Blame - docs/dyn/speech_v1p1beta1.speech.html - platform/external/python/google-api-python-client

2019-06-14 16:50:42 -0700

[diff] [blame]

100

# process the request.

101

# request.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

102

"metadata": { # Description of audio data to be recognized. # Metadata regarding this request.

103

"originalMediaType": "A String", # The original media the speech was recorded on.

104

"obfuscatedId": "A String", # Obfuscated (privacy-protected) ID of the user, to identify number of

105

# unique users using the service.

106

"recordingDeviceType": "A String", # The type of device the speech was recorded with.

107

"interactionType": "A String", # The use case most closely describing the audio content to be recognized.

108

"recordingDeviceName": "A String", # The device used to make the recording. Examples 'Nexus 5X' or

109

# 'Polycom SoundStation IP 6000' or 'POTS' or 'VoIP' or

110

# 'Cardioid Microphone'.

111

"originalMimeType": "A String", # Mime type of the original audio file. For example `audio/m4a`,

112

# `audio/x-alaw-basic`, `audio/mp3`, `audio/3gpp`.

113

# A list of possible audio mime types is maintained at

114

# http://www.iana.org/assignments/media-types/media-types.xhtml#audio

115

"audioTopic": "A String", # Description of the content. Eg. "Recordings of federal supreme court

116

# hearings from 2012".

117

"industryNaicsCodeOfAudio": 42, # The industry vertical to which this speech recognition request most

118

# closely applies. This is most indicative of the topics contained

119

# in the audio. Use the 6-digit NAICS code to identify the industry

120

# vertical - see https://www.naics.com/search/.

121

"microphoneDistance": "A String", # The audio type that most closely describes the audio being recognized.

122

},

123

"sampleRateHertz": 42, # Sample rate in Hertz of the audio data sent in all

124

# `RecognitionAudio` messages. Valid values are: 8000-48000.

125

# 16000 is optimal. For best results, set the sampling rate of the audio

126

# source to 16000 Hz. If that's not possible, use the native sample rate of

127

# the audio source (instead of re-sampling).

128

# This field is optional for FLAC and WAV audio files, but is

129

# required for all other audio formats. For details, see AudioEncoding.

130

"enableSeparateRecognitionPerChannel": True or False, # This needs to be set to `true` explicitly and `audio_channel_count` > 1

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

131

# to get each channel recognized separately. The recognition result will

132

# contain a `channel_tag` field to state which channel that result belongs

133

# to. If this is not true, we will only recognize the first channel. The

134

# request is billed cumulatively for all channels recognized:

135

# `audio_channel_count` multiplied by the length of the audio.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

136

"enableAutomaticPunctuation": True or False, # If 'true', adds punctuation to recognition result hypotheses.

137

# This feature is only available in select languages. Setting this for

138

# requests in other languages has no effect at all.

139

# The default 'false' value does not add punctuation to result hypotheses.

140

"adaptation": { # Speech adaptation configuration. # Speech adaptation configuration improves the accuracy of speech

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

141

# recognition. When speech adaptation is set it supersedes the

142

# `speech_contexts` field. For more information, see the [speech

143

# adaptation](https://cloud.google.com/speech-to-text/docs/context-strength)

144

# documentation.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

145

"phraseSets": [ # A collection of phrase sets. To specify the hints inline, leave the

146

# phrase set's `name` blank and fill in the rest of its fields. Any

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

147

# phrase set can use any custom class.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

148

{ # Provides "hints" to the speech recognizer to favor specific words and phrases

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

149

# in the results.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

150

"boost": 3.14, # Hint Boost. Positive value will increase the probability that a specific

151

# phrase will be recognized over other similar sounding phrases. The higher

152

# the boost, the higher the chance of false positive recognition as well.

153

# Negative boost values would correspond to anti-biasing. Anti-biasing is not

154

# enabled, so negative boost will simply be ignored. Though `boost` can

155

# accept a wide range of positive values, most use cases are best served with

156

# values between 0 (exclusive) and 20. We recommend using a binary search

157

# approach to finding the optimal value for your use case. Speech recognition

158

# will skip PhraseSets with a boost value of 0.

159

"name": "A String", # The resource name of the phrase set.

160

"phrases": [ # A list of word and phrases.

161

{ # A phrases containing words and phrase "hints" so that

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

162

# the speech recognition is more likely to recognize them. This can be used

163

# to improve the accuracy for specific words and phrases, for example, if

164

# specific commands are typically spoken by the user. This can also be used

165

# to add additional words to the vocabulary of the recognizer. See

166

# [usage limits](https://cloud.google.com/speech-to-text/quotas#content).

167

#

168

# List items can also include pre-built or custom classes containing groups

169

# of words that represent common concepts that occur in natural language. For

170

# example, rather than providing a phrase hint for every month of the

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

171

# year (e.g. "i was born in january", "i was born in febuary", ...), use the

172

# pre-built `$MONTH` class improves the likelihood of correctly transcribing

173

# audio that includes months (e.g. "i was born in $month").

174

# To refer to pre-built classes, use the class' symbol prepended with `$`

175

# e.g. `$MONTH`. To refer to custom classes that were defined inline in the

176

# request, set the class's `custom_class_id` to a string unique to all class

177

# resources and inline classes. Then use the class' id wrapped in $`{...}`

178

# e.g. "${my-months}". To refer to custom classes resources, use the class'

179

# id wrapped in `${}` (e.g. `${my-months}`).

180

"value": "A String", # The phrase itself.

181

"boost": 3.14, # Hint Boost. Overrides the boost set at the phrase set level.

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

182

# Positive value will increase the probability that a specific phrase will

183

# be recognized over other similar sounding phrases. The higher the boost,

184

# the higher the chance of false positive recognition as well. Negative

185

# boost values would correspond to anti-biasing. Anti-biasing is not

186

# enabled, so negative boost will simply be ignored. Though `boost` can

187

# accept a wide range of positive values, most use cases are best served

188

# with values between 0 and 20. We recommend using a binary search approach

189

# to finding the optimal value for your use case. Speech recognition

190

# will skip PhraseSets with a boost value of 0.

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

191

},

192

],

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

193

},

194

],

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

195

"customClasses": [ # A collection of custom classes. To specify the classes inline, leave the

196

# class' `name` blank and fill in the rest of its fields, giving it a unique

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

197

# `custom_class_id`. Refer to the inline defined class in phrase hints by its

198

# `custom_class_id`.

199

{ # A set of words or phrases that represents a common concept likely to appear

200

# in your audio, for example a list of passenger ship names. CustomClass items

201

# can be substituted into placeholders that you set in PhraseSet phrases.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

202

"name": "A String", # The resource name of the custom class.

203

"customClassId": "A String", # If this custom class is a resource, the custom_class_id is the resource id

204

# of the CustomClass. Case sensitive.

205

"items": [ # A collection of class items.

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

206

{ # An item of the class.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

207

"value": "A String", # The class item's value.

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

208

},

209

],

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

210

},

211

],

212

},

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

213

"maxAlternatives": 42, # Maximum number of recognition hypotheses to be returned.

214

# Specifically, the maximum number of `SpeechRecognitionAlternative` messages

215

# within each `SpeechRecognitionResult`.

216

# The server may return fewer than `max_alternatives`.

217

# Valid values are `0`-`30`. A value of `0` or `1` will return a maximum of

218

# one. If omitted, will return a maximum of one.

219

"diarizationSpeakerCount": 42, # If set, specifies the estimated number of speakers in the conversation.

220

# Defaults to '2'. Ignored unless enable_speaker_diarization is set to true.

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

221

# Note: Use diarization_config instead.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

222

"encoding": "A String", # Encoding of audio data sent in all `RecognitionAudio` messages.

223

# This field is optional for `FLAC` and `WAV` audio files and required

224

# for all other audio formats. For details, see AudioEncoding.

225

"speechContexts": [ # Array of SpeechContext.

226

# A means to provide context to assist the speech recognition. For more

227

# information, see

228

# [speech

229

# adaptation](https://cloud.google.com/speech-to-text/docs/context-strength).

230

{ # Provides "hints" to the speech recognizer to favor specific words and phrases

231

# in the results.

232

"phrases": [ # A list of strings containing words and phrases "hints" so that

233

# the speech recognition is more likely to recognize them. This can be used

234

# to improve the accuracy for specific words and phrases, for example, if

235

# specific commands are typically spoken by the user. This can also be used

236

# to add additional words to the vocabulary of the recognizer. See

237

# [usage limits](https://cloud.google.com/speech-to-text/quotas#content).

238

#

239

# List items can also be set to classes for groups of words that represent

240

# common concepts that occur in natural language. For example, rather than

241

# providing phrase hints for every month of the year, using the $MONTH class

242

# improves the likelihood of correctly transcribing audio that includes

# months.

"A String",

],

"boost": 3.14, # Hint Boost. Positive value will increase the probability that a specific

247

# phrase will be recognized over other similar sounding phrases. The higher

248

# the boost, the higher the chance of false positive recognition as well.

249

# Negative boost values would correspond to anti-biasing. Anti-biasing is not

250

# enabled, so negative boost will simply be ignored. Though `boost` can

251

# accept a wide range of positive values, most use cases are best served with

252

# values between 0 and 20. We recommend using a binary search approach to

253

# finding the optimal value for your use case.

254

},

255

],

256

"enableWordConfidence": True or False, # If `true`, the top result includes a list of words and the

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

257

# confidence for those words. If `false`, no word-level confidence

258

# information is returned. The default is `false`.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

259

"model": "A String", # Which model to select for the given request. Select the model

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

260

# best suited to your domain to get best results. If a model is not

261

# explicitly specified, then we auto-select a model based on the parameters

262

# in the RecognitionConfig.

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

263

# <table>

264

# <tr>

265

# <td><b>Model</b></td>

266

# <td><b>Description</b></td>

267

# </tr>

268

# <tr>

269

# <td><code>command_and_search</code></td>

270

# <td>Best for short queries such as voice commands or voice search.</td>

271

# </tr>

272

# <tr>

273

# <td><code>phone_call</code></td>

274

# <td>Best for audio that originated from a phone call (typically

275

# recorded at an 8khz sampling rate).</td>

276

# </tr>

277

# <tr>

278

# <td><code>video</code></td>

279

# <td>Best for audio that originated from from video or includes multiple

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

280

# speakers. Ideally the audio is recorded at a 16khz or greater

281

# sampling rate. This is a premium model that costs more than the

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

282

# standard rate.</td>

283

# </tr>

284

# <tr>

285

# <td><code>default</code></td>

286

# <td>Best for audio that is not one of the specific audio models.

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

287

# For example, long-form audio. Ideally the audio is high-fidelity,

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

288

# recorded at a 16khz or greater sampling rate.</td>

289

# </tr>

290

# </table>

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

291

"audioChannelCount": 42, # The number of channels in the input audio data.

292

# ONLY set this for MULTI-CHANNEL recognition.

293

# Valid values for LINEAR16 and FLAC are `1`-`8`.

294

# Valid values for OGG_OPUS are '1'-'254'.

295

# Valid value for MULAW, AMR, AMR_WB and SPEEX_WITH_HEADER_BYTE is only `1`.

296

# If `0` or omitted, defaults to one channel (mono).

297

# Note: We only recognize the first channel by default.

298

# To perform independent recognition on each channel set

299

# `enable_separate_recognition_per_channel` to 'true'.

300

"enableWordTimeOffsets": True or False, # If `true`, the top result includes a list of words and

301

# the start and end time offsets (timestamps) for those words. If

302

# `false`, no word-level time offset information is returned. The default is

303

# `false`.

304

"alternativeLanguageCodes": [ # A list of up to 3 additional

305

# [BCP-47](https://www.rfc-editor.org/rfc/bcp/bcp47.txt) language tags,

306

# listing possible alternative languages of the supplied audio.

307

# See [Language

308

# Support](https://cloud.google.com/speech-to-text/docs/languages) for a list

309

# of the currently supported language codes. If alternative languages are

310

# listed, recognition result will contain recognition in the most likely

311

# language detected including the main language_code. The recognition result

312

# will include the language tag of the language detected in the audio. Note:

313

# This feature is only supported for Voice Command and Voice Search use cases

314

# and performance may vary for other use cases (e.g., phone call

# transcription).

"A String",

],

"diarizationConfig": { # Config to enable speaker diarization. # Config to enable speaker diarization and set additional

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

319

# parameters to make diarization better suited for your application.

320

# Note: When this is enabled, we send all the words from the beginning of the

321

# audio for the top alternative in every consecutive STREAMING responses.

322

# This is done in order to improve our speaker tags as our models learn to

323

# identify the speakers in the conversation over time.

324

# For non-streaming requests, the diarization results will be provided only

325

# in the top alternative of the FINAL SpeechRecognitionResult.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

326

"minSpeakerCount": 42, # Minimum number of speakers in the conversation. This range gives you more

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

327

# flexibility by allowing the system to automatically determine the correct

328

# number of speakers. If not set, the default value is 2.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

329

"maxSpeakerCount": 42, # Maximum number of speakers in the conversation. This range gives you more

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

330

# flexibility by allowing the system to automatically determine the correct

331

# number of speakers. If not set, the default value is 6.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

332

"speakerTag": 42, # Output only. Unused.

333

"enableSpeakerDiarization": True or False, # If 'true', enables speaker detection for each recognized word in

334

# the top alternative of the recognition result using a speaker_tag provided

335

# in the WordInfo.

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

336

},

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

337

"languageCode": "A String", # Required. The language of the supplied audio as a

338

# [BCP-47](https://www.rfc-editor.org/rfc/bcp/bcp47.txt) language tag.

339

# Example: "en-US".

340

# See [Language

341

# Support](https://cloud.google.com/speech-to-text/docs/languages) for a list

342

# of the currently supported language codes.

343

"profanityFilter": True or False, # If set to `true`, the server will attempt to filter out

344

# profanities, replacing all but the initial character in each filtered word

345

# with asterisks, e.g. "f***". If set to `false` or omitted, profanities

346

# won't be filtered out.

347

"enableSpeakerDiarization": True or False, # If 'true', enables speaker detection for each recognized word in

348

# the top alternative of the recognition result using a speaker_tag provided

349

# in the WordInfo.

350

# Note: Use diarization_config instead.

351

"useEnhanced": True or False, # Set to true to use an enhanced model for speech recognition.

352

# If `use_enhanced` is set to true and the `model` field is not set, then

353

# an appropriate enhanced model is chosen if an enhanced model exists for

354

# the audio.

355

#

356

# If `use_enhanced` is true and an enhanced version of the specified model

357

# does not exist, then the speech is recognized using the standard version

358

# of the specified model.

359

},

360

"audio": { # Contains audio data in the encoding specified in the `RecognitionConfig`. # Required. The audio data to be recognized.

361

# Either `content` or `uri` must be supplied. Supplying both or neither

362

# returns google.rpc.Code.INVALID_ARGUMENT. See

363

# [content limits](https://cloud.google.com/speech-to-text/quotas#content).

364

"content": "A String", # The audio data bytes encoded as specified in

365

# `RecognitionConfig`. Note: as with all bytes fields, proto buffers use a

366

# pure binary representation, whereas JSON representations use base64.

367

"uri": "A String", # URI that points to a file that contains audio data bytes as specified in

368

# `RecognitionConfig`. The file must not be compressed (for example, gzip).

369

# Currently, only Google Cloud Storage URIs are

370

# supported, which must be specified in the following format:

371

# `gs://bucket_name/object_name` (other URI formats return

372

# google.rpc.Code.INVALID_ARGUMENT). For more information, see

373

# [Request URIs](https://cloud.google.com/storage/docs/reference-uris).

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

},

}

x__xgafv: string, V1 error format.

Allowed values

1 - v1 error format

2 - v2 error format

Returns:

An object of the form:

384

385

{ # This resource represents a long-running operation that is the result of a

386

# network API call.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

387

"name": "A String", # The server-assigned name, which is only unique within the same service that

388

# originally returns it. If you use the default HTTP mapping, the

389

# `name` should be a resource name ending with `operations/{unique_id}`.

390

"error": { # The `Status` type defines a logical error model that is suitable for # The error result of the operation in case of failure or cancellation.

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

391

# different programming environments, including REST APIs and RPC APIs. It is

392

# used by [gRPC](https://github.com/grpc). Each `Status` message contains

393

# three pieces of data: error code, error message, and error details.

394

#

395

# You can find out more about this error model and how to work with it in the

396

# [API Design Guide](https://cloud.google.com/apis/design/errors).

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

397

"code": 42, # The status code, which should be an enum value of google.rpc.Code.

398

"message": "A String", # A developer-facing error message, which should be in English. Any

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

399

# user-facing error message should be localized and sent in the

400

# google.rpc.Status.details field, or localized by the client.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

401

"details": [ # A list of messages that carry the error details. There is a common set of

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

402

# message types for APIs to use.

403

{

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

404

"a_key": "", # Properties of the object. Contains field @type with type URL.

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

405

},

406

],

407

},

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

408

"metadata": { # Service-specific metadata associated with the operation. It typically

409

# contains progress information and common metadata such as create time.

410

# Some services might not provide such metadata. Any method that returns a

411

# long-running operation should document the metadata type, if any.

412

"a_key": "", # Properties of the object. Contains field @type with type URL.

413

},

414

"done": True or False, # If the value is `false`, it means the operation is still in progress.

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

415

# If `true`, the operation is completed, and either `error` or `response` is

416

# available.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

417

"response": { # The normal response of the operation in case of success. If the original

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

418

# method returns no data on success, such as `Delete`, the response is

419

# `google.protobuf.Empty`. If the original method is standard

420

# `Get`/`Create`/`Update`, the response should be the resource. For other

421

# methods, the response should have the type `XxxResponse`, where `Xxx`

422

# is the original method name. For example, if the original method name

423

# is `TakeSnapshot()`, the inferred response type is

424

# `TakeSnapshotResponse`.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

425

"a_key": "", # Properties of the object. Contains field @type with type URL.

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

426

},

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

}</pre>

</div>

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

431

<code class="details" id="recognize">recognize(body=None, x__xgafv=None)</code>

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

432

<pre>Performs synchronous speech recognition: receive results after all audio

433

has been sent and processed.

434

435

Args:

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

436

body: object, The request body.

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

437

The object takes the form of:

438

439

{ # The top-level message sent by the client for the `Recognize` method.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

440

"config": { # Provides information to the recognizer that specifies how to process the # Required. Provides information to the recognizer that specifies how to

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

441

# process the request.

442

# request.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

443

"metadata": { # Description of audio data to be recognized. # Metadata regarding this request.

444

"originalMediaType": "A String", # The original media the speech was recorded on.

445

"obfuscatedId": "A String", # Obfuscated (privacy-protected) ID of the user, to identify number of

446

# unique users using the service.

447

"recordingDeviceType": "A String", # The type of device the speech was recorded with.

448

"interactionType": "A String", # The use case most closely describing the audio content to be recognized.

449

"recordingDeviceName": "A String", # The device used to make the recording. Examples 'Nexus 5X' or

450

# 'Polycom SoundStation IP 6000' or 'POTS' or 'VoIP' or

451

# 'Cardioid Microphone'.

452

"originalMimeType": "A String", # Mime type of the original audio file. For example `audio/m4a`,

453

# `audio/x-alaw-basic`, `audio/mp3`, `audio/3gpp`.

454

# A list of possible audio mime types is maintained at

455

# http://www.iana.org/assignments/media-types/media-types.xhtml#audio

456

"audioTopic": "A String", # Description of the content. Eg. "Recordings of federal supreme court

457

# hearings from 2012".

458

"industryNaicsCodeOfAudio": 42, # The industry vertical to which this speech recognition request most

459

# closely applies. This is most indicative of the topics contained

460

# in the audio. Use the 6-digit NAICS code to identify the industry

461

# vertical - see https://www.naics.com/search/.

462

"microphoneDistance": "A String", # The audio type that most closely describes the audio being recognized.

463

},

464

"sampleRateHertz": 42, # Sample rate in Hertz of the audio data sent in all

465

# `RecognitionAudio` messages. Valid values are: 8000-48000.

466

# 16000 is optimal. For best results, set the sampling rate of the audio

467

# source to 16000 Hz. If that's not possible, use the native sample rate of

468

# the audio source (instead of re-sampling).

469

# This field is optional for FLAC and WAV audio files, but is

470

# required for all other audio formats. For details, see AudioEncoding.

471

"enableSeparateRecognitionPerChannel": True or False, # This needs to be set to `true` explicitly and `audio_channel_count` > 1

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

472

# to get each channel recognized separately. The recognition result will

473

# contain a `channel_tag` field to state which channel that result belongs

474

# to. If this is not true, we will only recognize the first channel. The

475

# request is billed cumulatively for all channels recognized:

476

# `audio_channel_count` multiplied by the length of the audio.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

477

"enableAutomaticPunctuation": True or False, # If 'true', adds punctuation to recognition result hypotheses.

478

# This feature is only available in select languages. Setting this for

479

# requests in other languages has no effect at all.

480

# The default 'false' value does not add punctuation to result hypotheses.

481

"adaptation": { # Speech adaptation configuration. # Speech adaptation configuration improves the accuracy of speech

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

482

# recognition. When speech adaptation is set it supersedes the

483

# `speech_contexts` field. For more information, see the [speech

484

# adaptation](https://cloud.google.com/speech-to-text/docs/context-strength)

485

# documentation.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

486

"phraseSets": [ # A collection of phrase sets. To specify the hints inline, leave the

487

# phrase set's `name` blank and fill in the rest of its fields. Any

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

488

# phrase set can use any custom class.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

489

{ # Provides "hints" to the speech recognizer to favor specific words and phrases

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

490

# in the results.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

491

"boost": 3.14, # Hint Boost. Positive value will increase the probability that a specific

492

# phrase will be recognized over other similar sounding phrases. The higher

493

# the boost, the higher the chance of false positive recognition as well.

494

# Negative boost values would correspond to anti-biasing. Anti-biasing is not

495

# enabled, so negative boost will simply be ignored. Though `boost` can

496

# accept a wide range of positive values, most use cases are best served with

497

# values between 0 (exclusive) and 20. We recommend using a binary search

498

# approach to finding the optimal value for your use case. Speech recognition

499

# will skip PhraseSets with a boost value of 0.

500

"name": "A String", # The resource name of the phrase set.

501

"phrases": [ # A list of word and phrases.

502

{ # A phrases containing words and phrase "hints" so that

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

503

# the speech recognition is more likely to recognize them. This can be used

504

# to improve the accuracy for specific words and phrases, for example, if

505

# specific commands are typically spoken by the user. This can also be used

506

# to add additional words to the vocabulary of the recognizer. See

507

# [usage limits](https://cloud.google.com/speech-to-text/quotas#content).

508

#

509

# List items can also include pre-built or custom classes containing groups

510

# of words that represent common concepts that occur in natural language. For

511

# example, rather than providing a phrase hint for every month of the

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

512

# year (e.g. "i was born in january", "i was born in febuary", ...), use the

513

# pre-built `$MONTH` class improves the likelihood of correctly transcribing

514

# audio that includes months (e.g. "i was born in $month").

515

# To refer to pre-built classes, use the class' symbol prepended with `$`

516

# e.g. `$MONTH`. To refer to custom classes that were defined inline in the

517

# request, set the class's `custom_class_id` to a string unique to all class

518

# resources and inline classes. Then use the class' id wrapped in $`{...}`

519

# e.g. "${my-months}". To refer to custom classes resources, use the class'

520

# id wrapped in `${}` (e.g. `${my-months}`).

521

"value": "A String", # The phrase itself.

522

"boost": 3.14, # Hint Boost. Overrides the boost set at the phrase set level.

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

523

# Positive value will increase the probability that a specific phrase will

524

# be recognized over other similar sounding phrases. The higher the boost,

525

# the higher the chance of false positive recognition as well. Negative

526

# boost values would correspond to anti-biasing. Anti-biasing is not

527

# enabled, so negative boost will simply be ignored. Though `boost` can

528

# accept a wide range of positive values, most use cases are best served

529

# with values between 0 and 20. We recommend using a binary search approach

530

# to finding the optimal value for your use case. Speech recognition

531

# will skip PhraseSets with a boost value of 0.

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

532

},

533

],

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

534

},

535

],

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

536

"customClasses": [ # A collection of custom classes. To specify the classes inline, leave the

537

# class' `name` blank and fill in the rest of its fields, giving it a unique

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

538

# `custom_class_id`. Refer to the inline defined class in phrase hints by its

539

# `custom_class_id`.

540

{ # A set of words or phrases that represents a common concept likely to appear

541

# in your audio, for example a list of passenger ship names. CustomClass items

542

# can be substituted into placeholders that you set in PhraseSet phrases.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

543

"name": "A String", # The resource name of the custom class.

544

"customClassId": "A String", # If this custom class is a resource, the custom_class_id is the resource id

545

# of the CustomClass. Case sensitive.

546

"items": [ # A collection of class items.

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

547

{ # An item of the class.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

548

"value": "A String", # The class item's value.

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

549

},

550

],

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

551

},

552

],

553

},

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

554

"maxAlternatives": 42, # Maximum number of recognition hypotheses to be returned.

555

# Specifically, the maximum number of `SpeechRecognitionAlternative` messages

556

# within each `SpeechRecognitionResult`.

557

# The server may return fewer than `max_alternatives`.

558

# Valid values are `0`-`30`. A value of `0` or `1` will return a maximum of

559

# one. If omitted, will return a maximum of one.

560

"diarizationSpeakerCount": 42, # If set, specifies the estimated number of speakers in the conversation.

561

# Defaults to '2'. Ignored unless enable_speaker_diarization is set to true.

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

562

# Note: Use diarization_config instead.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

563

"encoding": "A String", # Encoding of audio data sent in all `RecognitionAudio` messages.

564

# This field is optional for `FLAC` and `WAV` audio files and required

565

# for all other audio formats. For details, see AudioEncoding.

566

"speechContexts": [ # Array of SpeechContext.

567

# A means to provide context to assist the speech recognition. For more

568

# information, see

569

# [speech

570

# adaptation](https://cloud.google.com/speech-to-text/docs/context-strength).

571

{ # Provides "hints" to the speech recognizer to favor specific words and phrases

572

# in the results.

573

"phrases": [ # A list of strings containing words and phrases "hints" so that

574

# the speech recognition is more likely to recognize them. This can be used

575

# to improve the accuracy for specific words and phrases, for example, if

576

# specific commands are typically spoken by the user. This can also be used

577

# to add additional words to the vocabulary of the recognizer. See

578

# [usage limits](https://cloud.google.com/speech-to-text/quotas#content).

579

#

580

# List items can also be set to classes for groups of words that represent

581

# common concepts that occur in natural language. For example, rather than

582

# providing phrase hints for every month of the year, using the $MONTH class

583

# improves the likelihood of correctly transcribing audio that includes

# months.

"A String",

],

"boost": 3.14, # Hint Boost. Positive value will increase the probability that a specific

588

# phrase will be recognized over other similar sounding phrases. The higher

589

# the boost, the higher the chance of false positive recognition as well.

590

# Negative boost values would correspond to anti-biasing. Anti-biasing is not

591

# enabled, so negative boost will simply be ignored. Though `boost` can

592

# accept a wide range of positive values, most use cases are best served with

593

# values between 0 and 20. We recommend using a binary search approach to

594

# finding the optimal value for your use case.

595

},

596

],

597

"enableWordConfidence": True or False, # If `true`, the top result includes a list of words and the

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

598

# confidence for those words. If `false`, no word-level confidence

599

# information is returned. The default is `false`.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

600

"model": "A String", # Which model to select for the given request. Select the model

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

601

# best suited to your domain to get best results. If a model is not

602

# explicitly specified, then we auto-select a model based on the parameters

603

# in the RecognitionConfig.

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

604

# <table>

605

# <tr>

606

# <td><b>Model</b></td>

607

# <td><b>Description</b></td>

608

# </tr>

609

# <tr>

610

# <td><code>command_and_search</code></td>

611

# <td>Best for short queries such as voice commands or voice search.</td>

612

# </tr>

613

# <tr>

614

# <td><code>phone_call</code></td>

615

# <td>Best for audio that originated from a phone call (typically

616

# recorded at an 8khz sampling rate).</td>

617

# </tr>

618

# <tr>

619

# <td><code>video</code></td>

620

# <td>Best for audio that originated from from video or includes multiple

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

621

# speakers. Ideally the audio is recorded at a 16khz or greater

622

# sampling rate. This is a premium model that costs more than the

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

623

# standard rate.</td>

624

# </tr>

625

# <tr>

626

# <td><code>default</code></td>

627

# <td>Best for audio that is not one of the specific audio models.

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

628

# For example, long-form audio. Ideally the audio is high-fidelity,

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

629

# recorded at a 16khz or greater sampling rate.</td>

630

# </tr>

631

# </table>

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

632

"audioChannelCount": 42, # The number of channels in the input audio data.

633

# ONLY set this for MULTI-CHANNEL recognition.

634

# Valid values for LINEAR16 and FLAC are `1`-`8`.

635

# Valid values for OGG_OPUS are '1'-'254'.

636

# Valid value for MULAW, AMR, AMR_WB and SPEEX_WITH_HEADER_BYTE is only `1`.

637

# If `0` or omitted, defaults to one channel (mono).

638

# Note: We only recognize the first channel by default.

639

# To perform independent recognition on each channel set

640

# `enable_separate_recognition_per_channel` to 'true'.

641

"enableWordTimeOffsets": True or False, # If `true`, the top result includes a list of words and

642

# the start and end time offsets (timestamps) for those words. If

643

# `false`, no word-level time offset information is returned. The default is

644

# `false`.

645

"alternativeLanguageCodes": [ # A list of up to 3 additional

646

# [BCP-47](https://www.rfc-editor.org/rfc/bcp/bcp47.txt) language tags,

647

# listing possible alternative languages of the supplied audio.

648

# See [Language

649

# Support](https://cloud.google.com/speech-to-text/docs/languages) for a list

650

# of the currently supported language codes. If alternative languages are

651

# listed, recognition result will contain recognition in the most likely

652

# language detected including the main language_code. The recognition result

653

# will include the language tag of the language detected in the audio. Note:

654

# This feature is only supported for Voice Command and Voice Search use cases

655

# and performance may vary for other use cases (e.g., phone call

# transcription).

"A String",

],

"diarizationConfig": { # Config to enable speaker diarization. # Config to enable speaker diarization and set additional

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

660

# parameters to make diarization better suited for your application.

661

# Note: When this is enabled, we send all the words from the beginning of the

662

# audio for the top alternative in every consecutive STREAMING responses.

663

# This is done in order to improve our speaker tags as our models learn to

664

# identify the speakers in the conversation over time.

665

# For non-streaming requests, the diarization results will be provided only

666

# in the top alternative of the FINAL SpeechRecognitionResult.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

667

"minSpeakerCount": 42, # Minimum number of speakers in the conversation. This range gives you more

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

668

# flexibility by allowing the system to automatically determine the correct

669

# number of speakers. If not set, the default value is 2.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

670

"maxSpeakerCount": 42, # Maximum number of speakers in the conversation. This range gives you more

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

671

# flexibility by allowing the system to automatically determine the correct

672

# number of speakers. If not set, the default value is 6.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

673

"speakerTag": 42, # Output only. Unused.

674

"enableSpeakerDiarization": True or False, # If 'true', enables speaker detection for each recognized word in

675

# the top alternative of the recognition result using a speaker_tag provided

676

# in the WordInfo.

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

677

},

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

678

"languageCode": "A String", # Required. The language of the supplied audio as a

679

# [BCP-47](https://www.rfc-editor.org/rfc/bcp/bcp47.txt) language tag.

680

# Example: "en-US".

681

# See [Language

682

# Support](https://cloud.google.com/speech-to-text/docs/languages) for a list

683

# of the currently supported language codes.

684

"profanityFilter": True or False, # If set to `true`, the server will attempt to filter out

685

# profanities, replacing all but the initial character in each filtered word

686

# with asterisks, e.g. "f***". If set to `false` or omitted, profanities

687

# won't be filtered out.

688

"enableSpeakerDiarization": True or False, # If 'true', enables speaker detection for each recognized word in

689

# the top alternative of the recognition result using a speaker_tag provided

690

# in the WordInfo.

691

# Note: Use diarization_config instead.

692

"useEnhanced": True or False, # Set to true to use an enhanced model for speech recognition.

693

# If `use_enhanced` is set to true and the `model` field is not set, then

694

# an appropriate enhanced model is chosen if an enhanced model exists for

695

# the audio.

696

#

697

# If `use_enhanced` is true and an enhanced version of the specified model

698

# does not exist, then the speech is recognized using the standard version

699

# of the specified model.

700

},

701

"audio": { # Contains audio data in the encoding specified in the `RecognitionConfig`. # Required. The audio data to be recognized.

702

# Either `content` or `uri` must be supplied. Supplying both or neither

703

# returns google.rpc.Code.INVALID_ARGUMENT. See

704

# [content limits](https://cloud.google.com/speech-to-text/quotas#content).

705

"content": "A String", # The audio data bytes encoded as specified in

706

# `RecognitionConfig`. Note: as with all bytes fields, proto buffers use a

707

# pure binary representation, whereas JSON representations use base64.

708

"uri": "A String", # URI that points to a file that contains audio data bytes as specified in

709

# `RecognitionConfig`. The file must not be compressed (for example, gzip).

710

# Currently, only Google Cloud Storage URIs are

711

# supported, which must be specified in the following format:

712

# `gs://bucket_name/object_name` (other URI formats return

713

# google.rpc.Code.INVALID_ARGUMENT). For more information, see

714

# [Request URIs](https://cloud.google.com/storage/docs/reference-uris).

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

715

},

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

716

}

717

718

x__xgafv: string, V1 error format.

Allowed values

1 - v1 error format

2 - v2 error format

Returns:

An object of the form:

725

726

{ # The only message returned to the client by the `Recognize` method. It

727

# contains the result as zero or more sequential `SpeechRecognitionResult`

728

# messages.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

729

"results": [ # Sequential list of transcription results corresponding to

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

730

# sequential portions of audio.

731

{ # A speech recognition result corresponding to a portion of the audio.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

732

"channelTag": 42, # For multi-channel audio, this is the channel number corresponding to the

Dan O'Meara

2020-05-01 07:42:23 -0700

[diff] [blame]

733

# recognized result for the audio from that channel.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

734

# For audio_channel_count = N, its output values can range from '1' to 'N'.

735

"languageCode": "A String", # Output only. The [BCP-47](https://www.rfc-editor.org/rfc/bcp/bcp47.txt) language tag

736

# of the language in this result. This language code was detected to have

737

# the most likelihood of being spoken in the audio.

738

"alternatives": [ # May contain one or more recognition hypotheses (up to the

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

739

# maximum specified in `max_alternatives`).

740

# These alternatives are ordered in terms of accuracy, with the top (first)

741

# alternative being the most probable, as ranked by the recognizer.

742

{ # Alternative hypotheses (a.k.a. n-best list).

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

743

"confidence": 3.14, # The confidence estimate between 0.0 and 1.0. A higher number

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

744

# indicates an estimated greater likelihood that the recognized words are

745

# correct. This field is set only for the top alternative of a non-streaming

746

# result or, of a streaming result where `is_final=true`.

747

# This field is not guaranteed to be accurate and users should not rely on it

748

# to be always provided.

749

# The default of 0.0 is a sentinel value indicating `confidence` was not set.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

750

"transcript": "A String", # Transcript text representing the words that the user spoke.

751

"words": [ # A list of word-specific information for each recognized word.

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

752

# Note: When `enable_speaker_diarization` is true, you will see all the words

753

# from the beginning of the audio.

754

{ # Word-specific information for recognized words.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

755

"speakerTag": 42, # Output only. A distinct integer value is assigned for every speaker within

756

# the audio. This field specifies which one of those speakers was detected to

757

# have spoken this word. Value ranges from '1' to diarization_speaker_count.

758

# speaker_tag is set if enable_speaker_diarization = 'true' and only in the

759

# top alternative.

760

"endTime": "A String", # Time offset relative to the beginning of the audio,

761

# and corresponding to the end of the spoken word.

762

# This field is only set if `enable_word_time_offsets=true` and only

763

# in the top hypothesis.

764

# This is an experimental feature and the accuracy of the time offset can

765

# vary.

766

"confidence": 3.14, # The confidence estimate between 0.0 and 1.0. A higher number

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

767

# indicates an estimated greater likelihood that the recognized words are

768

# correct. This field is set only for the top alternative of a non-streaming

769

# result or, of a streaming result where `is_final=true`.

770

# This field is not guaranteed to be accurate and users should not rely on it

771

# to be always provided.

772

# The default of 0.0 is a sentinel value indicating `confidence` was not set.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

773

"startTime": "A String", # Time offset relative to the beginning of the audio,

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

774

# and corresponding to the start of the spoken word.

775

# This field is only set if `enable_word_time_offsets=true` and only

776

# in the top hypothesis.

777

# This is an experimental feature and the accuracy of the time offset can

778

# vary.

Bu Sun Kim

2020-05-20 12:08:20 -0700

[diff] [blame]

779

"word": "A String", # The word corresponding to this set of information.

Bu Sun Kim

2019-06-14 16:50:42 -0700

[diff] [blame]

},

],

},

],

Bu Sun Kim