Whisper.WhisperManager Class Reference

Manages Whisper model lifecycle in Unity scene. More...

Inheritance diagram for Whisper.WhisperManager:

Public Member Functions
async Task	InitModel ()
	Load model and default parameters. Prepare it for text transcription.
bool	IsMultilingual ()
	Checks if currently loaded whisper model supports multilingual transcription.
async Task< WhisperResult >	GetTextAsync (AudioClip clip)
	Start async transcription of audio clip.
async Task< WhisperResult >	GetTextAsync (float[] samples, int frequency, int channels)
	Start async transcription of audio buffer.
async Task< WhisperStream >	CreateStream (int frequency, int channels)
	Create a new instance of Whisper streaming transcription.
async Task< WhisperStream >	CreateStream (MicrophoneRecord microphone)
	Create a new instance of Whisper streaming transcription from microphone input.

Public Attributes
LogLevel	logLevel = LogLevel.Log
string	language = "en"
bool	translateToEnglish
bool	noContext = true
bool	singleSegment
bool	enableTokens
string	initialPrompt
float	stepSec = 3f
float	keepSec = 0.2f
float	lengthSec = 10f
bool	updatePrompt = true
bool	dropOldBuffer
bool	useVad = true
bool	tokensTimestamps
bool	speedUp
int	audioCtx

Properties
string	ModelPath [get, set]
bool	IsModelPathInStreamingAssets [get, set]
bool	IsLoaded [get]
	Checks if whisper weights are loaded and ready to be used.
bool	IsLoading [get]
	Checks if whisper weights are still loading and not ready.

Events
OnNewSegmentDelegate	OnNewSegment
	Raised when whisper transcribed a new text segment from audio.
OnProgressDelegate	OnProgress
	Raised when whisper made some progress in transcribing audio. Progress changes from 0 to 100 included.

Detailed Description

Manages Whisper model lifecycle in Unity scene.

Member Function Documentation

CreateStream() [1/2]

async Task< WhisperStream > Whisper.WhisperManager.CreateStream ( int frequency, int channels )

inline

Create a new instance of Whisper streaming transcription.

Parameters

frequency	Audio sample rate.
channels	Audio channels count.

Returns

New streaming transcription. Null if failed.

CreateStream() [2/2]

async Task< WhisperStream > Whisper.WhisperManager.CreateStream ( MicrophoneRecord microphone )

inline

Create a new instance of Whisper streaming transcription from microphone input.

Returns

New streaming transcription. Null if failed.

GetTextAsync() [1/2]

async Task< WhisperResult > Whisper.WhisperManager.GetTextAsync ( AudioClip clip )

inline

Start async transcription of audio clip.

Returns

Full audio transcript. Null if transcription failed.

GetTextAsync() [2/2]

async Task< WhisperResult > Whisper.WhisperManager.GetTextAsync ( float[] samples, int frequency, int channels )

inline

Start async transcription of audio buffer.

Parameters

samples	Raw audio buffer.
frequency	Audio sample rate.
channels	Audio channels count.

Returns

Full audio transcript. Null if transcription failed.

InitModel()

async Task Whisper.WhisperManager.InitModel ( )

inline

Load model and default parameters. Prepare it for text transcription.

IsMultilingual()

bool Whisper.WhisperManager.IsMultilingual ( )

inline

Checks if currently loaded whisper model supports multilingual transcription.

Member Data Documentation

audioCtx

int Whisper.WhisperManager.audioCtx

dropOldBuffer

bool Whisper.WhisperManager.dropOldBuffer

enableTokens

bool Whisper.WhisperManager.enableTokens

initialPrompt

string Whisper.WhisperManager.initialPrompt

keepSec

float Whisper.WhisperManager.keepSec = 0.2f

language

string Whisper.WhisperManager.language = "en"

lengthSec

float Whisper.WhisperManager.lengthSec = 10f

logLevel

LogLevel Whisper.WhisperManager.logLevel = LogLevel.Log

noContext

bool Whisper.WhisperManager.noContext = true

singleSegment

bool Whisper.WhisperManager.singleSegment

speedUp

bool Whisper.WhisperManager.speedUp

stepSec

float Whisper.WhisperManager.stepSec = 3f

tokensTimestamps

bool Whisper.WhisperManager.tokensTimestamps

translateToEnglish

bool Whisper.WhisperManager.translateToEnglish

updatePrompt

bool Whisper.WhisperManager.updatePrompt = true

useVad

bool Whisper.WhisperManager.useVad = true

Property Documentation

IsLoaded

bool Whisper.WhisperManager.IsLoaded

get

Checks if whisper weights are loaded and ready to be used.

IsLoading

bool Whisper.WhisperManager.IsLoading

get

Checks if whisper weights are still loading and not ready.

IsModelPathInStreamingAssets

bool Whisper.WhisperManager.IsModelPathInStreamingAssets

getset

ModelPath

string Whisper.WhisperManager.ModelPath

getset

Event Documentation

OnNewSegment

OnNewSegmentDelegate Whisper.WhisperManager.OnNewSegment

Raised when whisper transcribed a new text segment from audio.

OnProgress

OnProgressDelegate Whisper.WhisperManager.OnProgress

Raised when whisper made some progress in transcribing audio. Progress changes from 0 to 100 included.

---

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

• /github/workspace/Packages/FMAR/ThirdParty/WhisperUnity/Runtime/WhisperManager.cs