HUH3805A - Speaking information system - Google Patents

Abstract

Field of the Invention The present invention relates to a speaker information system having a processor unit (10) having a speech storage device (30), a memory unit (20) associated with a memory unit (30), and a sound generating unit (20), with its own program storage memory, registers, control outputs, and data buses. (80). The processor unit (10) connects to a bus counting unit (20) "(20)", a memory unit (30) on a bus driver unit (40), and a memory unit (30) on the sound generating unit (80h). connected to a loudspeaker (104) 'and an internal speaker (105), and the processor unit (10) has an "analog input (13)" to which an external microphone (101) ζ "and / or an internal microphone (70) is passed through the microphone unit (70). a microphone (102) and / or a conductive microphone (103) is connected, and a conductive microphone (103) "to the sound generating unit (80) X, and a processing unit (lOJ ^ line interface (90)" "" control instructions for the soccer station and output line output) (106) "and connected to line input (107) (Figure 1) 25 30 35

Description

SPEAKING INFORMATION SYSTEM

BACKGROUND OF THE INVENTION The present invention relates to a speech information system having a processor unit having its own program storage memory, registers, control outputs, data buses, a memory unit for recording speech information, an audio signaling unit associated with a memory unit, and an audio unit. In various areas, such as public transport, so-called voice response devices are used which are computer controlled. Such computers have their own program memory, registers, control outputs and data buses. Such a processor unit further comprises recorded data of speech or some speech elements on a special medium, i.e. a memory unit. From this, an audio signal generating unit generates an audio frequency signal and converts it into an intelligible speech supplemented by a loudspeaker (see Technical Lexicon, Akadémiai Kiadó, Budapest 1978, Volume IV: Voice Response Equipment).

However, the disadvantage of such a voice-generating, or speech-information system, is that managing a computer is a relatively complex task and therefore cannot be handled in the traffic.

To address this shortcoming, we have developed a talking information system that is well-suited to informing passengers on board public transport vehicles, but can also be used on other devices that require the storage of large amounts of speech and the simple reproduction of certain details of speech. The device integrates with a program for organizing, storing and encoding data that can properly reproduce properly prepared speech information.

According to the present invention, the talking information system mentioned in the introduction is further improved by connecting the processor unit via a bus drive unit to an interrelated address counter, a tone generator, a memory unit, and a sound generator unit through an audio unit to an external speaker and an internal speaker. they have analog inputs, · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·

DANUBIA 81492-6009 to which an external microphone and / or an internal microphone and / or a master microphone are connected via a microphone unit, and the leading microphone is connected to the voice generator unit and to the line output and line input receiving and output instructions controlling the processor unit through a line interface.

In a preferred embodiment of the speaking information system of the invention, the memory unit comprises at least one flash memory card connected to the address counter by address buses.

In a preferred embodiment of the speaking information system of the present invention, the audio signaling unit comprises a bar code sensor unit for comparing the values output by the memory unit and counting the number of matches, a step register for receiving data output from the memory unit and a D / A converter. connected to a clock generator via a frequency divider, and the start input of the clock generator and the landmark sensor unit is connected to a tilt circuit, the audio generator unit has a busy signal output which is connected to the programmable logic circuit and the processor unit the frequency output of the frequency divider output is connected to the clock input of the address counter.

In a preferred embodiment of the speaking information system according to the invention, the audio unit comprises a low pass filter connected to the audio output of the audio output unit, the output of which is a volume control (and a mixing unit and mixing unit through an output stage connected to their output, it is connected to an external speaker on the one hand and to an internal speaker on the other, and to a microphone leading to an input of a mixing unit (ja).

In a preferred embodiment of the speaking information system according to the invention, the memory unit comprises two flash memory cards which are connected to the address counter via address buses and the data from one flash memory card to the processor unit to the other flash memory card.

DANUBIA 81492-6009 • · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·

The invention will now be described in more detail with reference to the exemplary embodiment shown in the accompanying drawings. In the drawing:

Figure 1 is a block diagram of a talking information system according to the invention, Figure 2 is a more detailed block diagram of a voice signal generating unit of Figure 1, Figure 3 shows an arrangement of information stored in a memory unit 30, Figure 5 is a flowchart illustrating the operation of the speech information system of Figure 1.

The talking information system of the present invention is provided with its own program storage memory, registers, analogue-to-digital converters, appropriate outputs and not shown separately. The system includes a processor unit 10 and a memory unit 30 for storing a large amount of speech information, and a 20 t milk milk connection via data bus 44. / audio signaling unit, which is connected to 80 sound generating units. The sound generating unit 80 comprises a low pass filter 81, a first and a second volume regulator 82 and 83 connected to its output, the outputs of which are connected to one of the inputs of the first and second mixing units 84 and 85. The volume controllers 82 and 83 are provided with control inputs which are connected to the control output 14 of the processor unit 10. The mixing units 84 and 85 are connected via end stages 86 and 87 to an external speaker 104 and to an internal speaker 105, respectively. The address inputs of the memory unit 30 are connected to the address counter 50 via 51 address lines. The address counter 50 has an input 53 which is connected to the bus drive unit 40 via the data bus 44 and to the processor unit 10 via the data bus 41. The address counter 50 is connected to 60 programmable logic circuits via address rails 43 (A23-A26) which transmit the highest local value bits. The programmable logic circuit 60 is further coupled to the processor unit 10 via control bus 42. Outputs of 60 programmable logic circuits

DANUBIA 81492-6009 - 4 forming control wires 61 are connected to the respective control inputs of the address counter 50 and the memory unit 30. The 20 voice signal generating unit has and is active 21 Busy output during phonation Duration ⁵ tama which is connected to a part of the 60 programmable logic circuit, on the other hand the processor 10 units indítóbemenete associated with the processor 10. Unit 15 trigger output. The audio signal generating unit 20 also outputs a clock output 23 that schedules the rate of voice generation, which is connected to the clock input 52 of the address counter 50. The processor unit 10 is provided with analog inputs 13 to which rectifiers 74, 75 and 76 are connected. They are connected via input preamplifiers 71, 72 and 73 respectively to an external microphone 101, an internal microphone 102 and a guide microphone 103. The rectifiers 74-76 and the preamplifiers 71-73 together form 70 microphone units. The processor unit 10 is provided on line adapter 90 for line outputs 106 and line input 107 for receiving and issuing control commands. > TT06 | Zes'

the. The 40T line input is preferably adapted to or is part of a party line. The processor unit 10 is further provided with a control switch 11 and a copying switch which initiate the appropriate processes. (Ib)

Optionally, the memory unit 30 comprises one or more, in this case two flash memory cards 31 and 32, into which the speech information must be loaded separately before use.

The use of the speech information system according to the invention also requires a separate computer workstation to run a program controlling the recording, organizing, storing and encoding of the speech information. The function of this program is that the speech information system according to the invention is to be reproduced at a desired time and in the order of timing, instruction given by the operator or some sensor. This speech or these speech details must be digitized at this computer workplace and recorded in the appropriate order and resolution. In order for each detail to be identified later, the position of each detail within the memory must be ____, (¼) The 2-0 42 fveMépLCLC oJuA Chfet-eA / tS out escape path k> u. aunt Á- -Yh Luutuet-e · · · · · · · · · · · · · ·

DANUBIA 81492-6009 gyis start address. Each speech fragment can be placed in a common file. In this case, the relative position of these details within the file, ie their relative address, must also be provided. Thus, a database containing the individual texts, the titles of the texts, the details within the texts separately and the relative titles of the details can be compiled. In order to avoid having a separate title at the end of the details, the details should preferably be separated by a set of codes that do not occur in speech.

The speech information can be loaded into the memory unit 30 as follows. At a separate and unpublished workplace, a person with the right organ will cancel the desired text with a microphone and record it in digital form. With the help of a suitable sound management program, unnecessary silence and noises are cut off from the beginning and end of the text. If the result is correct, this speech snippet can be stored with the appropriate identifier. In the same way, any number of speech fragments can be recorded and stored. The related parts of speech can be merged into a single file, which of course has to be given the beginning of each part relative to the beginning of the file, ie its relative title.

The speech parts are separated from each other by a series of codes that are also checked separately by the program. This is made possible by the use of so-called delta code modulation for coding. A feature of this is that silence generates evenly alternating bits during encoding (..., 1, 0, 1, 0, ...), which are written in hexadecimal format, 55h or AAh. From this it can be seen that both codes are actually silent, and one of them can be replaced with the other one without any problems. One of these multiple repetitions can be used to separate speech parts, i.e., to indicate the end of speech parts. For this purpose, six AAh codes were selected and added to the end of each speech fragment after recording the speech details. Although this sequence of codes is unlikely to occur randomly, it can occur in speech, so every single part of speech must be passed through.

DANUBIA 81492-6009 to determine if there is such a code sequence and if so replace it with a 55h code sequence. After the exchange is completed, such a sequence of codes can no longer occur in the speech sample, but must be placed at the end of each speech fragment to separate them within the file.

The files thus prepared must be copied to a memory unit 30 which can be inserted into the speaking information system of the present invention.

For ease of use, so-called flash memory cards, which are designed to be connected to PCMCIA sockets on laptops, are conveniently used to store speech. The embodiment shown in Figure 1 comprises two such flash memory cards 31, 32, one removable and a constant. They are equivalent in function. The placement of the information is illustrated in Figure 3. Memory is essentially divided into five areas. The FAT area 1 contains the relative addresses of the files containing the speech details. This is followed by the first 2 barcodes, which is a series of identifiable codes, in this case a code of sixteen bytes in length. Then there are 3 relative addresses, which are sixteen bytes long, suitable for identifying the speech parts within each file. kh / gu followed by masodikvhat bar code. Next, the files containing the speech details are actually located.

The figure does not show the so-called attribute memory of the flash memory cards 31, 32 in which the factory-programmed and card-related data are written in an indelible manner. Apart from this area, the remainder of the memory can be read and written bytes, and these functions are controlled by programmable logic circuit 60 via control wires 61. This circuit provides addressing of the flash memory cards 30, 31, rotation of the bus drive unit 40 and disabling during speech (reading).

The construction of the sound signal generating unit 20 is illustrated in Figure 3. The tone generator 20 includes a landmark sensor 24 that compares the values output by the memory unit 30 and counts the number of matches, a 25-register register that receives data output from the memory unit 30 and is connected to a D / A converter 29, and 25-step ·· ·· · · · · ·······••••••••••••••••••

: ...: ..... ·: · .. ·

DANUBIA 81492-6009

The register clock input is connected to a clock generator 26 via a frequency divider 28, and the start input of the clock generator 26 and the barcode detector unit 24 is connected to a tilt circuit 27; the audio generator 20 has a busy signal output 21 connected to the programmable logic circuitry and, at tilt 27, it is coupled to the start output 15 of the processor unit 10 and the clock output of the frequency divider 28 is coupled to the input of the address counter 50.

The signal generator 20 operates as follows. The 26 clock generator determines the rate of sound generation. The frequency is set by a capacitor C connected to the appropriate input of the 26 clock generator. Sound generation occurs at a rate similar to the sampling rate used in coding, but is not sensitive to slow rate changes and therefore does not require the use of an accurate quartz oscillator. With this frequency, the information entered in the shift register 25 via the data bus 44 is transmitted at bit rate to the D / A converter 29. In fact, it is designed as a coupled integrator whose output, for example, is a linearly increasing signal with a value of 1 per input, and a linearly decreasing signal with a value of 0. This output forms the audio signal output 22. In the step 5 register, the data is entered at the rate of reading from the memory unit 30, i.e. bytes. To schedule this, a frequency divider 28 is connected to the output of the clock generator 26, which is actually the output of the clock signal 23. The audio signal generation is enabled at tilt 27, which is in fact a RS bistable multivibrator, the enable activation input of which is connected to the start output 15 of the processor unit 10. At start-up, tilt 27 starts the clock generator 26 and signals on the output indicator 21 the processor unit 10 and the programmable logic circuit 60 that an audio signal is being generated, and during this period, nothing but the step counter 50 is incremented. generating an audio signal based on data read from a memory unit.

Simultaneously with the generation of the audio signal, the same weep $ a is made at the end of the speech portions read from the 30 memory units.

• · · · · · · ··· · · ·

DANUBIA 81492-6009

As mentioned above, the end of speech details is indicated by AAh (hexadecimal) codes. To this end, the barcode sensor unit 24 compares the bytes on the data bus 44 with the code AAh and, when matched, increments the content of a counter included therein and not shown separately by one. If the comparator indicates a successive number, in our case six matches, the bar code sensor unit 24 signals and tilts the tilt loop 27. This indicates that the speech portion you want to speak is complete.

All these operations are illustrated in Figure 4. In the start step, the processor unit 10 tilts the tilt loop 27. In the next step, the busy signal output 21 informs the processor unit 10 that an audio signal is being generated. The first data byte is then read from the memory unit 30. In the next step, this data is compared by the bar code sensor unit 24 with the preferred code, and if it does not match, it is cleared by its counter. If it does, it increases the content of the counter while allowing the data byte to be converted to an analog signal, that is, an audio signal. It then checks that the detected AA code is sequentially queued. When it reaches six, it will tilt the 27 tilt circles to indicate that the speech segment is over. If not, it returns to retrieving the next data byte.

Figure 5 illustrates the operation of the entire system. After powering on, the processor unit 10 performs a system initial setup based on a program written in its program memory and then checks the presence and contents of the memory unit 30. It also checks the type and capacity of the flash memory cards 31, 32. If an error is found during the check, the error is displayed and, for example, an alarm sounds. If no error is detected, it scans the contents of the memory unit 30 for sixteen bytes first and second boundary codes 2 and 4 · Stores the end of these codes for relative addressing as its first and second offset values - in its internal registers. When searching for codes, you decide whether or not to find the codes. If you do not find the correct code, it will give you an error message and then wait for it. In doing so, examine whether the line input 107 is receiving a command. The command contains the identification of the speech part to be announced • · · • ·

DANUBIA 8149.2-6009 • · · · · · · · · · · · · · · · · · · · · · · · · · This data is usually a 4 byte relative address to which the first offset value is added from the relative address range 3 of the flash memory card 31 or 32, typically based on four bytes of data to find the relative address of the speech snippet within the file range 5. From the sum of the previously stored second offset value and the relative address, the absolute address of the location of the speech portion within the range 5 of the flash memory card 31 or 32 can be determined in the next step. This absolute address is entered by the processor unit 10 under the control of the bus drive unit 40 via the data bus 44 into the address counter 50, i.e., the address counter 50 starts from the address given to the input 53 and contains its contents during the speech signal at eight times the sampling rate grow. This value is transmitted to the addressees 51 on the memory unit 30, and the next byte containing the speech portion to be spoken is digitally addressed, respectively. At the same time, the trigger output 15 tilts the tilt loop 27, which is waiting for the speech to end. This is indicated by the bar code sensor unit 24. When the speech portion is over, the system returns to the standby mode. In this case, the processor unit 10 can write and read flash memory cards 31, 32 per byte.

If a malfunction occurs in the system during startup or startup, a general reset can be performed by actuating the control switch 11 and a system check can be performed by starting a program permanently written to the processor unit 10. By actuating the copying switch 12, the contents of the flash memory cards 31 and 32 of the memory unit 30 can be copied. For example, vehicles with the same flight number need to be charged with a single flash memory card and their contents can be copied to the remaining flash memory in the vehicle by plugging them into the correct location when starting each vehicle. However, a temporary change does not necessarily require copying, but reads the contents of the temporarily plugged in flash memory card, which, for example, contains the changed name of the stops according to the temporary diversion.

• · · ·

DANUBIA 81492-6009

Of course, you may also need to utter a text other than pre-recorded speech. The conductor microphone 103 serves for this purpose, and the level of the audio signal received therefrom is detected by a rectifier 76, and the detected level is output to one of the 13 analog inputs of the processor unit 10. When the driver speaks to the driver microphone 103, the processor unit 10, by controlling the volume controls 82 and 83, mutes or mutes the sound transmitted by the tone generator unit 20. In this case, the driver's voice is heard from the outer speaker 104 and the internal speaker 105.

Claims (5)

PATENT CLAIMS A talking information system having a processor unit having its own program storage memory, registers, control outputs, data buses, a voice information recording unit, an audio signaling unit associated with the memory unit, characterized in that the processor unit / bus drive unit (40) connected to an interrelated address counter (50), a tone generator (20), a memory unit (£ 3F), and the audio generator (20) is connected to an external speaker (104) and an internal speaker (105) via a voice generator (80); the processor unit (10) having analogue inputs (13) to which an external microphone (101) and / or an internal microphone (102) and / or a leading microphone (103) are connected via a microphone unit (70), and the leading microphone (103). ) controls the voice acquisition unit (80) and the processor unit (10) via a line interface (90) instructions for receiving and issuing line-out (106) and line input (107). A talking information system according to claim 1, characterized in that the memory unit (30) comprises at least one flash memory card (31, 32) connected to the address counter (50) by address buses (51). The speech information system according to claim 1 or 2, characterized in that the tone generator (20) compares the values output by the memory unit (30) with a prescribed code and counts the number of matches by a bar code detector (24), the memory unit (30). ) includes a step register (25) connected to a D / A converter (25), and the clock input of the step register (25) is connected via a frequency divider (28) to a clock generator (26) and a clock generator (26). ) and the trigger input of the landmark sensing unit (24) is connected to a tilt circuit (27) and the acoustic signal generator (£ 20 unit occupancy indicator (21 / output / Van) connected to the programmable logic circuit (60) and the processor unit (10)), and the tilt lever (27) is connected to the start output (15) of the processor unit (10) and clocked by the output of the frequency divider (28) signal output (23) clock of address counter (50) The DANUBIA 81492-6009 processor is connected to its input (52). 4. A speech information system according to any one of claims 1 to 6, characterized in that the voice generator (80) comprises a low pass filter (81) connected to the audio signal output (22) of the audio signal generating unit (20). your controllers (82jE volume - sabbath as well as your mixing units (8 £ 4, * 9L) and connected to an external speaker (104) through an output stage (86 and 87) and an internal speaker (105) - is connected to a microphone (103) through one of its inputs (8 & y) via a preamplifier (73). 5. A talking information system according to any one of claims 1 to 5, characterized in that the memory unit (30) comprises two flash memory cards (31, 32) connected via address buses (51) to the address counter (50) and to the processor unit (10) from one flash memory card. (31) a copying switch (12) for copying the data on it to the other flash memory card (32).