Parameters of whisper streaming processing. More...

Public Member Functions
	WhisperStreamParams (WhisperParams inferenceParam, int frequency, int channels, float stepSec=3f, float keepSec=0.2f, float lengthSec=10f, bool updatePrompt=true, bool dropOldBuffer=false, bool useVad=false)

Public Attributes
readonly WhisperParams	InferenceParam
	Regular whisper inference params.

readonly int	Frequency
	Audio stream frequency. Can't change during transcription.

readonly int	Channels
	Audio stream channels count. Can't change during transcription.

readonly float	StepSec
	Minimal portions of audio that will be processed by whisper stream in seconds.

readonly int	StepSamples
	Minimal portions of audio that will be processed by whisper in audio samples.

readonly float	KeepSec
	How many seconds of previous segment will be used for current segment.

readonly int	KeepSamples
	How many samples of previous audio chunk will be used for current chunk.

readonly float	LengthSec
	How many seconds of audio will be recurrently transcribe until context update.

readonly int	LengthSamples
	How many samples of audio will be recurrently transcribe until context update.

readonly bool	UpdatePrompt
	Should stream modify whisper prompt for better context handling?

readonly int	StepsCount
	How many recurrent iterations will be used for one chunk?

readonly bool	DropOldBuffer
	If false stream will use all information from previous iteration.

readonly bool	UseVad
	If true stream will ignore audio chunks with no detected speech.

Detailed Description

Parameters of whisper streaming processing.

Constructor & Destructor Documentation

◆ WhisperStreamParams()

Whisper.WhisperStreamParams.WhisperStreamParams	(	WhisperParams	inferenceParam,
		int	frequency,
		int	channels,
		float	stepSec = 3f,
		float	keepSec = 0::2f,
		float	lengthSec = 10f,
		bool	updatePrompt = true,
		bool	dropOldBuffer = false,
		bool	useVad = false )

inline

Member Data Documentation

◆ Channels

readonly int Whisper.WhisperStreamParams.Channels

Audio stream channels count. Can't change during transcription.

◆ DropOldBuffer

readonly bool Whisper.WhisperStreamParams.DropOldBuffer

If false stream will use all information from previous iteration.

◆ Frequency

readonly int Whisper.WhisperStreamParams.Frequency

Audio stream frequency. Can't change during transcription.

◆ InferenceParam

readonly WhisperParams Whisper.WhisperStreamParams.InferenceParam

Regular whisper inference params.

◆ KeepSamples

readonly int Whisper.WhisperStreamParams.KeepSamples

How many samples of previous audio chunk will be used for current chunk.

◆ KeepSec

readonly float Whisper.WhisperStreamParams.KeepSec

How many seconds of previous segment will be used for current segment.

◆ LengthSamples

readonly int Whisper.WhisperStreamParams.LengthSamples

How many samples of audio will be recurrently transcribe until context update.

◆ LengthSec

readonly float Whisper.WhisperStreamParams.LengthSec

How many seconds of audio will be recurrently transcribe until context update.

◆ StepSamples

readonly int Whisper.WhisperStreamParams.StepSamples

Minimal portions of audio that will be processed by whisper in audio samples.

◆ StepsCount

readonly int Whisper.WhisperStreamParams.StepsCount

How many recurrent iterations will be used for one chunk?

◆ StepSec

readonly float Whisper.WhisperStreamParams.StepSec

Minimal portions of audio that will be processed by whisper stream in seconds.

◆ UpdatePrompt

readonly bool Whisper.WhisperStreamParams.UpdatePrompt

Should stream modify whisper prompt for better context handling?

◆ UseVad

readonly bool Whisper.WhisperStreamParams.UseVad

If true stream will ignore audio chunks with no detected speech.

The documentation for this class was generated from the following file:

/github/workspace/Packages/FMAR/ThirdParty/WhisperUnity/Runtime/WhisperStream.cs

Public Member Functions

Public Attributes