Go to MPAI-MMC V2.5 AI Modules
Function
Ref. Model
I/O Data
SubAIMs
JSON MData
Profiles
Ref. Software
Conformance
Performance
1 Functions
The Voice Command Interface (MMC‑VCI) is a Composite AIM that enables a person to issue vocal commands to an application – first, to Audio Scene Management (CAE‑ASM) – and, in the future, to other Modules. It has no application-specific code: the application is known to it by its Application Context, the standing description of the application – the schema of its commands, the schema of its responses and the rest of what they do not say – which JSON Interface takes from its Access or receives on the Application Context Port. When the application supplies its Application Context on that Port it is used; otherwise JSON Interface takes it from its Access. The person says what is wanted; the AIM turns it into a command for the application, a User Command request (CAE‑UCM), and, when the application answers with a User Command response, says the answer aloud.
The speech is recognised as text; JSON Interface turns the text into the command, a User Command request that conforms to the schema of the commands of the application, knowing what it remembers of the state of the application; the application answers with a User Command response; JSON Interface turns the response into the Response Hint, a text that tells what the application did and with which values, or what it did not do and why, or why no command was formed; Entity Dialogue Processing composes the reply from the recognised text and the Response Hint; Text to Speech speaks it. Commands carry numbers: a request that lacks a number the application requires does not become a command, and the person is asked for it.
2 Reference Model
Figure 1 depicts the Reference Model of the Voice Command Interface (MMC‑VCI) AIM.
Figure 1 – Reference Model of the Voice Command Interface (MMC‑VCI) AIM
The operation of the Voice Command Interface (MMC‑VCI) AIM develops in the following way:
- The person speaks. The speech that carries commands is not the speech an application edits: they arrive on different Ports.
- The application receives the command and answers it with a response, which names the command it answers.
- The AIM provides the command for the application, the response to the person as speech and as text, and the text of what it heard. A response is said when it arrives: the application may take time.
3 I/O Data
Table 1 specifies the Input and Output Data of the Voice Command Interface (MMC‑VCI) AIM. A Port is addressed by its Data Type and its Port Number (#); where the metadata gives none, none is shown.
| Input | Description |
|---|---|
| Basic Speech Object (OSD-BSO) #1 | The speech of the person. Optional. |
| Selector (OSD-SEL) | The language of the speech. Optional. |
| User Command RS (CAE-UCM) #1 | The response of the application to a command: a User Command response, which carries the UserCommandID of the command it answers. Optional. |
| Application Context (MMC-APC) | The Application Context: the standing description of the application, when the application supplies it. Optional. |
| Output | Description |
| Basic Speech Object (OSD-BSO) #1 | The response to the person, spoken (Response Speech). Optional. |
| User Command RQ (CAE-UCM) #1 | The command for the application: a User Command request that conforms to the schema of the commands of the application. Optional. |
| Basic Text Object (OSD-BTO) #1 | What the person said, as text (Heard Text). Optional. |
| Basic Text Object (OSD-BTO) #2 | The response to the person, as text (Response Text). Optional. |
4 SubAIMs
4.1 Reference Model
Figure 2 depicts the Reference Model of the Voice Command Interface (MMC‑VCI) Composite AIM.
Figure 2 – Reference Model of the Voice Command Interface (MMC‑VCI) Composite AIM
4.2 Operation
The Voice Command Interface (MMC‑VCI) AIM receives the speech of the person and, with it, the language it is in. Automatic Speech Recognition turns the speech into text. JSON Interface turns the text into a command, a User Command request, using the Application Context of the application – the schema of the commands, the schema of the responses and the rest of it – which it receives on the Application Context Port or, when none arrives, takes from its Access, and what it remembers of the state of the application; a request that does not become a valid command gives no command and a Response Hint that says why. The application answers with a User Command response, which names the command it answers; JSON Interface turns it into the Response Hint. Entity Dialogue Processing, which holds the dialogue, says what is to be said from the recognised text and the Response Hint: an answer, a question when the request is not clear, what was done or why not. Text to Speech speaks it. Acquisition and delivery of speech are the User Agent’s, not SubAIMs.
4.3 Functions of SubAIMs
Table 2 provides the functions of the Voice Command Interface (MMC‑VCI) SubAIMs.
| SubAIM | Function |
|---|---|
| Automatic Speech Recognition | Recognises as text the speech of the person. |
| JSON Interface | Turns the text of the request into a command for the application – a User Command request that conforms to the schema of the commands of the application, which it takes from its Access – knowing the application from its Application Context, which it receives or takes from its Access; turns the response of the application into the Response Hint, a text. |
| Entity Dialogue Processing | Holds the dialogue: from the recognised text and the Response Hint it says what is to be said – an answer, a question when the request is not clear, what was done or why not. |
| Text-To-Speech | Speaks the text that Entity Dialogue Processing says. |
4.4 I/O Data of SubAIMs
Table 3 gives the Input and Output Data of the Voice Command Interface (MMC‑VCI) SubAIMs.
4.5 AIMs and JSON Metadata
Table 4 provides the links to the AIM specifications and JSON schemas. AIM1 indicates the Composite AIM and AIM2 its SubAIMs.
| AIM1 | AIM2 | Name | JSON |
|---|---|---|---|
| MMC‑VCI | Voice Command Interface | X | |
| MMC‑ASR | Automatic Speech Recognition | X | |
| MMC‑JSI | JSON Interface | X | |
| MMC‑EDP | Entity Dialogue Processing | X | |
| MMC‑TTS | Text-To-Speech | X |
5 JSON Metadata
https://schemas.mpai.community/MMC/V2.5/AIMs/VoiceCommandInterface.json
6 Profiles
No Profiles.
7 Reference Software
Under development.
8 Conformance Testing
Table 5 provides the Conformance Testing Method for the Voice Command Interface (MMC‑VCI) AIM. Conformance Testing of the individual SubAIMs is given by the individual AIM specifications.
If a schema contains references to other schemas, conformance of data for the primary schema implies that any data referencing a secondary schema shall also validate against the relevant schema, if present, and conform with the Qualifier, if present.
| Receives | Basic Speech Object (OSD-BSO) #1 | Shall validate against Basic Speech Object schema. Speech Data shall conform with Speech Qualifier. |
| Selector (OSD-SEL) | Shall validate against the Selector schema. | |
| User Command RS (CAE-UCM) #1 | Shall validate against the User Command schema and contain a UserCommandRS whose UserCommandID is that of the command answered. | |
| Application Context (MMC-APC) | Shall validate against the Application Context schema. | |
| Produces | Basic Speech Object (OSD-BSO) #1 | Shall validate against Basic Speech Object schema. |
| User Command RQ (CAE-UCM) #1 | Shall validate against the User Command schema, contain a UserCommandRQ, and conform to the schema of the commands of the application. | |
| Basic Text Object (OSD-BTO) #1 | Shall validate against Basic Text Object schema. | |
| Basic Text Object (OSD-BTO) #2 | Shall validate against Basic Text Object schema. |
9 Performance Assessment
Not part of this specification.
