KR20000003717A - Voice communicating method using voice recognition/synthesis and apparatus performing the same - Google Patents

Abstract

PURPOSE: A voice communication method is provided for a user to recognize a voice signal at a receiving side. CONSTITUTION: The voice communication method using a voice recognition/synthesis comprises the steps of: transmitting a voice data packet to a data network from a receiving side, wherein the voice data packet has a header to which at least transmission frame number is written(500); comparing the transmission frame number with a predetermined transmission frame number at a transmitting side to calculate frame loss rate(504 to 510); transmitting the header, to which the calculated frame loss rate is written, to the transmitting side; and transmitting character data packet from the receiving side by use of voice recognition/synthesis instead of the voice data packet when the frame loss rate is over a predetermined threshold value(512).

Description

Voice communication method and apparatus using voice recognition / synthesis

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to multimedia communications, and more particularly, to an apparatus and method for speech communication using speech recognition / synthesis when heavy traffic is present.

1 is a block diagram illustrating a conventional multimedia communication environment, in which a system integrating a video and audio communication device for video and audio communication is schematically a video input / output device 100 and 102 and a voice input / output device 110. 112, image encoders / decoders 120 and 122, speech coders / decoders 130 and 132, and multiplexers (MUX) and demultiplexers (DEMUX) 150.

Referring to FIG. 1, referring to a transmission part for transmitting a conventional video and audio, the video and audio input to the video input device 100 and the audio input device 110, respectively, may be an image encoder 120 and an audio encoder 130. Are each encoded through The MUX / DEMUX 150 generates a data stream in consideration of synchronization of two encoded signals and transmits the data stream to a target system through a predetermined communication network. As such, in general, in a communication network that transmits video and audio, traffic changes frequently, and in consideration of this characteristic, the data amount of the stream is adjusted according to the situation. On the contrary, the signal received by the MUX / DEMUX 150 is divided into an image data stream and an audio data stream, and the original image and the original audio are obtained through the image decoder 122 and the audio decoder 132, respectively. .

The technology applied to the network traffic has accumulated technology as the multimedia communication is in the spotlight. However, when the traffic is heavy, the quality of communication is inevitably deteriorated. As a result, the sender does not receive all of the data sent by the sender or the sender adapts the traffic and sends only a small amount of data. In this case, if a certain scene is lost in the received image, the user will see unnatural image scenes, but generally the overall information can be perceived. However, if a certain amount of data is lost in the received voice, the user may not be able to recognize the information and may feel uncomfortable compared to the image.

The technical problem to be achieved by the present invention is to check the frame loss ratio during the data transmission process, by using voice recognition / voice synthesis in case of heavy traffic, so that the user can recognize the voice signal at the receiving side, voice recognition / The present invention provides a voice communication method using synthesis.

Another object of the present invention is to provide a voice communication apparatus using voice recognition / synthesis for performing the voice communication method.

1 is a block diagram illustrating a conventional multimedia communication environment.

2 is a diagram conceptually illustrating data transmission between two systems to which the present invention is applied.

3 is a block diagram illustrating a voice communication apparatus according to the present invention.

4 (a) and (c) are diagrams for explaining the transmission data structure used in the present invention.

5 is a flowchart for explaining a voice communication method according to the present invention.

In order to achieve the above object, in the case of heavy traffic, the voice communication method according to the present invention using voice recognition / synthesis comprises the steps of: transmitting a voice data packet having a header with a specified transmission frame number at a transmitting side to a data communication network; Comparing the transmission frame number of the received voice data packet with the predetermined transmission frame number to calculate the frame loss ratio, transmitting the header indicating the calculated frame loss ratio back to the transmitting side, and the frame loss ratio being predetermined. When the threshold value is exceeded, thereafter, the transmitting side transmits a text data packet using voice recognition / synthesis in place of the voice data packet.

In order to achieve the above object, the voice communication apparatus according to the present invention using voice recognition / synthesis when the traffic is heavy, processes the voice data to be transmitted in the voice mode in response to the control signal or in the voice recognition / synthesis mode. When the voice mode is selected in the transmission / reception mode mode switching unit for selecting whether or not to be selected, the voice coder / decoder for encoding / decoding the voice data, and the voice recognition / synthesis mode is selected in the mode switching unit, In the transmission, the voice data is converted into text data and transmitted, the frame recognition rate is calculated from the voice recognition / synthesis unit and the result of the transmission of the previously transmitted voice data. And a frame loss ratio calculation and comparison unit for generating the control signal when the threshold is exceeded.

Hereinafter, a voice communication method using voice recognition / synthesis according to the present invention and a device thereof will be described with reference to the accompanying drawings.

2 is a diagram conceptually illustrating data communication between two systems to which the present invention is applied.

Referring to FIG. 2, for example, when data communication is performed between the first system 200 and the second system 210, the video / audio / text data packet is transmitted in the present invention, and additionally, a frame loss ratio (FLR) is added. ) Is sent. The frame loss rate (FLR) is determined by traffic on a data communication network, and text is transmitted in place of voice when the frame loss rate is larger than a predetermined threshold. That is, video and text data packets are generated according to the FLR instead of video and audio data packets.

3 is a block diagram illustrating a voice communication apparatus according to the present invention. In a system integrating a video and audio communication apparatus for video and voice communication, the video communication apparatus includes an image input / output device (300, 302), an image encoder. / Decoder (310,312) and multiplexer (MUX) / demultiplexer (DEMUX) 360, the voice communication device characterized in the present invention is a voice input / output device (320/322), the first and second modes Switching unit 330,332, speech coder / decoder 340,342, speech recognition module / Text To Speech (TTS) 350, word list 352, multiplexer (MUX) / demultiplexer (DEMUX) 360 and FLR calculation And a comparator 370. The system shown in FIG. 3 represents the system on one side shown in FIG. 2, compared to conventional devices in terms of voice communication devices.

Referring to FIG. 3, video and audio input to the video input device 300 and the audio input device 320 are respectively encoded through the video encoder 310 and the voice encoder 340, respectively. The MUX / DEMUX 360 generates a data stream in consideration of synchronization of two encoded signals and transmits the data stream to a target system through a predetermined communication network.

However, when the traffic is heavy, the voice input to the voice input device 320 is transmitted by the first mode switching unit 330 to the voice recognition module / TTS 350 instead of the voice encoder 340. The first mode switching unit 330 controls whether the input voice is processed in a normal voice mode or a voice recognition mode for text conversion by the FLR calculator and comparator 370. The FLR calculation and comparison unit 370 calculates the FLR according to the previous data transmission result transmitted through the MUX / DEMUX 360 and generates a control signal for converting to the voice recognition mode when the calculated value exceeds a predetermined threshold. do. In response to the control signal generated from the FLR calculation and comparison unit 370, the first mode switching unit 330 at the transmitting side and the second mode switching unit 332 at the receiving side are sent to the voice recognition module / TTS 350. Switch the transmission path.

When voice communication is performed using the voice recognition module / TTS 350, the receiving side converts the input voice, that is, a word found in the word list 352 among the words spoken by the user, into text and transmits the converted word. In general, transmission in this manner results in less data than the amount of data compressed by the audio signal. The reason for referring to the word list 352 is that it is impossible to guarantee 100% speech recognition rate. Thus, the user adds a suitable number of phrases to the word list 352 so that they can be transmitted. On the other hand, at the receiving side, the text received in place of the voice data stream is converted from text to voice using voice synthesis in the voice recognition module / TTS 350 selected by the second mode switching unit 322, or is heard by the user. Or show it to the user as text itself. In this case, the receiving side receives a small amount of data, and in particular, when receiving the information in text, the receiving party correctly recognizes the information sent.

4 (a) and 4 (c) are diagrams for explaining a transmission data structure used in the present invention, and FIG. 4 (a) shows a video / audio packet structure, and FIG. 4 (b) shows a video / text packet structure. 4 (c) shows the details of the header, respectively.

Referring to FIG. 4, in general, when the voice mode is selected by the first and second mode switching units 330 and 332, the data stream transmitted through the MUX / DEMUX 360 may be as shown in FIG. 4A. It has a video / audio packet structure. Meanwhile, when the traffic is heavy and the first and second mode switching units 330 and 332 select the voice recognition / TTS mode, the data stream transmitted through the MUX / DEMUX 360 is as shown in FIG. 4 (b). It has a video / text packet structure. The data packet includes a header before the frame to represent information related to the frames transmitted. In the data packet according to the present invention, the header includes at least a mode flag, a frame number, and an FLR indicating whether the video / audio mode or the video / audio recognition / TTS mode is used.

5 is a flowchart for explaining a voice communication method according to the present invention.

A voice communication method between the two systems shown in FIG. 3 will be described in detail with reference to FIGS. 4 and 5. First, in the environment in which traffic is not checked, packet 1 in which the transmission frame number is specified in the packet header is transmitted from the first system to the second system (step 500). At this time, the transmission data structure has a video / audio packet structure as shown in Fig. 4A. The second system receives the packet 1 (step 502). Thereafter, the frame number specified in the packet header of the received packet 1 is checked to calculate the FLR (step 504). That is, the second system on the receiving side calculates the FLR by comparing the frame number included in the header of the received data packet with the frame number to be actually received.

After the FLR is calculated, when the packet is transmitted from the second system to the first system, packet 2 in which the FLR is specified in the packet header is transmitted (step 506). The first system receives the packet 2 (step 508). The first system on the transmitting side compares the FLR included in the header of the data packet with a predetermined threshold, that is, a loss ratio threshold, to determine whether the threshold is exceeded (step 510). When the threshold is exceeded, the structure of the data transmitted after switching to the voice recognition / TTS mode is a video / text packet structure as shown in FIG. 4 (b). Continue switching.

As described above, the voice communication method and apparatus using voice recognition / synthesis according to the present invention use voice recognition / voice synthesis in case of heavy traffic, so that the user can recognize the voice signal at the receiving side. This has the advantage.

Claims (3)

In the case of heavy traffic, a voice communication method using voice recognition / synthesis, Transmitting, at the transmitting side, a voice data packet having a header, at least in which transmission frame numbers are specified, to the data communication network; Calculating a frame loss ratio by comparing the transmission frame number of the voice data packet received at the receiving side with a predetermined transmission frame number; Transmitting a header specifying the calculated frame loss rate back to the transmitting side; And And when the frame loss ratio exceeds a predetermined threshold, transmitting a text data packet using voice recognition / synthesis on behalf of the voice data packet. In the case of heavy traffic, a voice communication device using voice recognition / synthesis, A transmission side / reception side mode switching unit for selecting whether to process the voice data to be transmitted in a voice mode or a voice recognition / synthesis mode in response to a control signal; A speech encoder / decoder for encoding / decoding the speech data when the speech mode is selected by the mode switching unit; A voice recognition / synthesis unit for converting the voice data into text data when transmitting and transmitting the text data and vice versa when the mode switching unit selects the voice recognition / synthesis mode; And And a frame loss ratio calculation and comparison unit for calculating a frame loss ratio from a transmission result of previously transmitted voice data and generating the control signal when the calculated value exceeds a predetermined threshold. The method of claim 2, wherein the speech recognition / synthesis unit, And a word found in the word list among the phrases of the voice data to be transmitted is converted into a text using a word list trained on a predetermined phrase that can be voice recognized.