicated by “|”) and recalculation (indicated by “~”) according to the structural rule responsible for the transmission of the corresponding events, similar to Example 6B:

Start of frame: 1/1

... ;

End of frame: ... -> 4/1 -> [T] ^£ ~° -> 4/1 -> CT*] ¹ -> 1/1.

Suppose that it is necessary (for some reason) to increase the degree of stability of recognition of the event corresponding to the beginning of the frame of transmitted data, for which extra-dictionary signaling signals (□) are available, with which the corresponding rule can be strengthened:

Start of frame:

Let us assume that it is necessary (again, for some reason) to provide the ability to interrupt (“cancel”, “reset”) the transmission of the current data frame in an arbitrary place, and with an increased degree of stability in recognizing such an event, for which we introduce into the process under consideration compaction is a new rule, clearly different from the existing ones and initially reinforced by signal symbols:

Frame break: (any echo) - "ppp] ³

1/1.

Thus, the use of extra-dictionary signaling in the implementation of the present invention is subject to the same general principles that are discussed in detail in the examples earlier.

Note - Recall that / in the entry

] ^z —» ... means how many words of the sequence the contents in parentheses occupy.

Example 10B. Let's take example 6B as a basis and again assume, by analogy with example 10A, that the compaction process has access to extra-dictionary signaling signals (□), with the help of which it is necessary to increase the degree of stability in recognizing the boundaries of data frames.

First of all, let us strengthen the replacements carried out for existing compacting events with signal symbols:

“Beginning of frame P” xy -> □ g*, where x, y belong to T, “End of frame P” xy z -> s*on where x, y, z belong to T; “Beginning of frame X” xy - > □ t*, where , y belong to T, “End of frame X” xy z - > u* □ □, where x, y, z belong to T.

Next, we will strengthen the replacement of the echo by force, carried out when the event “Start of frame X” hits the position of any of the words of the replaced or recalculated vocabulary group in the interrupted process, knowing that for the event “Start of frame P” the first one has increased by one word:

Thus, the interruptible compaction process that forms the preemptible (interruptible, P) data frame must follow the following structural rule, now reinforced:

Start of frame P:

End of frame P:

Hence, the nested compaction process that forms the urgent (interrupting, X) data frame must follow the following structural rule, also now reinforced:

The beginning of frame X is in positions located between the echo, 1/£ _P ,

TO

The dependence of the number of recalculation approaches of the interrupting process, R, on the current echo of the interrupted process, l/E'p, when the resulting echo is suppressed to a value equal to 2 ¹⁶ /1, is still linear in the entry above:

If it is necessary to interrupt the transmission of a data frame in an arbitrary place, with an increased degree of stability in recognizing the location of such a break, then a new structural rule, initially reinforced with signal symbols, can be introduced into the compaction process under consideration:

Frame break P: (any echo) -> [,[,□□□□] ⁴ ->1/1;

Frame break X : (any echo) ->

— * 2 ¹⁶ /1 .

Example 10B. Let's take example 7A as a basis, where the process of condensing five different events (one in each of the five categories) is correlated with five different sets of word positions in the sequence: event “1”: set of positions m = 5m + 0; event “2”: set of positions t ₂ = 5t + 1; event “3”: set of positions t ₂ = 5t + 2; event “4”: set of positions t ₄ = 5t + 3; event “5”: set of positions t ₅ = 5t + 4.

Let us assume, by analogy with example 10A, that the process has access to extra-dictionary signaling signals (□), using which it is necessary to increase the degree of stability of recognition of the event “1” - by transmitting five such signals in five adjacent positions of the sequence.

Consequently, the click caused by the event "1" is carried out by replacing words in all five sets of positions - one for each set - according to a rule reinforced by signal symbols:

@ t : v wxy z - > □ □ □ □ □ , where: v belongs to T, position v belongs to pu; w belongs to T, T*, or T**, position w belongs to t ₂ ; x belongs to T, T*, or T**, position x belongs to m ₂ ; y belongs to T, T*, or T**, position y belongs to pc; z belongs to T, T*, or T**, position z belongs to pc.

This cross-click generates an echo, new or additional, the subsequent cancellation of which is carried out in each of the five sets of positions separately (in isolation) and independently of each other - within the framework of the corresponding compaction process.

For event “1” the compaction process is the simplest and is expectedly carried out according to the following (single) rule:

For event “2” the compaction process is more complex and in general is carried out according to the following (initial) rule:

If (and when) the cross-click falls on the word position of the recalculated vocabulary group related to the event “2” itself, the echo is replaced by force according to the following rule - depending on of the position of which word of the group such a click occurs and therefore must place an extra-dictionary signaling symbol:

@ t ₂ : ME [~T**] ⁴ 2/E

or ME - CT** T** T** p] ⁴ 1/42? .

(Here: N _a = d*.. *d = [<2] ⁶ = 2 ⁶ = 64 < 125 = [TV**] ³ , therefore the additional echo module E _i = [7V] ⁴ / [< ] ⁶ = 256 / 64 = 4, hence E _a * E = E.)

This means that such a replacement of the echo by force leads to a lengthening of the process by two additional recalculation approaches (log^ = log ₂ 4 = 2) and an approach for which the recalculation result is replaced by the force itself.

If (and when) the cross-click falls on a word position outside (not from) the replaced and recalculated vocabulary groups related to event “2”, the compaction process is carried out according to the rule:

As you can see, in this case the event "2" itself is missing, so here the compaction process (the compaction process for event "2") processes the event, as opposed to the "valid" one - as in the cases above and below, when the event "2" is present, - conventionally called “false” (prev. English, real and false, respectively). The specificity of a false event is that it participates in the compaction process on the transmitting side, but cannot be restored on the receiving side. The position of the false event in the case under consideration coincides with the position of the word at which the cross-click occurs, therefore the further (after replacement) course of the compaction process is similar to the initial rule.

If the cross-click occurs on the position of a single word of the replaced word group associated with event "2", such word will be "forcibly replaced" by the non-word signaling symbol:

Consequently, by analogy with replacing an echo with a force, a click is replaced with a force, that is, the click of event “2” is replaced by a click (or, figuratively speaking, the “force” of the click) of event “1”.

If and when the cross-click occurs on the word position of the exact vocabulary group being replaced, related to the event "2", or the position pre-, pre-pre-, pre-pre-, or pre-pre-pre-pre-pre-, or pre-pre-pre-pre-pre-, the position of the single word of this word group, which means that there is a valid event “2”, subsequent to the cross-click, falls on the position of one of the four (n = 4) words of the first recalculation approach in the recalculated group related to the corresponding false event “2”, the echo is replaced by force according to the rule:

**] ⁴ - [ ] ⁴ , additional echo module E _d = [TV] ⁴ / [d] ⁴ = 256 / 16 = 16, hence 4 x 16 = 64.)

This means that such a replacement of the echo by force leads to a lengthening of the process by four additional recalculation approaches (logjE _;( = log ₂ 16 = 4) and an approach for which the recalculation result is replaced by the force itself.

If and when the cross-click occurs at a word position in the sequence such that the subsequent actual event occurs at the position of one of the four (n = 4) words of the second approach, a similar replacement of the echo by force occurs according to the rule:

(Here already: _i>2 = n^d'...*d) = *[d\ ⁴ = 446 = 64 < 113 = [TV**] ⁴ - [TV] ⁴ , additional echo module E _i = [ TV] ⁴ / [d] ⁴ = 256 / 16 = 16, hence 2 x 16 = 32.)

This means that such a replacement of the echo by force leads to a lengthening of the process, also by four additional recalculation approaches (logj/T = log ₂ 16 = 4) and an approach for which the recalculation result is replaced by the force itself.

For events “3”, “4”, “5” the compaction process is similar to the compaction process considered for event “2”. Summarizing the above, let us compare separate (isolated) compaction processes in the context of the rules imposed on them: event “1” @ mi ■ the only rule, typical and linear; event "2" @ t ₂ : a set of several rules, including complex ones; event "3" @ t ₃ : set of rules as for event "2" @ t ₂ ; event "4" @ t ₄ : set of rules as for event "2" @ t ₂ ; event "5" @ t ₅ : set of rules as for event "2" @ t ₂ .

Note - It is worth noting that the replacement of a click by force, introduced in example 10B, and the replacement of an echo by force, introduced in example 6 A, represent - in essence - the same basic operation (replacement) of the same name, known from example 1 A, and have only a syntactically varying designation (replacement by force), which aims to associatively indicate the conditions of the replacement being carried out, significant in its context.

The basic operation, syntactically designated precisely as a replacement, is determined by the event that caused the process in question of compacting the data of the original sequence with data about this event.

In turn, the basic operation, syntactically designated as force substitution, is caused by an event different from the event that caused the compaction process in question. In other words, syntactically, such a designation indicates that the compaction process caused by one event changes due to the occurrence of another event.

Replacement increases the echo module, recalculation reduces the echo module, while both basic operations represent a certain operation-corresponding transformation of a certain numeric code corresponding to the data, excluding (conversion) data loss.

Of course, the same replacement can reset the echo module, as shown in examples 10A and 10B, but it is worth understanding that in such cases data loss is inevitable, since after such a replacement the echo is simply neglected. Thus, to implement the proposed invention, only two basic operations are sufficient, the designation of one of which varies syntactically depending on the context of its disclosure.

Implementation of the invention. The ways of implementing the proposed invention are disclosed in detail and described earlier - using examples 1 A, 1B, IB, 2A, 2B, 3, 4A, 4B, 4B, 5, 6A, 6B, 6B, 7A, 7B, 7B, 8, 9, 10A , 10B, 10B.

In all the examples listed above, sets are denoted in bold capital letters, elements of such sets are denoted in roman capital letters, and scalar quantities are denoted in italics.

As can be seen from the examples, the feasibility of the proposed invention is based on the arithmetic redundancy of the accompanying coding, which is expressed by limitation (2), and is therefore achievable in practice.

The author considers the implementation of the proposed invention in the form of a digital electrical circuit unit to be a practical target.

Industrial applicability. Based on the foregoing, the author believes that the proposed invention is industrially applicable in various telecommunications applications where uncompressed digital data transmission channels are used.

Taking into account the textual nature of the approach underlying the model for describing the proposed method of compaction, the author also believes that this method can be called “linguistic multiplexing”.

Claims

— 47 —

FORMULA OF THE INVENTION A method for compressing a digital data transmission channel, called “linguistic compaction,” characterized by the fact that the compressed vocabulary sequence, composed exclusively of words that are clear in meaning, whose location in it is uniquely determined by their positions, is modified in such a way that (firstly, so-called “click” phase), one vocabulary group, called replaceable, associated with the occurrence of a compacting event, composed of a selected number of initial words located in selected positions relative to the event, is replaced by a vocabulary group corresponding to the combination of initial words and the event itself, located on in the same positions the same number of words, of which at least one is figurative in meaning, (secondly, the so-called “echo” phase) another vocabulary group, called recalculated, associated with compensation for the compacting event, composed of a selected number of initial words, - 48 - located in positions selected relative to the event, different from the positions of the words of the group being replaced, are recalculated according to the combination of the original words located in the selected positions that belong to the recalculated group, the original words located in the selected positions that belong to the group being replaced, the event itself , stipulating that after recalculation it is permissible to place both clear and figurative words in the positions of the words of the group being recounted. The method according to claim 1, characterized in that the recalculation is carried out in several approaches. The method according to claim 1 or claim 2, characterized in that - in the context of a single event - the positions following the positions selected for replacement are selected for recalculation. The method according to claim 1 or claim 2, characterized in that - in the context of a single event - positions that come before the positions selected for replacement are selected for recalculation. - 49 - The method according to claim 1, claim 2, claim 3, or claim 4, characterized in that - in the context of a single event - an event for which at least one word position from any vocabulary group corresponding to this event , coincides with the position of a word from a vocabulary group corresponding to another event, is ignored. The method according to claim 1, claim 2, claim 3, or claim 4, characterized in that - in the context of a single event - an event for which at least one word position from any vocabulary group corresponding to this event coincides with the position of a word from a vocabulary group corresponding to another event is taken into account. The method according to clause 1, clause 2, clause 3, clause 4, clause 5, or clause 6, characterized in that - given the variety of events available - the events are divided into categories, and each of the categories is compared with its own set of word positions in the sequence, which (the set) does not intersect element-wise with any of the sets associated with categories other than this one, and within which - 50 - it is allowed to select word positions for vocabulary groups corresponding to events in this category. 8. The method according to clause 1, clause 2, clause 3, clause 4, clause 5, clause 6, or clause 7, characterized in that - having sufficient redundancy - words from a single set of corresponding words. 9. The method according to clause 1, clause 2, clause 3, clause 4, clause 5, clause 6, or clause 7, characterized in that - having sufficient redundancy - words from different sets of corresponding words. 10. The method according to clause 1, clause 2, clause 3, clause 4, clause 5, clause 6, clause 7, clause 8, or clause 9, characterized in that it is in focus on a single substitutions—when at least one selected word is replaced, out-of-dictionary signaling is used.