JP7093025B2 - Computer system and how to generate audio information - Google Patents

Description

The present disclosure relates to computer systems and methods of generating audio information.

When distributing a gameplay video, live audio information of a live person may be provided (see, for example, Patent Document 1). In the example of Patent Document 1, a game play video including live audio information is distributed according to the operation of a live player who has lived the game.

Japanese Unexamined Patent Publication No. 2020-044086

By outputting audio such as live commentary to the game, not only the player but also the viewer of the game play video and the spectator of e-sports (Electric Sports) can feel the game more interesting. ..

It is an object of the present disclosure to enable a computer system to output audio such as a live commentary along the flow of conversation such as a game situation or a live commentary.

The first aspect is an input unit that accepts voice input from an information sender for a game played via a computer, a status collection unit that collects status data regarding the status of the game, voice input to the input unit, and an input unit. It is a computer system characterized in that it includes a voice information generation unit having a function of generating voice information according to the situation data collected by the situation collection unit, and an output unit for outputting the voice information. ..

Further, in the first aspect, the voice information generation unit is configured to function based on a trained model of machine learning, and a plurality of types of the trained model are prepared, and the computer system is a computer system. A switching unit for switching the learned model used by the voice information generation unit may be provided.

Further, in the above aspect, a user interface control unit for designating the learned data used by the voice information generation unit may be provided.

Further, in the above aspect, the voice information generation unit has a function of generating voice information according to the situation data, and the output unit has the situation data when the voice input has not been performed for a predetermined period. Audio information may be output according to the above.

Further, in the above aspect, the voice information generation unit has a function of generating voice information according to the situation data, and the output unit has the situation when the situation data indicates a predetermined state. Audio information according to the data may be output.

Further, in the above aspect, when the situation data indicates a predetermined state, the output unit outputs voice information according to the situation data while the input unit is accepting the voice input. You may output it.

Further, in the above aspect, an information display unit for displaying information about the game may be provided.

Further, in the above aspect, an evaluation unit that evaluates the voice input may be provided based on at least one of the voice information and the situation data collected by the situation collection unit.

The second aspect is an input step for receiving voice input of an information sender for a game played via a computer by a computer system, a situation collection step for collecting status data regarding the status of the game, and the input. Voice information characterized by executing a voice information generation step for generating voice information and an output step for outputting the voice information according to the voice input in the step and the situation data collected in the situation collection step. It is a generation method of.

According to the present invention, the computer system makes it possible to output audio such as a live commentary along the flow of conversation such as a game situation or a live commentary.

It is a block diagram which shows the structure of a game system. It is a block diagram which shows the structure of a server apparatus. It is a block diagram which shows the structure of a game apparatus. It is a block diagram which shows the structure of an information originator terminal. An example of a game image is shown. An example of a game image and a user interface on an information sender terminal is shown.

Hereinafter, the computer system and the method of generating voice information in the present disclosure will be described with reference to the drawings. The computer system of the present disclosure is realized as a game system. The method of generating voice information in the present disclosure is implemented as a function of a game system.

In the following description, the "game" means an activity to achieve some purpose according to a predetermined rule and is carried out by using a computer. In some of the games, a plurality of participants compete with each other (which determines the outcome), and in some cases, each participant is evaluated based on some index such as score and time. Participants here include both individuals and teams.

[Embodiment]
<< Overview of the game system >>
FIG. 1 is a block diagram showing a configuration of a game system 1 according to an embodiment. The game system 1 is used, for example, to run an e-sports competition. In the game in the game system 1, the player character is activated in a three-dimensional virtual game space (hereinafter, simply referred to as a virtual space) in response to an operation of a game player (participant), and various types of player characters are used. Let them take action.

The game described in this embodiment is a battle-type action game in which a player character is played against another player character or a non-player character in a virtual space. More specifically, the game of the present embodiment is a so-called fighting game.

In the fighting game of the present embodiment, two game players play a one-on-one battle. In the battle, each game player selects and operates one character (player character). In a match, for example, the victory or defeat is determined when the physical strength value of any of the characters becomes 0.

In the e-sports tournament of the present embodiment, the tournament progresses by, for example, a tournament method. In this e-sports tournament, the situation of the competition will be played by a commentator. In addition, in this e-sports competition, the content of the battle (for example, the technique used) will be explained by the commentator. The live commentary and commentary will be distributed to spectators via communication networks. The live commentary and commentary may be broadcast by speakers installed at the venue.

In the following, a person who transmits information related to a game as voice information is referred to as an "information sender". "Voice information" is information related to at least one of the game commentary and the game commentary. The commentator and the commentator are examples of "information senders", respectively.

《Hardware configuration》
<Overview>
The game system 1 includes a server device 2, a game device 5, and an information sender terminal 10. The server device 2 is configured to be able to communicate with the game device 5 via a communication network 6 (for example, the Internet).

The information sender terminal 10 is a terminal used by an information sender. The information sender terminal 10 is connected to the communication network 6. The information sender terminal 10 and the server device 2 communicate with each other via the communication network 6.

The server device 2 manages the e-sports tournament and supports information transmission. The server device 2 stores a game program and game data. The server device 2 manages the game data of the game device 5. The server device 2 stores various programs and data for managing the e-sports competition.

The game device 5 executes the game based on the operation of the game player. In the present embodiment, a competitive action game is played in the game device 5. In the present embodiment, one game device 5 accepts the operation of a set of players (two players in this example).

As the game device 5, for example, a commercially available device such as a personal computer, PlayStation (registered trademark), XBox (registered trademark), PlayStation Vita (registered trademark), Nintendo Switch (registered trademark) can be used.

The game device 5 may receive (download) a game program and game data from the server device 2 via the communication network 6. When the game device 5 receives the game program or the like, the game device 5 installs the received game program or the like.

The game device 5 starts a battle by a battle-type action game based on the operation of each player (hereinafter, players P1 and P2). The game device 5 advances the game while outputting game images and sounds to the display 61 and the speaker 62 (each described later) based on the operation of each player.

When the battle is completed, the game device 5 transmits the result information indicating the battle result to the server device 2. The result information includes, for example, the stage of the match (information on the first round, the final match, etc.) and information for identifying the winner in the match.

<Configuration of server device 2>
FIG. 2 is a block diagram showing the configuration of the server device 2. As shown in FIG. 2, the server device 2 includes a network interface 21, a storage unit 22, and a control unit 23. The network interface 21 and the storage unit 22 are electrically connected to the control unit 23 via the bus 29.

The network interface 21 is communicably connected to the game device 5, the information sender terminal 10, and the like via a communication network 6 (communication network) such as the Internet and a LAN.

The storage unit 22 is composed of an HDD, SSD, RAM, ROM, and the like. The storage unit 22 stores game data and various programs including a part of the game program according to the present embodiment.

The control unit 23 controls the operation of the server device 2. The control unit 23 includes a CPU (microcomputer) and a semiconductor memory (not shown). The semiconductor memory stores a program for operating the CPU and the like.

<Structure of game device 5>
FIG. 3 is a block diagram showing the configuration of the game device 5. As shown in FIG. 3, the game device 5 has a display 61, a speaker 62, and a controller 63 externally connected or built-in. Two displays 61 are provided. Further, the game device 5 is also provided with two controllers 63. In the following, a plurality of components such as the display 61 are distinguished by adding a branch number to a reference code as necessary (for example, 61-1, 61-2, etc., and the same applies hereinafter).

The display 61-1 and the controller 63-1 are used by the player P1. The display 61-2 and the controller 63-2 are used by the player P2. The controller 63 may be connected to one prepared by the players P1 and P2 itself instead of the one provided as standard in the game system 1.

The game device 5 has a network interface 51, a graphic processing unit 52, an audio processing unit 53, an operation unit 54, a storage unit 55, and a control unit 56. The network interface 51, the graphic processing unit 52, the audio processing unit 53, the operation unit 54, and the storage unit 55 are electrically connected to the control unit 56 via the bus 59.

The network interface 51 is communicably connected to the communication network 6 in order to transmit and receive various data between the game device 5 and the server device 2.

The graphic processing unit 52 draws a game image including various objects related to the virtual space in a moving image format according to the game image information output from the control unit 56. The graphic processing unit 52 is connected to, for example, a liquid crystal type display 61 (61-1, 61-2). The game image drawn in the moving image format is displayed on the display 61 as a game screen.

The audio processing unit 53 is connected to the speaker 62. The audio processing unit 53 reproduces and synthesizes the game sound according to the instruction of the control unit 56, and outputs the game sound from the speaker 62.

The operation unit 54 is connected to the controller 63 (63-1, 63-2). The operation unit 54 transmits / receives data related to the operation input to / from the controller 63. The players P1 and P2 input operation signals to the game device 5 by operating various controls such as buttons on the controller 63.

The storage unit 55 is composed of an HDD, SSD, RAM, ROM, and the like. The storage unit 55 stores game data downloaded from the server device 2, various programs including a part of the game program, and the like.

The control unit 56 controls the operation of the game device 5. The control unit 56 includes a CPU (microcomputer) and a semiconductor memory (not shown). The semiconductor memory stores a program for operating the CPU and the like.

<Configuration of information sender terminal 10>
The information sender terminal 10 can be configured, for example, by installing predetermined software on a personal computer. FIG. 4 is a block diagram showing the configuration of the information sender terminal 10. As shown in FIG. 4, a display 91, a speaker 92, a microphone 94, and a keyboard 93 are connected to the information sender terminal 10.

The information sender terminal 10 has a network interface 81, a graphic processing unit 82, an audio processing unit 83, an operation unit 84, a storage unit 85, and a control unit 86. The network interface 81, the graphic processing unit 82, the audio processing unit 83, the operation unit 84, and the storage unit 85 are electrically connected to the control unit 86 via the bus 87.

The network interface 81 is communicably connected to the communication network 6 in order to transmit and receive various data. As a result, the information sender terminal 10 can communicate with the server device 2.

The graphic processing unit 82 is connected to the display 91. The graphic processing unit 82 displays various images on the display 91 according to the image information output from the control unit 86.

The audio processing unit 83 is connected to the speaker 92. The audio processing unit 83 outputs (reproduces) the voice from the speaker 92 according to the voice information output from the control unit 86.

A microphone 94 is connected to the audio processing unit 83. The microphone 94 is used to collect the voice of the information sender as an electric signal. The audio processing unit 83 converts the signal input from the microphone 94 into a digital signal (audio signal). The audio processing unit 83 transmits an audio signal to the control unit 86.

The operation unit 84 is connected to the keyboard 93. An operation signal is input to the operation unit 84 in response to an operation of the keyboard 93 by a user (for example, an information sender). For example, the user can execute various programs stored in the storage unit 85 by operating the keyboard 93. The user can input various information by operating the keyboard 93.

The storage unit 85 is composed of an HDD, SSD, RAM, ROM, and the like. Various programs and the like are stored in the storage unit 85.

The control unit 86 controls the information sender terminal 10. The control unit 86 includes a CPU (microcomputer) and a semiconductor memory (not shown). The semiconductor memory stores a program for operating the CPU and the like.

<< Functional configuration of the control unit 56 in the game device 5 >>
The control unit 56 functions as a battle execution unit 561 and a transmission unit 562 by the CPU executing various programs.

The battle execution unit 561 reads data such as objects and textures included in the game data and executes the game program according to the operation of the controller 63 by the player. By executing the game program, two-dimensional or three-dimensional game image information is generated.

The game image information is processed by the graphic processing unit 52. By the processing of the graphic processing unit 52, the processed game image is displayed on the display 61. For example, the game image Sc as shown in FIG. 5 is displayed on the display 61. The game image Sc is also displayed on the audience display installed at the venue. The game image Sc is also distributed to the spectators via the communication network 6. Voice input from the information sender (described later) is also distributed to the spectators via the communication network 6.

The battle execution unit 561 starts a battle by a battle-type action game when a predetermined condition is satisfied (for example, when a predetermined operation is performed on the controller 63). The battle execution unit 561 ends the battle when the victory or defeat of the battle is decided (for example, when the physical strength value of the character becomes 0).

The transmission unit 562 transmits information (character name, costume used, etc.) of the character selected by each opponent. This information can be used for live commentary and commentary.

The transmission unit 562 transmits the history information of the input by the controller 63 (hereinafter referred to as input history information) to the server device 2 via the network interface 51 during the battle. By analyzing the input history information, it is possible to identify the technique performed by the players P1 and P2.

The transmission unit 562 transmits the status information in the game to the server device 2 during the battle. As the status information, the physical strength value of each character can be exemplified.

<< Functional configuration of the control unit 23 in the server device 2 >>
The control unit 23 has an input unit 231, a status collection unit 232, a display unit 233, a user interface control unit 234, an information generation unit 235, an output unit 236, a switching unit 237, and an evaluation unit when the CPU executes various programs. It functions as a unit 238.

The input unit 231 receives a voice input (digital data) from the information sender. More specifically, the input unit 231 receives the voice input transmitted from the information sender terminal 10. The input unit 231 transmits the received voice input to the information generation unit 235.

The situation collecting unit 232 collects information about the situation of the game (hereinafter referred to as situation data). For example, the status collection unit 232 collects input history information, status information, and character information as status data. The status collection unit 232 transmits the collected status data to the information generation unit 235.

The display unit 233 outputs the game image Sc to the information sender terminal 10. This allows the information sender to see the match. That is, the information sender can grasp the situation of the battle not only when the battle is held at the battle venue but also when the battles are played between the opponents via the communication network 6.

The user interface control unit 234 (information display unit) causes the information sender terminal 10 (display 91) to display the user interface UI via the network interface 21. The user interface control unit 234 displays the user interface UI on the information sender terminal 10 by superimposing it on the game image Sc, or displays the user interface UI on the information sender terminal 10 independently.

For example, the user interface control unit 234 causes the information sender terminal 10 to display an area for displaying text information (hereinafter referred to as a text area) as a user interface UI. Further, the user interface control unit 234 causes the information sender terminal 10 to display a selection screen for switching the learned model (described later) as the user interface UI.

The information generation unit 235 generates audio information as a commentator or audio information as a commentator. In this embodiment, the information generation unit 235 is implemented by AI (artificial intelligence). The information generation unit 235 functions based on the trained model of machine learning. The information generation unit 235 determines whether to generate the voice information as a commentator or the voice information as a commentator according to the contents of the trained model.

The information generation unit 235 is configured by using a neural network. This neural network (information generation unit 235) uses voice input to the input unit 231 and situation data collected by the situation collection unit 232 as inputs. The voice input can be easily input to the information generation unit 235 by converting it into text data by the voice recognition technique.

The neural network (information generation unit 235) has a function (first function) of generating voice information corresponding to both voice input and input of situation data. The information generation unit 235 also has a function (second function) of generating voice information according to the situation data. That is, the information generation unit 235 can also generate voice information without using voice input. Further, the information generation unit 235 has a function (third function) of generating voice information by using only voice input.

In this embodiment, a plurality of types of trained models are prepared. The trained model can be built, for example, using commentary and commentary made in the game in the past. AI with various characteristics (personality) can be constructed by appropriately selecting the voice information used in machine learning.

For example, letting AI perform machine learning using an input consisting of the voice (speech) of a commentator (person) and situation data, and the voice (output) of a commentator (person) corresponding to these inputs as teacher data. Then, the information generation unit 235 can be made to perform an action (speech) as a commentator. Similarly, let AI perform machine learning using the input consisting of the voice (speech) of the commentator (person) and the situation data, and the voice (output) of the commentator (person) corresponding to these inputs as teacher data. As a result, the information generation unit 235 can be made to perform an operation (speech) as a commentator.

Further, by causing AI to perform machine learning using the voice of a female information sender, the information generation unit 235 can generate voice information as a woman.

Further, the AI (information generation unit 235) may perform machine learning so as to behave as a guest. Guests are free to speak, sometimes with commentary. The information generation unit 235 (hereinafter referred to as guest AI) that behaves as a guest can be constructed by, for example, performing machine learning using the following teacher data.

For example, the teacher data includes guest remarks that disturb the information sender. Specifically, insert a horizontal spear by interfering with information senders (people such as commentators) such as "○○ like!", "Chance!", "Dangerous!", "Good" by guests. It is conceivable to include the remarks in the teacher data. Remarks that disturb and insert a horizontal spear have nothing to do with playing the guest's voice during the voice input of the information sender, talking to the information sender, the progress of the game, the voice input of the information sender, or the ups and downs of emotions. It also includes playing various voices.

The teacher data may include statements made by the guest that are different from the emotional ups and downs of the information sender. Specifically, in a soccer game, after the voice input (statement of the information sender) is "Goal !!!", "Don't let your guard down (in a tone that is not happy at all). It is conceivable to include guest remarks such as "Let's tighten up" in the teacher data.

In addition, the teacher data may include statements made by guests that are not true. For example, if something close to a foul is done in the game (not really a foul), the teacher data could include an angry remark such as "That's a foul! It's a red card!".

The spectator can feel more interesting when the commentator (person) or the like responds well to the remarks of the guest AI (information generation unit 235). AI's remarks are a kind of issue for information senders. The information sender inputs voice in consideration of the response to the AI's remarks (issues). Guest AI's remarks are also a kind of issue for information senders. The method of responding to the problem of the information sender can be used for the evaluation item by the evaluation unit 238. Specifically, it can be used as an evaluation item (described later) of an information sender by a spectator (viewer).

The information generation unit 235 generates information about the game as text information (information expressed in characters). In the present embodiment, the information generation unit 235 generates, as text information, the name of the most recently played technique, the progress of the match (success rate of the technique in this match, etc.), the profile of the opponent, and the like. Further, the text information is not limited to the information displayed to the player, but includes information that is not displayed to the player during the game.

The profile of the opponent can be obtained by creating a database of profiles of various players. The latest technique name can be obtained by analyzing the input history information. The success rate of the technique or the success rate of the guard can be obtained, for example, by statistically processing the technique and the degree of change in the physical strength value during the battle. The method of judging the success or failure of the technique and the method of judging the success or failure of the guard will be illustrated later.

In addition, in the text information, the characters are facing each other and glaring at each other, the gap after the technique or technique is activated has been canceled, the facing position (direction of the character) has changed, and advanced attacks are performed. The state of the character, such as that, may be included. It is desirable to generate this textual information using technical terms (eg, terms specific to the game being played).

In order to generate text information including technical terms, it is conceivable to configure a database and an information generation unit 235 so that the technical terms can be extracted from the database in the storage unit 22. Further, the information generation unit 235 may be made to perform machine learning so that technical terms can be output.

The information generation unit 235 transmits the generated text information to the user interface control unit 234. The user interface control unit 234 displays the received text information in the text area. The information generation unit 235 updates the text information as appropriate.

The information generation unit 235 ends the operation when an end instruction is input via the user interface UI. As a result, the information sender can end the transmission of information such as the actual situation.

The output unit 236 distributes the voice input from the information sender via the communication network 6. The output unit 236 also transmits the game voice to the information sender terminal 10 via the network interface 21. Further, the output unit 236 outputs the voice information generated by the information generation unit 235 to the information sender terminal 10 at a predetermined timing (described later) via the network interface 21. The output unit 236 also distributes the voice information generated by the information generation unit 235 via the network interface 21.

Specifically, the output unit 236 selects the voice information to be output according to the presence or absence of voice input in a predetermined period (described later) after the situation data has changed a predetermined number of times (one or a plurality of times). Here, as the "change in situation data", in the case of a fighting game, examples include cases where parameters such as a physical strength gauge and a special move gauge are changed, and cases where a special attack such as a special move is executed. can.

For example, the success or failure of a continuous technique (so-called combo technique) can be determined by referring to the input key information (input history information) and the input frame (information on whether or not the input corresponding to the technique was performed within a predetermined period). .. The output unit 236 can generate and output voice information depending on the success or failure of the technique.

For example, when the gauge of one character's special move becomes the maximum value and the special move is executed by the player's operation, and the physical strength gauge of the other character decreases, the special move (for example, continuous move) is hit. It becomes a situation. Also, when the gauge of one character's special move becomes the maximum value and the special move is executed by the player's operation, the other character executes the guard by the player's operation and the physical strength gauge does not decrease. It becomes a situation where the special move did not hit (a situation where the guard or evasion was successful). In addition, it is possible to compare the damage given by the special move or the like with the damage actually given (the degree of decrease in the physical strength gauge), and determine how much the special move has hit.

Further, the "change in situation data" includes the case where the output unit 236 outputs (speaks) a predetermined number of times. The content of this "statement" may be a question.

“No voice input” can be exemplified as a state in which the input unit 231 is not receiving a signal. "No voice input" includes a state in which the input unit 231 receives a signal but the signal cannot be recognized as a word. The "predetermined period" is a period in which the information sender is recognized to be silent. The "predetermined period" in this embodiment is a few seconds. The "predetermined period" may be determined by, for example, the number of frames of a moving image.

In the following, the "predetermined period after the status data has changed a predetermined number of times (one or more times)" is referred to as a "determination period". When there is a voice input during the determination period, the output unit 236 outputs voice information (voice information generated based on the first function) corresponding to both the voice input and the situation data.

On the other hand, when there is no voice input during the determination period, the output unit 236 outputs voice information (voice information generated based on the second function) according to the situation data. Further, the output unit 236 receives voice information according to the voice input (voice generated based on the third function) when there is a voice input even if there is no change in the situation data for a predetermined period. Information) is output.

When the status data indicates a specific state (hereinafter referred to as a special state), the output unit 236 has a function of outputting voice information (voice information generated based on the second function) according to the situation data. Have. In the special state, the output unit 236 may output voice information according to the situation data even when the input unit 231 is in a state of accepting voice input. The special state can be exemplified when the situation of the game is a so-called chance scene, a scoring scene, or the like.

The above functions of the output unit 236 can be realized by AI. The AI that realizes the output unit 236 may be the same as the AI that realizes the information generation unit 235, or it may be a separate AI. When the output unit 236 is implemented by AI, it may be trained every time the game is repeated and the definition of the special state may be updated. Of course, which state is regarded as a special state may be set in advance (for example, information is stored in a predetermined variable).

The switching unit 237 switches the trained model used by the information generation unit 235. In the present embodiment, the user of the information sender terminal 10 (for example, the information sender) operates the user interface UI, so that the request for switching the learned model is transmitted to the switching unit 237. As a result, the user of the information sender terminal 10 can select the character (characteristic) of the AI.

The switching unit 237 notifies the information generation unit 235 of the trained model to be used in response to the switching request. The information generation unit 235 uses the trained model notified at the time of operation.

The evaluation unit 238 evaluates the voice input (that is, the statement of the information sender) based on at least one of the voice information and the situation data. Here, the evaluation unit 238 compares the voice information converted into the text data and the situation data, and evaluates the voice input using the degree of conformity between the two as an index.

The evaluation unit 238 also uses the evaluation by the spectator (viewer) for the evaluation of the audio information. For evaluation from spectators, evaluation via SNS (social networking service) etc. can be used. Further, for the evaluation by the spectator, it is conceivable to use the evaluation input from the UI provided in the terminal device for viewing. The terminal device for viewing can be configured by a personal computer or a game device.

<< Functional configuration of the control unit 86 in the information sender terminal 10 >>
The control unit 86 functions as a voice transmission unit 861 and a UI processing unit 862 when the CPU executes various programs.

The voice transmission unit 861 transmits the voice signal generated by the audio processing unit 83 to the server device 2 via the network interface 81. The audio signal here is input by the information sender to the audio processing unit 83 using the microphone 94. The voice signal transmitted from the voice transmission unit 861 is received by the input unit 231 of the server device 2.

The UI processing unit 862 displays the user interface UI on the display 91 in response to an instruction from the user interface control unit 234 of the server device 2. The UI processing unit 862 displays the game image Sc output by the display unit 233 on the display 91. FIG. 6 shows an example of a screen displayed by the UI processing unit 862. The UI processing unit 862 transmits the operation result of the user to the user interface UI to the server device 2.

<< Operation example >>
In order to transmit voice information, the information sender accesses the game system 1 (specifically, the server device 2) from the information sender terminal 10 and logs in. As a result, the user interface control unit 234 of the server device 2 functions, and the user interface UI is displayed on the information sender terminal 10.

By operating the user interface UI, the information sender can select an AI that functions as a commentator or an AI that functions as a commentator. Here, it is assumed that the information sender selects the AI that functions as a commentator. In other words, the information sender shall speak as a commentator.

When the preparation is completed in the information sender terminal 10, the information generation unit 235 functions. For example, the information generation unit 235 transmits text information such as a player's profile to the user interface control unit 234 before the start of the match. As a result, the user interface control unit 234 causes the received information to be displayed in the text area (display 91). The information sender may start the explanation or the like using the information displayed in the text area, or may prepare the content to be spoken during the match based on the displayed information.

During the battle, the display unit 233 displays the state of the battle as a game image Sc on the display 91. The information sender can see the game image Sc and give an explanation. The voice emitted by the information sender is transmitted from the information sender terminal 10 to the input unit 231 of the server device 2. The voice input to the input unit 231 is sent to the information generation unit 235 and the evaluation unit 238. The evaluation unit 238 evaluates the voice input as appropriate.

During the battle, the situation collecting unit 232 collects the input history information, the state data such as the state of the game (parameters such as the physical strength gauge and the special move gauge, etc.) and sends it to the information generation unit 235. The information generation unit 235 generates voice information as a commentator based on at least one of the voice input and the situation data. In other words, the information generation unit 235 appropriately exerts the first to third functions. The information generation unit 235 sends the generated voice information to the output unit 236.

For example, the output unit 236 outputs the voice information generated based on the first function when there is a voice input during the determination period. Further, the output unit 236 outputs the voice information generated based on the second function when there is no voice input during the determination period and in the case of a special state. In a special state, the output unit 236 may output while accepting voice input. That is, the output unit 236 may interrupt the remark of the information sender and output (speak).

If there is information that can be displayed in the text area, the information generation unit 235 transmits the information to the user interface control unit 234. For example, the information generation unit 235 displays the latest technique name or the like in the text area. The information sender can use the information displayed in the text area when giving an explanation.

The information sender may continue to send information such as commentary even after the match is over. The information generation unit 235 generates voice information in response to voice input even after the battle is over. When the information sender wants to end the information transmission, he / she can notify the server device 2 of the end of the information transmission training via the user interface UI. When the information transmission of the information sender is completed, the evaluation unit 238 stores the evaluation result of the voice information in the database in the storage unit 22.

Summarizing the above, the present disclosure discloses an input unit 231 that accepts voice input from an information sender for a game played via a game device 5 (computer), and a status collection unit 232 that collects status data related to the status of the game. And the voice information generation unit 235 having a function of generating voice information according to the voice input to the input unit 231 and the situation data collected by the situation collection unit 232, and the output unit 236 to output the voice information. It is a game system 1 (computer system) characterized by having.

<< Effect of this embodiment >>
As described above, in the present embodiment, the computer system makes it possible to output audio such as a live commentary along the flow of conversation such as a game situation or a live commentary.

Further, the game system 1 can be used for the practice of the commentator and the commentator. That is, according to the present embodiment, it is possible to train information senders by using a computer system.

[Other embodiments]
The information generation unit 235 may be configured to generate audio information as a commentator. However, the audio information generated by the information generation unit 235 is not limited to the commentator and the commentator. For example, the information generation unit 235 (AI) may be configured to output audio information as a spectator watching a game together with a live spectator.

The type of game is not limited to the illustrated fighting game. The game system 1 (computer system) can be used for various game plays. For example, games are not limited to interpersonal battles. Further, the game may be a game in which a plurality of players cooperate with each other or play individually (for example, time attack).

The "change in situational data" described in the embodiments is merely an example. For example, if the game is a soccer game, changes such as a change in the team that owns (dominates) the ball or a ball approaching the goal is an example of "change in situation data". If the game is a baseball game, changes such as runners appearing on the base and changes in offense and defense are examples of "changes in situation data." Similarly, the "special state" is just an example.

The server used for generating the voice information may be provided separately from the game management server (server that manages the battle execution).

It is desirable to update the trained model when the character or technique is updated or the game rules are changed due to the version upgrade of the game.

The information generation unit 235 is configured so that the data of the battle can be collected and the tendency of each character (player's operation) (recently, information such as the combo of Y after X is popular) can be explained. May be good. In addition, the information generation unit 235 may be configured so that the changes due to the version upgrade of the game can be explained. As the changes to be explained, the character A can exemplify a change such as a weakened special move, a change such that the gap after a strong punch is shortened by X frames, and the like.

A parameter (hereinafter referred to as a trust parameter) indicating the trust relationship between the AI (information generation unit 235) and the user (information sender) may be provided. For example, when the value of the trust parameter is low, the output unit 236 is made to output voice by covering the user's story, and when the value of the trust parameter is high, the output unit 236 is spoken by the user. It can be applied to make the sound output quieter.

Further, when the value of the trust parameter is high, when the voice input and the voice output by the information sender happen to be suffered, it can be applied to give the remark to the information sender. In this case, the output unit 236 interrupts the voice being output (instead of the voice being output) and outputs the voice indicating that the opportunity to speak is given (for example, the voice output "Mr. A please"). May be good.

The trust parameter may be held in a predetermined variable or may be used as teacher data.

The server device 2 may be equipped with a function of interviewing a player character (or a player). Specifically, the server device 2 is provided with a function of transmitting the voice of the information sender to the game device 5. The game device 5 is equipped with a function of transmitting a statement from a player character (or a player) to the server device 2.

By doing this, for example, if the commentator (information sender) asks the player character (player is also acceptable) that "it was a good match", the player character (player is also acceptable) will answer "Thank you." It becomes possible to exchange such things. The spectator (viewer) can feel more interesting about the game.

The server device 2 may be configured so that it can be controlled by the voice input of the information sender. For example, it is conceivable to display (deliver) the replay video when the information sender inputs a voice to the server device 2 asking "Can you start the replay?". In addition, when the information sender (for example, a commentator) inputs "Can you show player A?" By voice input, the server device 2 controls the virtual camera or switches the image to be displayed. The A player character may be displayed as the game image Sc.

Depending on the game, the game image Sc to be distributed can be selected by operating a virtual camera. In addition, in online games (including e-sports), in addition to players and information senders, there are some behind the scenes such as cameramen and broadcasting stations (described later).

A cameraman is a person who operates a virtual camera. A broadcasting station is responsible for selecting, for example, a plurality of virtual cameras to be used for distribution. The role of a broadcasting station and the role of a cameraman are played by a person through a terminal device (for example, a personal computer, a game device, etc.).

The game system 1 may be configured so that the commentator and the back side can communicate (communicate). By doing this, it becomes possible to make the distribution content more interesting.

The information generation unit 235 may be configured to point out in real time when the information sender makes an inaccurate statement. Examples of inaccurate statements include misuse of words (misunderstanding of the meaning of words), mistakes in player names and team names, and misunderstandings of rules. As a method of pointing out, for example, it is conceivable to display the information sender terminal 10 to indicate that it has been misused, to display an evaluation such as "bad", or to display correct information. In addition, when communicating the evaluation result to the information sender, the content of the inaccurate remark may be shown.

The content of remarks and wise sayings (famous scenes) that have been evaluated above a certain level may be recorded in the database. The "Quotations" may be automatically extracted by a computer (for example, the server device 2) or may be extracted based on the evaluation of the spectator.

In SNS, a mechanism for a user (viewer) to indicate his / her intention (for example, a button with a label such as "like") is often provided. It is conceivable to extract remarks that are highly evaluated by spectators (for example, remarks when many buttons such as "like" are pressed and get excited) as "Quotations".

The game system 1 may be configured to give advice to the information sender. Specifically, it is conceivable to refer to the latest information transmission history and evaluation of the information sender and give advice. It is desirable that the user (play-by-play person) can arbitrarily switch the form of advice (timing, display method, etc.).

Even when these other embodiments are adopted, the effects of the present invention are exhibited. Further, it is also possible to appropriately combine this embodiment with other embodiments and other embodiments.

1 Game system (computer system)
5 Game device (computer)
231 Input unit 232 Status collection unit 233 Display unit 234 User interface control unit (information display unit)
235 Voice information generation unit 236 Output unit 237 Switching unit 238 Evaluation unit

Claims (4)

An input unit that accepts voice input from information senders for games played via a computer,
A situation collection unit that collects situation data related to the game situation,
A voice information generation unit having a function of generating voice information according to the voice input to the input unit and the situation data collected by the situation collection unit.
A voice information based on the voice input, an output unit that outputs the voice information according to the situation data, and an output unit.
A player who plays the game, a display unit that displays a game screen to the information sender, and a display unit.
Equipped with
The situation data is a computer system characterized in that it is information on a game situation progressed by the player. An input unit that accepts voice input from information senders for games played via a computer,
A situation collection unit that collects situation data related to the game situation,
A voice information generation unit having a function of generating voice information according to the voice input to the input unit and the situation data collected by the situation collection unit.
An output unit that outputs the audio information and
Equipped with
The voice information generator is configured to function based on a trained model of machine learning.
Multiple types of the trained model are prepared ,
A computer system including a switching unit for switching the trained model used by the voice information generation unit. An input unit that accepts voice input from information senders for games played via a computer,
A situation collection unit that collects situation data related to the game situation,
A voice information generation unit having a function of generating voice information according to the voice input to the input unit and the situation data collected by the situation collection unit.
An output unit that outputs the audio information and
Equipped with
The voice information generation unit has a function of generating voice information according to the situation data.
When the status data indicates a predetermined state, the output unit outputs audio information corresponding to the status data.
The output unit is characterized in that, when the status data indicates a predetermined state, the input unit outputs voice information according to the status data while the input unit is accepting the voice input. Computer system. Depending on the computer system
An input step that accepts voice input from an information sender for a game played via a computer,
A status collection step for collecting status data regarding the status of the game, and
A voice information generation step that generates voice information according to the voice input in the input step and the situation data collected in the situation collection step.
A voice information based on the voice input, an output step for outputting the voice information according to the situation data, and an output step.
A player playing the game, a display step for displaying the game screen to the information sender, and a display step.
And run
The situation data is a method of generating voice information, characterized in that the situation data is information on a game situation progressed by the player.

Images