EP1164575A2 - Generation of vocal announcements - Google Patents

Abstract

The invention relates to a method for outputting a voice announcement, the content of which depends on the data content of at least one data record, in which an announcement type is selected on the basis of the data content by means of a first program segment, in which a second program segment is provided for program-controlled output, which is used for the selected announcement type is suitable, in which the parameters required for the second program segment provided are determined and in which the voice announcement corresponding to the selected announcement type is output, taking into account the determined parameters. The invention further relates to a voice announcement module and a program module therefor. <IMAGE>

Description

The invention relates to a method for outputting a Voice announcement according to the preamble of claim 1, a voice announcement module according to the preamble of claim 2 and a program module according to the preamble of claim 3.

Known voice announcement modules are designed to use announcements to output a simple structure. These mainly exist from a, possibly divided, fixed part and a variable part Part. Examples of this are time announcements ("It is now "," five "," clock "," thirty-two ".), also as part of one Timetable information, the announcement of a changed number or the cost of a long distance call just ended.

In the known voice announcement modules, announcements are more complicated Structure, for example with alternatives, cannot be automated possible.

The invention is based, with announcements More complicated structure, for example with alternatives, can be automated to enable.

According to the invention, this object is achieved by a method according to the teaching of claim 1, a voice announcement module according to the Teaching of claim 2 and a program module according to the teaching of Claim 3.

Ultimately, the invention is based on the idea of a programming language to create announcements and then use this to control the voice announcement.

Further embodiments of the invention are the subclaims and the following description.

The presented invention also has the advantage that the Program segments not only used for the control of announcements can be, but that also beyond that Documentation of a voice announcement module is possible directly.

Figur 1

zeigt eine Reihe von Beispielen von Ansagetypen, aus denen durch das erfindungsgemäße Verfahren in einem erfindungsgemäßen Sprachansagemodul ausgewählt und die entsprechend den aktuellen Parametern variiert und ausgegeben werden.

Figur 2

zeigt eine weitere Reihe von Beispielen wie in Figur 1.

Figur 3

zeigt mehr symbolisch ein Ausführungsbeispiel eines PCs mit einem erfindungsgemäßen Sprachansagemodul.

The invention is explained in more detail below with the aid of the accompanying drawings:

Figure 1

shows a series of examples of announcement types, from which the inventive method in a voice announcement module according to the invention is selected and which are varied and output according to the current parameters.

Figure 2

shows a further series of examples as in Figure 1.

Figure 3

shows more symbolically an embodiment of a PC with a voice announcement module according to the invention.

Using example 1 in FIG. 1, it is first shown briefly that how the invention can express itself.

But first a word about the language should be said. (This is also a note to those translators which later based on the English version in more Must translate languages.) On the one hand, the invention relates a voice announcement and therefore deals with language and on the other hand, the invention manifests itself in the form of a Programming language. So the language doesn't just occur here as a means of describing the invention, but also as content and expression of the invention in appearance.

Here, in Example 1 of Figure 1, as in most others Examples, it is assumed that the speech output in English language. The language elements to be output, e.g. B. Good morning, good afternoon, good evening therefore, as an expression of the invention, reproduced in English, although the description of the invention here in German Language takes place. The same applies even more to the program code as such, for example the variable names such as "<Area Code> ", or instructions like" loop infinite ". But it applies also for comments like "where ThreeToneSuite is defined like "in Example 8 in Figure 2, because according to the invention also non-executable "instructions" are allowed as comments and these are then usually written in English.

An exception is based on Examples 9 and 10 in Figure 2 shown and is in connection with the description of this Examples discussed.

So what in the original German language version in German language is expressed is description of the invention while what is expressed in English brought, content or form of expression of the invention is. (This can be done in the other translations as well can be distinguished.) But also in English Version will at least be what is in the drawing in the examples called and in the description from the examples of Drawing is cited as the content or form of expression of the invention be clearly recognizable.

Of course, as with inventions in others Areas of technology, country-specific adjustments not excluded. Here, however, such country-specific Adjustments primarily by replacing the syntax elements by expressing them in the respective national language based.

But now back to Example 1 of Figure 1:

First, "[Good morning, Good afternoon, Good evening]" a salutation selected according to the time of day and output, whereby the selection is made automatically. Subsequently will be a new number including the required area code announced by two fixed sentence components, "the number you have dialed has changed. Please dial the area code "and "followed by the new number", with two variables, "<Area Code>" and "<B-party number>" alternate. In an endless loop, "{loop infinite | <half second pause> I repeat, please dial the area code <area code> followed by the new number <B-party number> } ", the output will then be added as long a pause, "<half second pause>", and a reference to the Repeat, "I repeat", repeated until the addressee of the issue this breaks off. The break is here for the sake of simplicity as a variable with a predeterminable length, but without language content spent.

In this example, the syntax elements "Salutation", "are fixed Sentence components "," Variables "and" Loops "used and exemplary presented. Before going into further examples the syntax of this programming language is briefly summarized, but not necessarily completely, shown; a Supplement, modification or even a reduction should not be excluded.

When comparing with a common programming language like C, C ++ or Pascal shows that the programming language in which the method according to the invention expresses itself without express Commands ("explicit commands"). Rather, it exists from references and "separators". The References represent the connections (left) to where there is the required information during the disconnections define the context in which the references are placed are. The references, such as a text sequence, are only related with a context of importance. By using a certain separation, such as "<" or ">", the context and to change the meaning of the reference.

In practice, an automatic translator becomes the most natural References available using a language special table depending on the context in an application-specific Format transferred, then from the invention Voice announcement module can be read. To input This programming language supports making it more flexible and readable also punctuation symbols. These are unchangeable characters used in translation have no meaning. They can be added by the user or removed to improve the look and readability improve, but have no technical impact.

This way it can be very easy and even after a short one Incorporation of a voice announcement module can be programmed, whereby the result, i.e. the program, both directly for the documentation is suitable as well as after automatic translation translated into a lower application specific language can be.

An announcement is a collection of fixed elements ("elements"), like "no connection under this number" and complex Segments ("Complex_Segments"), for the specific application runtime-specific information, such as parameters, is still required are. For example, the announcement "The call costs <x> DM "the" <x> "for a parameter that is only known is when the announcement is really made. Where is "runtime" (English: runtime) an expression for the moment when the compilation of a current voice announcement actually done.

In the following (...) means an additional possibility ("option"), [... v ... v ...] a choice ("choice") and {...} one Extension.

An announcement thus has the following structure:
Announcement = (Prefix), Segment_List, (Suffix),
in which:

An expression, "Expression" is an algebraic expression, which is evaluated at runtime and has known operators, such as +, -, /, *,% (modulo),! (not), and, or, xor, =, <>, <=,> =, <,> or any other system specific Operators. For example, "(Month = 2) and (day> 15) "a valid expression.

1. eine Schleifenanweisung wie "loop 5 times" oder "loop infinite" (nur bei Ausgabe von Wörtern): ({loop|...}),

2. eine Anweisung für eine Dauer wie "10seconds" (nur bei Ausgabe von Wörtern): ({duration|...}),

3. eine Anweisung zur Wiederholung wie "loop as long the parameter buffer is not empty": ({repetitive|...}) oder

4. ein Ausdruck, der zur Laufzeit auszuwerten ist, wie "Euro=0 and Cent<>0", der zusammen mit dem IfThenElse_Segment ({Euro=0 and Cent<>0|(then)|(else)}) oder dem Branch_Segment ({Key pressed|1,2:...|5:...|ELSE:...}) benutzt wird.

A "Block_Reference" is a reference to a detail like

1. a loop instruction such as "loop 5 times" or "loop infinite" (only when words are output): ({loop | ...}),

2. an instruction for a duration like "10seconds" (only when words are output): ({duration | ...}),

3. an instruction for repetition such as "loop as long the parameter buffer is not empty": ({repetitive | ...}) or

4. an expression to be evaluated at runtime, such as "Euro = 0 and Cent <>0", which together with the IfThenElse_Segment ({Euro = 0 and Cent <> 0 | (then) | (else)}) or the Branch_Segment ({Key pressed | 1,2: ... | 5: ... | ELSE: ...}) is used.

1. ein Parameterwert und die dazugehörige Regel, nach der er erzeugt wird,

2. eine (große) "Element_List" oder

3. jede komplexe Einzelheit, die als Merkmal vorliegt.

A Complex_Reference is a reference to a complex detail like

1. a parameter value and the associated rule according to which it is generated,

2. a (large) "Element_List" or

3. every complex detail that is present as a characteristic.

A "Macro_Reference" is a reference to a macro, which in turn is a collection of elements and segments, as well any announcement in which an element or part of it is symbolic can be replaced by a parameter. A macro is at the compilation resolved, with each Macro_Parameter by its value taken from the parameter list "Parameter_List" is replaced. A macro works here just like in any known one higher programming language, for example C ++.

A "Parameter_Number" is a natural number that represents the nth parameter representing "Parameter_List".

A "punctuation sign" is any character that is not the character set "/; <> {} () [] |" listened to. It can be saved separately become and need not be part of an element (an element can Be part of various announcements). This sign is for the spoken announcement without meaning.

The "Flush sign" is intended as a sign for carriage return or be treated for a new line. It only matters if, for example in addition to the voice output, a text output he follows. For example, the so-called "flush sign" "Backslash", "\", a character used for other purposes are not needed.

Before going into technical implementation, should briefly refer to the other shown in Figures 1 and 2 Examples are included:

Example 2 according to FIG. 1 shows the programming for the announcement a price, assuming that the call is not is or was completely in vain. The syntax of the "Parameter_Fixed" used. First there is a fixed announcement component, "Your call costs", followed by an entry in Dollars and cents. The current values for this, <Dollar> and <Cent>, are used when this announcement is called becomes. The associated units, dollars, dollars, cents or Cents, are added afterwards, with each by value the singular form, dollars or cents, or the plural form, Dollars or cents that unit is spent. Is one of the two values zero, this is not output and also suppresses the associated unit. The connecting "and" is suppressed if the preceding value, <Dollar>, Is zero as well if the subsequent value, <Cent>, is zero is.

Example 3 in Figure 1 shows an announcement similar to that in the example 2, but only rounded up or down to whole dollars. Here Simply suppressing information is no longer sufficient. Therefore here by the Branch_Segment DollarUnequalZero distinguished the cases that something paid is "costs <Dollar> (Dollars / Dollars)" and that in total nothing to pay is "is free". The suppression of unity is not necessary here, because this branch just started anyway if the value is non-zero.

In the event that on the one hand you want to bill exactly, on the other hand, free calls are also possible of course, a combination of both examples but then the Branch_Segment DollarUnequalZero defines it differently or through another Branch_Segment, such as PriceUnequalZero needs to be replaced. The instructions DollarUnequalZero and PriceUnequalZero have to do this, as with Programming languages common, elsewhere, for example as a function or subroutine can be defined. Instead of Branch_Segments DollarUnequalZero could also do this directly the following equation (Dollar • Zero) can be entered. In such simple cases, it may not be worthwhile to have a separate one Provide function and call it up. To what extent this also applies to a modified branch segment, for example PriceUnequalZero would apply, would not depend on this here defined internal structure.

The following three examples, Example 4 and Example 5 Figure 1 and Example 6 from Figure 2 show examples of possible Confirmations of changes made by the user User profile. These three examples each show one first announcement part, in which the type of change made "Your calls will be forwarded" or "You entered a filter for the following directory numbers ", and one second part, which confirms the content of the change made, namely in the case of call forwarding or forwarding the respective Goal "to your mobile telephone!" or "to your answer machine. "and in the case of entering a number of phone numbers for a phone number filter are the Block_Reference {Repetitive | <DN>} all entered phone numbers are output for checking.

Using example 7 in FIG. 2, a more complex one is to be constructed Announcement. A statement is shown for a person waiting in a queue. First there is a Greeting: "Hello, welcome by XYZ. All lines are busy. Your call is being queued. ", to which is repeated by a Block_Reference, {loop infinite | ...}, pointed out the fact is that you are currently in a queue: "You are the <xth> in the queue". It is also used as a variable "<xth>" specified the number of times in this queue is. From the speed at which it counts down you can estimate how much longer you have to wait got to. To make the end of the queue closer is to be made by another Block_Reference, {pool | 1: <music1> | 2: <music2> | ELSE: <wait tone>}, for a variation the statement against the end. Every time the infinite loop "loop infinite" is still an acoustic Output character that is selected from a pool, "pool" and usually, "ELSE", a wait character, "<wait tone>", at the end of the queue by two different ones Music sequences, first "<music2>", then "<music1>" is replaced. The waiting sign could, as in the example below, consist of a predefined sequence of tones.

Example 8 in FIG. 2 shows a simple output for the In the event that there is no connection under the dialed number. The output is first through the execution of a macro instruction "« ThreeToneSuite »" and then by a simple one Edition "No abonnee on this number" completed. In this Example is hinted at another possibility, comments to use. As in any programming language, here too provision is made for comments to be inserted. One in possibility provided in many programming languages in that everything is ignored, which is after a quotation mark, "'" follows; this way the commentary is here 'where ThreeToneSuite is defined like "<Tone800Hz>; <Tone1000Hz>; <Tone1200Hz>; <Pause500ms> "'added to briefly explain the content of this macro. The macro itself must can of course be defined elsewhere.

The last two examples, Example 9 and Example 10 Figure 2, show the same output, namely the indication a cost amount in a Slavic language, namely in Russian. Slavic languages have two different plural formations, that must be taken into account in the output. (Strictly speaking the genitive singular is used here as a second plural formation used.)

To do this here in a not available in a Slavic language To show text, the currency information, "Roubel / Rublei / Roublia", which is the subject here, in Russian Language, but translated into English, reproduced. The Rest, both the announcements themselves, "Your call costs", as well Comments, ((Russian announcement)), and other program elements, <Costs>, are shown in English. Here the use of double parentheses as a further type considered adding a comment. Such a comment should just be overlooked while using a quotation mark "'" as a sign for a comment does not do any of the following more should be considered.

In example 10 in FIG. 2, both types of comments are closed see. In example 9 it is assumed here that with the call of the variable <Costs> automatically the macro for the selection of the correct currency unit is called while this is in Example 10 explicitly using the macro call «SlavicParFixed | <Costs>; Roubel; Rublei; Roublia ». In this example 10 more detailed comments are also added, for example the definition of the macro is attached as a comment.

Only if the value of the variable <Costs> is "1" is issued the singular form "Roubel" as currency unit. Ends if the value is not equal to "1", the value of the variable <Costs> on 1, 2, 3 or 4, but not on 11, 12, 13 or 14, so the form "Roublia" is used, while in all others Cases the form "rublei" is used.

Of course, the different announcements are not saved as shown in Figures 1 and 2. In particular serves the graphic representation used here, among others also with different fonts, for better understanding, cannot be saved in this way. A visualization in exactly this form but with an appropriately trained Editor possible. The actual storage takes place but in the form of tables.

For example, there is a first table for saving of announcements as such. This table shows one first column with unique identifiers, a second column with the names of the respective announcements, a third column with the language used, ultimately a comment, and a fourth column with the definitions of the announcements. A column with comments can also be provided. Instead However, a further table can also be provided contains only comments and about the unique identifiers with Rows from this table or with rows from other tables is linked.

A second table contains the "Block_Segments", the columns for unique identifiers, for the names that "Block_References", and for the definitions of these "Block_Segments" has. Further definitions are in others Tables saved.

In conclusion, let's briefly look at the more external form of realization To be received. 3 shows one for this purpose known PC, the appropriate equipment as inventive Voice announcement module with a program module according to the invention Is provided. The screen shown in this example as output means and the keyboard shown here as Input devices are not absolutely necessary in the basic configuration. The voice announcement module is then part of a so-called servers, i.e. a computer that only works in a network works with other computers and for certain tasks takes over.

For example, it can be an operator's fee calculator Telecommunications network, on the ongoing charge information be sent and not just for the monthly Fee statements should be used, but after Ending a conversation as the costs incurred Outputs language. These costs can be derived from the data content of one the connection data of this conversation containing data record be determined. A second may then become Data record that contains specific information about the connection which caused these costs, derived in which Language to be output. Examples of this are the Examples 2, 3, 9 and 10 from FIGS. 1 and 2.

This "server" can also be a switching computer, the incoming calls in a so-called call center the individual operator stations distributed. Access from Pages of the voice announcement module on the data records of the On hold waiting calls may be possible for each individual of which determine the current position and, as in the example 7 shown in Figure 2, to be able to output as language.

In both cases, as in any other case, are excluded the indicated data access means still voice output means required with which those formed by the voice announcement module Voice announcements can actually be output. In the implementation must be done in a manner known per se the voice announcement compiled in individual cases into spoken Language and its output via a suitable (analog or digital) connection to a telecommunications network.

The data set that is specific to those already mentioned above Contains information about a connection is often also called a "user profile" designated. A user can now be allowed at least parts of this record by some sort the remote input itself, and that the result of this Remote entry then as a language for control and confirmation is issued again. Examples are the examples 4, 5 and 6 of Figures 1 and 2.

Examples 1 and 8 of Figures 1 and 2 show two possible ones Announcements from a switching computer, in the event that under no subscriber to the dialed number via this switching computer can be achieved. Here then either, if known, a new subscriber number (example 1) issued or it is announced that this subscriber number there is no entry (example 8). The request in the Subscriber database of this switching computer then contains either a record that is under the previous, now selected Call number contains a reference to the new number or it comes in response to the data access means data access an error message as a special form of a data record back, which says that there is no entry.

Such computers are often used with the invention Voice announcement modules can be equipped, but also at the same time used to maintain at least part of the data, are the basis for the voice prompts. In all of these cases, for example in hotels or hospitals, they change Data is ongoing, including input and output means such as also shown in Figure 3 are required. Change here organizational processes, which in turn an adjustment of the voice prompts required. Especially with however, the present invention is programming newer or changing existing announcements is relatively easy and can also be carried out by less trained personnel. Therefore in such cases it is advantageous if the voice announcement module is also equipped with suitable editing tools. Therefor can usually those that already exist in almost all computers Editors or a writing program can be used. In In this case, the computer used should also have and output options (microphone, loudspeaker) for spoken Have language to record texts and programmed Test announcements.

The structure of the program module according to the invention for use for implementation in a voice announcement module according to the invention of the inventive method for outputting a voice announcement is common in itself. It’s part of normal programming, a program in segments (subroutines, procedures) split, with individual segments more for control of the process and others for carrying out individual activities are intended this program with drivers for data access and to connect to voice output and with the above to fill the inventive content shown.

Claims (4)

Method for outputting a voice announcement, the content of which depends on the data content of at least one data record, characterized in that an announcement type is selected on the basis of the data content by means of a first program segment, that a second program segment is provided for program-controlled output, which is available for the selected announcement type it is suitable that the parameters required for the second program segment provided are determined and that the voice announcement corresponding to the selected announcement type is output, taking into account the parameters determined. Voice announcement module for outputting a voice announcement, the content of which depends on the data content of at least one data record, characterized in that the voice announcement module has a program control, that a first program segment is present, by means of which an announcement type can be selected on the basis of the data content, that each announcement type second program segment from a plurality of second program segments is assigned that data access means are available to determine the parameters required for a selected second program segment from the at least one data record, and that speech output means are available to provide the voice announcement corresponding to the selected announcement type taking into account the determined parameters to spend. Program module for use in a voice announcement module for outputting a voice announcement, the content of which depends on the data content of at least one data record, characterized in that the program module has a program control segment, a first program segment and a plurality of second program segments that the program control segment for control of access to data access means and speech output means and for controlling the call of the first program segment and a program segment from the multiplicity of the second program segments is designed such that the first program segment is configured such that an announcement type and a second program segment assigned to this announcement type from the multiplicity of second program segments is selectable, and that each of the second program segments from the plurality of second program segments is designed in such a way that, with the aid of the required parameters, which are determined by means of the program control segment and the D Data access means are determined from the at least one data record, forms a current voice announcement, and that it uses the program control segment and the voice output means to output this voice announcement. Voice announcement module according to claim 2, characterized in that there are editing means by means of which existing announcement types with associated second program segments can be changed or deleted or new announcement types can be generated with associated second program segments, and that input and output means are available via which communication with the editing means can be.

Images