homebot / stc-speechkit-web

Javascript libraries for REST API (ASR, TTS, Diarization)

STC speech libraries for web

Libraries for the use of rest API recognition, synthesis and speech diarization

ASR

rest API https://cp.speechpro.com/vkasr/rest/
help https://cp.speechpro.com/vkasr/help

TTS

rest API https://cp.speechpro.com/vktts/rest/
help https://cp.speechpro.com/vktts/help

Diarization

rest API https://cp.speechpro.com/vkdiarization/rest/
help https://cp.speechpro.com/vkdiarization/help

Using library ASR

Load main script from HTML first.

<script src="SpeechProASR.js"></script>

Constructor

asr = new SpeechProASR(config);

• config (Parameters with default values are not required)
  • host: url rest API (default = "https://cp.speechpro.com/vkasr/rest/")
  • client: user data for authorization
    • .username: "username"
    • .password: "password"
    • .domain_id: "id"
  • recorder: record audio input and encodes to audio file image (Blob object). After the end of the recording, it sends the data to the server for recognition (default = false)
  • bufferLength: The size of the buffer sent to the server (only WebSockets)
  • package: package name for using recognize (default = "TelecomRus")
  • packageSocket: package name for using online recognize (WebSockets) (default = "CommonRus")

Methods

asr.getPackages();

Get all available packages

• Parameters
  • (none)
• Returns (Promise then or catch)
  • Array: array of objects with data about available packages

asr.setPackage(packageId);

Set package after initialization

• Parameters
  • .packageId: package identifier (type: 'String')
• Returns
  • (none)

asr.setPackageSocket(packageId);

Set package for WebSockets after initialization

• Parameters
  • .packageId: package identifier (type: 'String')
• Returns
  • (none)

asr.recognizeFile(file);

Recognize from wav file

• Note
  • The file must have a sampling frequency of 8000
• Parameters
  • .file: blob object (type: 'audio/wav')
• Returns
  • (none)

asr.startRecord();

Start recording from microphone (Can be used if the parameter recorder = true)

• Parameters
  • (none)
• Returns
  • (none)

asr.stopRecord();

Stop recording and send audio to server (Can be used if the parameter recorder = true)

• Parameters
  • (none)
• Returns
  • (none)

asr.startRecordSocket();

Start recording and send audio to server on WebSockets (automatically used recorder)

• Parameters
  • (none)
• Returns
  • (none)

asr.stopRecordSocket();

Stop recording and close WebSockets

• Parameters
  • (none)
• Returns
  • (none)

Events

complete - asr object has been initialized
recognizeComplete - Recognition result
recognizeSocketComplete - Recognition result web sockets
recognizeSocketCompleteFinal - The final result of recognition web sockets (Returns the result after the socket is closed)
recognizeError - Recognition error

Examples

let asr = new SpeechProASR({
  "client": {
    "username": "username",
    "password": "password",
    "domain_id": "id"
  },
  "recorder": true, //default false
});

asr.complete = function(){
  //  asr object initialized, here you can use the available methods

  //  Events
  
  asr.recognizeComplete = function(result){
    //  result = {
    //    "score": recognize score,
    //    "text": recognized text,
    //  }
  }

  asr.recognizeError = function(error){
    //  recognize error
  }

  asr.recognizeSocketComplete = function(result){
    //  result = recognized text
    
  }
  
  asr.recognizeSocketCompleteFinal = function(result){
    //  result = {
    //    "score": recognize score,
    //    "text": recognized text,
    //  }
  }
}

Using library TTS

Load main script from HTML first.

<script src="SpeechProTTS.js"></script>

Constructor

tts = new SpeechProTTS(config);

• config (Parameters with default values are not required)
  • host: url rest API (default = "https://cp.speechpro.com/vktts/rest/")
  • client: user data for authorization
    • .username: "username"
    • .password: "password"
    • .domain_id: "id"

Methods

tts.getLanguages();

Get all available languages

• Parameters
  • (none)
• Returns (Promise then or catch)
  • Array: array of objects with data about available languages

tts.getVoices(language);

Get all available voices

• Parameters
  • language: 'language name' (default = 'Russian')
• Returns (Promise then or catch)
  • Array: array of objects with data about available voices

tts.setLanguage(language);

Set language after initialization

• Parameters
  • .language: language identifier (type: 'String')
• Returns
  • (none)

tts.setVoice(voice);

Set voice after initialization

• Parameters
  • .voice: voice identifier (type: 'String')
• Returns
  • (none)

tts.synthesize(text, options);

Text synthesis

• Parameters
  • text: "your text"
  • options:
    • .voice: "voice_name" (default = 'Alexander')
    • .play: Synthesized text playing (default = 'false')
• Returns
  • (none)

tts.sendSocket(text, options);

Text synthesis for WebSockets

• Parameters
  • text: "your text"
  • options:
    • .voice: "voice_name" (default = 'Alexander')
    • .play: Synthesized text playing after socket closed (default = 'false')
• Returns
  • (none)

tts.getWavFromBuffer();

Returns the wav file from the accumulated socket message buffer. If the socket is closed, returns null

• Parameters
  • (none)
• Returns
  • blob: object Blob (type = 'audio/wav')

Events

complete - tts object has been initialized
synthesizeComplete - Synthesis result
synthesizeSocketComplete - Synthesis result WebSockets
synthesizeError - Synthesis error
synthesizeSocketOnline - Intermediate synthesis result (socket messages, type: 'ArrayBuffer')

Examples

let tts = new SpeechProTTS({
  "client": {
    "username": "username",
    "password": "password",
    "domain_id": "id"
  }
});

tts.complete = function() {
  //  tts object initialized, here you can use the available methods
  
  // Events 
  tts.synthesizeComplete = function(result) {
    //  result = {
    //    "b64Data": base64 data,
    //    "blob": object blob (type = 'audio/wav'),
    //    "blobUrl": blobUrl (object URL.createObjectURL)
    //  }
  }

  tts.synthesizeError = function(error) {
    // Synthesis error
  }

  tts.synthesizeSocketComplete = function(result) {
    //  result = {
    //    "blob": object blob (type = 'audio/wav'),
    //    "blobUrl": blobUrl (object URL.createObjectURL)
    //  }
  }

  tts.synthesizeSocketOnline = function(arraybuffer) {
    //  result socket message
  }
}

Using library Diarization

Load main script from HTML first.

<script src="SpeechProDiarization.js"></script>

Constructor

diariz = new SpeechProDiarization(config);

• config (Parameters with default values are not required)
  • host: url rest API (default = "https://cp.speechpro.com/vkasr/rest/")
  • client: user data for authorization
    • .username: "username"
    • .password: "password"
    • .domain_id: "id"
  • recorder: record audio input and encodes to audio file image (Blob object). After the end of the recording, it sends the data to the server for diarization (default = false)

Methods

diariz.diarization(file);

Diarization from wav file

• Parameters
  • .file: blob object (type: 'audio/wav')
• Returns
  • (none)

diariz.startRecord();

Start recording from microphone (Can be used if the parameter recorder = true)

• Parameters
  • (none)
• Returns
  • (none)

diariz.stopRecord();

Stop recording and send audio to server (Can be used if the parameter recorder = true)

• Parameters
  • (none)
• Returns
  • (none)

Events

complete - diarization object has been initialized
diarizationComplete - Diarization result
diarizationError - Diarization error

Examples

let diariz = new SpeechProDiarization({
  "client": {
    "username": "username",
    "password": "password",
    "domain_id": "id"
  },
  "recorder": true, //default false
});

diariz.complete = function(){
  //  diarization object initialized, here you can use the available methods

  //  Events
  diariz.diarizationComplete = function(result){
    //  result = {
    //    "data": {
    //      "speakers": [{
    //        "number": 0,
    //        "segments": [{
    //          "start": 0,
    //          "length": 0
    //        }]
    //      }]
    //    }
    //  }
  }

  diariz.diarizationError = function(error){
    //  recognize error
  }
}