Skip to main content

speaker_diarization.proto

path phonexia/grpc/technologies/speaker_diarization/v1/speaker_diarization.proto

package phonexia.grpc.technologies.speaker_diarization.v1

Phonexia Speaker Diarization gRPC API


Messages​

DiarizeConfig​

NameTypeDescription
max_speakersoptional int32

Upper limit for the number of speakers in the audio. The value must be larger than 0. If omitted, the default value is set to 100.

total_speakersoptional int32

Enforces exact number of speakers in the audio. If set, it overwrites max_speakers. If not set, the number of speakers is detected automatically, with upper limit defined by max_speakers.

DiarizeRequest​

The top-level message sent by the client for the Diarize method.

NameTypeDescription
audiophonexia.grpc.common.Audio

Audio to be diarized. The audio should have minimum length of 0.8 seconds. If the audio is shorter, the segments will be empty.

configDiarizeConfig

Speaker diarization configuration.

DiarizeResponse​

The top-level message returned to the client by the Diarize method.

NameTypeDescription
speakers_countint32

Detected number of speakers in the audio.

segmentsrepeated Segment

List of detected segments in the audio. Will be empty if the audio is shorter than 0.8 seconds.

processed_audio_lengthgoogle.protobuf.Duration

Total length of the processed audio.

speech_lengthgoogle.protobuf.Duration

Total Speech length of the processed audio.

Segment​

NameTypeDescription
speaker_idint32

Identifier of a speaker in the segment.

start_timegoogle.protobuf.Duration

Start time of the segment.

end_timegoogle.protobuf.Duration

End time of the segment.


Services​

SpeakerDiarization​

Service implementing speaker diarization.

Diarize​

MethodDiarize
RequestDiarizeRequest stream
ResponseDiarizeResponse
Description

Performs synchronous speaker diarization of an audio. Returns a result after completely sent and processed audio.