v2

latestSwagger 2.02026-07-1371728.8 KB

text-independent

Identify Single Speaker Profile

Identifies who is speaking in input audio among a list of candidate profiles.

Limitations:

Minimum audio input length is 1 second
Maximum audio input length is 120 seconds
Minimum candidate speakers count is 1
Maximum candidate speakers count is 50
Minimum effective speech length (excluding silence and other non-speech frames) is 4 seconds This limitation can be disabled by setting "ignoreMinLength" to true.
Minimum audio Signal-to-noise ratio (SNR) is 0dB

post/text-independent/profiles/identifySingleSpeaker

Query parameters

profileIdsstring required

Comma-delimited profile IDs. Maximum supported number is 50 IDs.

ignoreMinLengthboolean

If true, the minimum amount of speech needed for identification is skipped. Default is false.

Response

Example response

{
  "identifiedProfile": {
    "profileId": "111f427c-3791-468f-b709-fcef7660fff9",
    "score": 0.63
  },
  "profilesRanking": [
    {
      "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
    }
  ]
}