Microsoft Azure Text-Independent API

openapi: 3.2.0
info:
  version: 2020-08-01-preview
  title: Microsoft Azure AccessControlClient AccessConnector Text Independent API
tags:
- name: Text-Independent
paths:
  /text-independent/profiles:
    post:
      description: Creates a new speaker profile with specified locale.
      operationId: microsoftAzureTextindependentCreateprofile
      summary: Microsoft Azure Create Profile
      x-ms-examples:
        Successful Query:
          $ref: ./examples/createProfile.json
      tags:
      - Text-Independent
      parameters:
      - $ref: '#/components/parameters/apiVersionParam'
      responses:
        '201':
          description: Speaker profile created successfully. GUID is returned to reference the created profile.
          headers:
            location:
              description: url location of new resource
              schema:
                type: string
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/TiProfileInfo'
        default:
          $ref: '#/components/responses/SpeakerErrorResponse'
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/LocaleInfo'
        description: 'Provide following detail info when creating a new profile.

          Fields  | Description

          ------- | ------------

          locale  | Locale for the language of this speaker profile. A complete supported locale list is here: <ul><li>**en-US (American English)**</li><li>**es-ES (Castilian Spanish)**</li><li>**fr-FR (Standard French)**</li><li>**zh-CN (Mandarin Chinese)**</li></ul>'
    get:
      x-ms-pageable:
        nextLinkName: nextLink
        itemName: value
      description: Retrieves a set of profiles.<br>Profiles are sorted alphabetically by ProfileId
      operationId: microsoftAzureTextindependentListprofiles
      summary: Microsoft Azure List Profile
      x-ms-examples:
        Successful Query:
          $ref: ./examples/listProfiles.json
      tags:
      - Text-Independent
      parameters:
      - $ref: '#/components/parameters/apiVersionParam'
      - in: query
        name: maxpagesize
        description: The number of profiles to return. Default is 100 and the maximum is 500
        schema:
          type: integer
          default: 100
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/TiProfileInfoList'
        default:
          $ref: '#/components/responses/SpeakerErrorResponse'
  /text-independent/profiles/{profileId}:
    parameters:
    - $ref: '#/components/parameters/profileIdParam_2'
    get:
      description: Retrieves a single profile by ID.
      operationId: microsoftAzureTextindependentGetprofile
      summary: Microsoft Azure Retrieve Single Profile
      x-ms-examples:
        Successful Query:
          $ref: ./examples/fetchProfile.json
      tags:
      - Text-Independent
      parameters:
      - $ref: '#/components/parameters/apiVersionParam'
      - $ref: '#/components/parameters/profileIdParam'
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/TiProfileInfo'
        default:
          $ref: '#/components/responses/SpeakerErrorResponse'
    delete:
      description: Deletes an existing profile.
      operationId: microsoftAzureTextindependentDeleteprofile
      summary: Microsoft Azure Delete Profile
      x-ms-examples:
        Successful Query:
          $ref: ./examples/deleteProfile.json
      tags:
      - Text-Independent
      parameters:
      - $ref: '#/components/parameters/apiVersionParam'
      - $ref: '#/components/parameters/profileIdParam'
      responses:
        '204':
          description: OK
        default:
          $ref: '#/components/responses/SpeakerErrorResponse'
  /text-independent/profiles/{profileId}:reset:
    post:
      description: Resets existing profile to its original creation state. The reset operation does the following:<br>* Updates enrollmentStatus to Enrolling.<br>* Updates lastUpdatedDateTime.<br>* Updates enrollmentsCount to 0.<br>* Updates enrollmentsLength to 0.<br>* Updates enrollmentsSpeechLength to 0.<br>* Updates remainingEnrollmentsSpeechLength to the required number.<br>* Removes all associated enrollments from storage.<br>* Removes chosen passphrase association.<br>* Resets value of modelVersion.
      operationId: microsoftAzureTextindependentResetprofile
      summary: Microsoft Azure Reset Profile
      x-ms-examples:
        Successful Query:
          $ref: ./examples/resetProfile.json
      tags:
      - Text-Independent
      parameters:
      - $ref: '#/components/parameters/apiVersionParam'
      - $ref: '#/components/parameters/profileIdParam'
      responses:
        '200':
          description: Speaker profile reset successfully. Profile is returned with reset values.
          headers:
            location:
              description: url location of the resource
              schema:
                type: string
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/TiProfileInfo'
        default:
          $ref: '#/components/responses/SpeakerErrorResponse'
  /text-independent/profiles/{profileId}/enrollments:
    parameters:
    - $ref: '#/components/parameters/profileIdParam_2'
    post:
      description: Adds an enrollment to existing profile.<br>The first enrollment must be a predefined activation phrase which can be listed using the /phrases/{locale} api.<br>If the minimum number of requested enrollment audios is reached, a voice print is created.<br>Any further enrollment will be used to improve the voice print.<br><br>Limitations:<br>* Minimum audio input length per request is **1 second**<br>* Maximum audio input length per request is **120 seconds**<br>* Minimum total effective speech length (excluding silence and other non-speech frames) for creating a voiceprint is **20 seconds**<br>  This limitation can be disabled by setting ignoreMinLength to **true**.<br><br>* Maximum total audio input length allowed for creating a voiceprint is **300 seconds**<br>* Minimum audio Signal-to-noise ratio (SNR) is **2dB**
      operationId: microsoftAzureTextindependentCreateenrollment
      summary: Microsoft Azure Enroll Profile
      x-ms-examples:
        Successful Query:
          $ref: ./examples/createEnrollment.json
      tags:
      - Text-Independent
      parameters:
      - $ref: '#/components/parameters/apiVersionParam'
      - $ref: '#/components/parameters/profileIdParam'
      - in: query
        name: ignoreMinLength
        description: If true, a voice print will be created immediately for this profile regardless of how much speech is supplied or stored. Default is false.
        schema:
          type: boolean
          default: false
      responses:
        '201':
          description: Created
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/TiEnrollmentInfo'
        default:
          $ref: '#/components/responses/SpeakerErrorResponse'
      requestBody:
        content:
          audio/wav; codecs=audio/pcm:
            schema:
              type: object
              format: file
        description: Binary audio file. Supported formats are audio/wav; codecs=audio/pcm. Supports audio up to 5MB.
        required: true
  /text-independent/profiles:identifySingleSpeaker:
    post:
      description: Identifies who is speaking in input audio among a list of candidate profiles.<br><br>Limitations:<br>* Minimum audio input length is **1 second**<br>* Maximum audio input length is **120 seconds**<br>* Minimum candidate speakers count is **1**<br>* Maximum candidate speakers count is **50**<br>* Minimum effective speech length (excluding silence and other non-speech frames) is **4 seconds**<br>  This limitation can be disabled by setting "ignoreMinLength" to **true**.<br><br>* Minimum audio Signal-to-noise ratio (SNR) is **2dB**
      operationId: microsoftAzureTextindependentIdentifysinglespeaker
      summary: Microsoft Azure Identify Single Speaker Profile
      x-ms-examples:
        Successful Query:
          $ref: ./examples/identifySingleSpeaker.json
      tags:
      - Text-Independent
      parameters:
      - $ref: '#/components/parameters/apiVersionParam'
      - in: query
        name: profileIds
        required: true
        description: Comma-delimited profile IDs. Maximum supported number is 50 IDs.
        schema:
          type: array
          items:
            type: string
            format: uuid
          maxItems: 50
          minItems: 1
      - in: query
        name: ignoreMinLength
        description: If true, the minimum amount of speech needed for identification is skipped. Default is false.
        schema:
          type: boolean
          default: false
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/IdentifiedSingleSpeakerInfo'
        default:
          $ref: '#/components/responses/SpeakerErrorResponse'
      requestBody:
        content:
          audio/wav; codecs=audio/pcm:
            schema:
              type: object
              format: file
        description: Binary audio file. Supported formats are audio/wav; codecs=audio/pcm. Supports audio up to 5MB.
        required: true
  /text-independent/phrases/{locale}:
    get:
      description: Retrieves list of supported passphrases for a specific locale.
      operationId: microsoftAzureTextindependentListactivationphrases
      summary: Microsoft Azure Activation Phrases
      x-ms-examples:
        Successful Query:
          $ref: ./examples/listActivationPhrases.json
      tags:
      - Text-Independent
      parameters:
      - $ref: '#/components/parameters/apiVersionParam'
      - $ref: '#/components/parameters/localeParam'
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ActivationPhraseInfo'
        default:
          $ref: '#/components/responses/SpeakerErrorResponse'
  /text-independent/profiles/{profileId}/reset:
    parameters:
    - $ref: '#/components/parameters/profileIdParam_2'
    post:
      description: Resets existing profile to its original creation state. The reset operation does the following:<br>* Updates enrollmentStatus to Enrolling.<br>* Updates lastUpdatedDateTime.<br>* Updates enrollmentsCount to 0.<br>* Updates enrollmentsLength to 0.<br>* Updates enrollmentsSpeechLength to 0.<br>* Updates remainingEnrollmentsSpeechLength to the required number.<br>* Removes all associated enrollments from storage.<br>* Removes chosen passphrase association.<br>* Resets value of modelVersion.
      operationId: microsoftAzureTextindependentResetprofile
      summary: Microsoft Azure Reset Profile
      x-ms-examples:
        Successful Query:
          $ref: ./examples/resetProfile.json
      tags:
      - Text-Independent
      responses:
        '204':
          description: OK
        '401':
          $ref: '#/components/responses/Unauthorized'
        '404':
          $ref: '#/components/responses/NotFound'
        '429':
          $ref: '#/components/responses/RateLimit'
        '500':
          $ref: '#/components/responses/ServerError'
  /text-independent/profiles/{profileId}/verify:
    parameters:
    - $ref: '#/components/parameters/profileIdParam_2'
    post:
      description: Verifies existing profiles against input audio.<br><br>Limitations:<br>* Minimum audio input length is **1 second**<br>* Maximum audio input length is **120 seconds**<br>* Minimum effective speech length (excluding silence and other "non-speech" frames) is **4 seconds**<br>  This limitation can be disabled by setting "ignoreMinLength" to **true**.<br>* Minimum audio Signal-to-noise ratio (SNR) is **0dB**
      operationId: microsoftAzureTextindependentVerifyprofile
      summary: Microsoft Azure Verify Profile
      x-ms-examples:
        Successful Query:
          $ref: ./examples/verify.json
      tags:
      - Text-Independent
      parameters:
      - in: query
        name: ignoreMinLength
        description: If true, a voice print will be created immediately for this profile regardless of how much speech is supplied or stored. Default is false.
        schema:
          type: boolean
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/RecognitionInfo'
        '400':
          description: 'Bad Request

            * InvalidRequest - Invalid audio length. Maximum allowed length per recognition request is 120 seconds.

            * InvalidRequest - Invalid audio length. Minimum allowed length per recognition request is 1 second.

            * InvalidRequest - Invalid audio format. Supported format is 16Khz 16bit Mono PCM.

            * InvalidArgument - Invalid ignoreMinLength parameter value. Valid values are "true" and "false".

            * InvalidRequest - Profile is not enrolled.'
          x-ms-error-response: true
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
              example:
                error:
                  code: InvalidRequest
                  message: Profile is not enrolled.
        '401':
          $ref: '#/components/responses/Unauthorized'
        '404':
          $ref: '#/components/responses/NotFound'
        '415':
          description: UnsupportedMediaType - Unsupported media type. Only "audio/wav; codecs=audio/pcm" is accepted.
          x-ms-error-response: true
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
              example:
                error:
                  code: UnsupportedMediaType
                  message: Unsupported media type. Only 'audio/wav; codecs=audio/pcm' is accepted.
        '429':
          $ref: '#/components/responses/RateLimit'
        '500':
          $ref: '#/components/responses/ServerError'
      requestBody:
        content:
          audio/wav; codecs=audio/pcm:
            schema:
              type: object
              format: file
        description: Binary audio file. Supported formats are audio/wav; codecs=audio/pcm. Supports audio up to 5MB.
        required: true
components:
  schemas:
    LocaleInfo:
      description: Speaker profile locale
      type: object
      required:
      - locale
      properties:
        locale:
          $ref: '#/components/schemas/Locale'
    Error:
      description: Speaker error message
      required:
      - error
      properties:
        error:
          properties:
            code:
              type: string
            message:
              type: string
    TiProfileInfoList:
      description: Text-Independent Speaker profile info list
      type: object
      required:
      - value
      properties:
        value:
          type: array
          items:
            $ref: '#/components/schemas/TiProfileInfo'
        nextLink:
          type: string
          example: '{opaqueUrl}'
    ProfileId:
      type: string
      description: Unique identifier for profile id (guid).
      format: uuid
      example: 49a36324-fc4b-4387-aa06-090cfbf0064f
    Locale:
      type: string
      pattern: ^[a-zA-Z]{2}-?[a-zA-Z]{2}$
      description: Language identifier consisting of a combination of language code and country code.
      example: en-US
    SpeakerErrorInfo:
      description: Speaker error message
      type: object
      required:
      - error
      properties:
        error:
          required:
          - code
          - message
          type: object
          properties:
            code:
              type: string
            message:
              type: string
    CreatedDateTime:
      type: string
      format: date-time
      description: Profile creation datetime.
      example: '2015-04-23T18:25:43.41Z'
    EnrollmentsCount:
      type: integer
      description: Number of enrollment audios accepted for this profile.
      example: 1
    TiEnrollmentInfo:
      description: Speaker profile enrollment info
      type: object
      properties:
        profileId:
          $ref: '#/components/schemas/ProfileId'
        enrollmentStatus:
          $ref: '#/components/schemas/EnrollmentStatus'
        enrollmentsCount:
          $ref: '#/components/schemas/EnrollmentsCount'
        enrollmentsLengthInSec:
          $ref: '#/components/schemas/EnrollmentsLengthInSec'
        enrollmentsSpeechLengthInSec:
          $ref: '#/components/schemas/EnrollmentsSpeechLengthInSec'
        remainingEnrollmentsSpeechLengthInSec:
          $ref: '#/components/schemas/RemainingEnrollmentsSpeechLengthInSec'
        audioLengthInSec:
          type: number
          description: This enrollment audio length in seconds.
          example: 1.83
        audioSpeechLengthInSec:
          type: number
          description: This enrollment audio pure speech (which is the amount of audio after removing silence and non-speech segments) length in seconds.
          example: 1.35
    EnrollmentsSpeechLengthInSec:
      type: number
      description: Summation of pure speech (which is the amount of audio after removing silence and non-speech segments) across all profile enrollments in seconds.
      example: 1.35
    ActivationPhraseInfo:
      description: Activation phrase list
      type: object
      required:
      - value
      properties:
        value:
          type: array
          items:
            type: object
            properties:
              activationPhrase:
                $ref: '#/components/schemas/ActivationPhrase'
    IdentifyInfo:
      description: Identified speaker info
      type: object
      properties:
        profileId:
          type: string
          description: ID of identified of profile. If no candidate is identified as the right speaker, the value is set to empty GUID.
          example: 111f427c-3791-468f-b709-fcef7660fff9
          format: uuid
        score:
          type: number
          description: A float number indicating the similarity between input audio and targeted voice print. This number must be between 0 and 1. A higher number means higher similarity.
          example: 0.63
    ActivationPhrase:
      type: string
      description: Activation phrases available to activate a profile.
      example: this is my activation phrase to identify me
    IdentifiedSingleSpeakerInfo:
      type: object
      properties:
        identifiedProfile:
          description: Object containing data of identified profile.
          $ref: '#/components/schemas/IdentifyInfo'
        profilesRanking:
          description: Object containing data of the top 5 profiles (including identified profile) sorted in descending order by score.
          type: array
          items:
            $ref: '#/components/schemas/IdentifyInfo'
          example:
          - profileId: 111f427c-3791-468f-b709-fcef7660fff9
            score: 0.63
          - profileId: 3669fa29-1bf3-45ad-beea-6b348d058d7e
            score: 0.49
          - profileId: 0e196cd9-32d5-4883-8631-54a0e7c7cb3d
            score: 0.4
          - profileId: 726e57d9-04e0-4214-b482-7f786fa83560
            score: 0.1
          - profileId: f95189fd-1bf5-4485-9c2e-e5897e0c98ca
            score: 0.03
    RemainingEnrollmentsSpeechLengthInSec:
      type: number
      description: Amount of pure speech (which is the amount of audio after removing silence and non-speech segments) needed to complete profile enrollment in seconds.
      example: 18.65
    EnrollmentStatus:
      type: string
      description: 'Status representing the current state of the profile enrollment. Available values are:

        * Enrolling: profile has no voice print and not ready for recognition requests.

        * Training: voice print of profile is being created and can’t be used for recognition at the moment.

        * Enrolled: profile has a voice print and ready for recognition requests.'
      enum:
      - Enrolling
      - Training
      - Enrolled
      x-ms-enum:
        name: TrainingStatusType
        modelAsString: false
      example: Enrolling
    ModelVersion:
      type: string
      format: date
      description: Date specifying the model assigned to this profile. Format is yyyy-mm-dd. If profile has no enrollments, this value will be empty.
      example: '2019-12-05'
    ProfileStatus:
      type: string
      description: 'Status representing the current state of the profile activation. Available values are:

        * Active: profile is active and can be used if the enrollment status is ''Enrolled''.

        * Inactive: profile has not been activated and an activation phrase must be submitted.'
      enum:
      - Active
      - Inactive
      x-ms-enum:
        name: ProfileStatusType
        modelAsString: false
      example: Inactive
    LastUpdatedDateTime:
      type: string
      format: date-time
      description: Last datetime when the profile was updated.
      example: '2015-04-23T19:34:51.52Z'
    RecognitionInfo:
      description: Speaker recognition result
      properties:
        recognitionResult:
          type: string
          description: Does audio belong to targeted profile?
          enum:
          - accept
          - reject
          x-ms-enum:
            name: ResultType
            modelAsString: false
          example: Accept
        score:
          type: number
          description: A float number indicating the similarity between input audio and targeted voice print. This number must be between 0 and 1. A higher number means higher similarity.
          example: 0.63
    TiProfileInfo:
      description: Text-Independent Speaker profile info
      type: object
      properties:
        profileId:
          $ref: '#/components/schemas/ProfileId'
        locale:
          $ref: '#/components/schemas/Locale'
        profileStatus:
          $ref: '#/components/schemas/ProfileStatus'
        enrollmentStatus:
          $ref: '#/components/schemas/EnrollmentStatus'
        createdDateTime:
          $ref: '#/components/schemas/CreatedDateTime'
        lastUpdatedDateTime:
          $ref: '#/components/schemas/LastUpdatedDateTime'
        enrollmentsCount:
          $ref: '#/components/schemas/EnrollmentsCount'
        enrollmentsLengthInSec:
          $ref: '#/components/schemas/EnrollmentsLengthInSec'
        enrollmentsSpeechLengthInSec:
          $ref: '#/components/schemas/EnrollmentsSpeechLengthInSec'
        remainingEnrollmentsSpeechLengthInSec:
          $ref: '#/components/schemas/RemainingEnrollmentsSpeechLengthInSec'
        modelVersion:
          $ref: '#/components/schemas/ModelVersion'
    EnrollmentsLengthInSec:
      type: number
      description: Total length of enrollment audios accepted for this profile in seconds.
      example: 1.83
  responses:
    SpeakerErrorResponse:
      description: Failure
      x-ms-error-response: true
      headers:
        x-ms-error-code:
          description: Error code
          schema:
            type: string
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/SpeakerErrorInfo'
    Unauthorized:
      description: Request is not authorized. Make sure subscription key is included and valid.
      x-ms-error-response: true
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
          example:
            error:
              code: Unauthorized
              message: Request is not authorized. Make sure subscription key is included and valid.
    ServerError:
      description: Internal Server Error.
      x-ms-error-response: true
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
          example:
            error:
              code: InternalServerError
              message: Internal Server Error.
    RateLimit:
      description: Rate limit is exceeded.
      x-ms-error-response: true
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
          example:
            error:
              code: RateLimit
              message: Rate limit is exceeded.
    NotFound:
      description: NotFound - Requested profile doesn't exist
      x-ms-error-response: true
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
          example:
            error:
              code: Not Found
              message: Requested profile doesn't exist
  parameters:
    profileIdParam:
      in: path
      name: profileId
      required: true
      description: Unique identifier for profile id (guid).
      x-ms-parameter-location: method
      schema:
        type: string
        format: uuid
    apiVersionParam:
      name: api-version
      x-ms-parameter-location: client
      in: query
      required: true
      description: Specifies the version of the operation to use for this request.
      schema:
        type: string
        enum:
        - '2021-09-05'
        default: '2021-09-05'
    profileIdParam_2:
      in: path
      name: profileId
      required: true
      description: Unique identifier for profile id (guid).
      x-ms-parameter-location: method
      schema:
        type: string
        format: uuid
        pattern: ^([0-9a-fA-F]){8}-?([0-9a-fA-F]){4}-?([0-9a-fA-F]){4}-?([0-9a-fA-F]){4}-?([0-9a-fA-F]){12}$
    localeParam:
      in: path
      name: locale
      required: true
      description: A combination of language code and country code.
      x-ms-parameter-location: method
      schema:
        type: string
        pattern: ^[a-zA-Z]{2}-?[a-zA-Z]{2}$
x-ms-parameterized-host:
  hostTemplate: '{endpoint}'
  useSchemePrefix: false
  parameters:
  - $ref: '#/components/parameters/Endpoint'