Cubic Protobuf API Docs

Unique identifier of the model to compile the context information for. The model chosen needs to support context which can be verified by checking its ModelAttributes.ContextInfo obtained via ListModels.

The token that is associated with the provided list of phrases or words (e.g “menu_item”, “airport” etc.). Must be one of the tokens included in the model being used, which can be retrieved by calling the ListModels method.

List of phrases and/or words to be compiled.

Context information in a compact form that is efficient for use in subsequent recognition requests. The size of the compiled form will depend on the amount of text that was sent for compilation. For 1000 words it’s generally less than 100 kilobytes.

The context information compiled by the CompileContext method.

Word in the recognized transcript

Confidence estimate between 0 and 1. A higher number represents a higher likelihood that the word was correctly recognized.

Time offset relative to the beginning of audio received by the recognizer and corresponding to the start of this link

Duration of the current link in the confusion network

Arcs between this link

If this is set to true, the model supports taking context information into account to aid speech recognition. The information may be sent with with recognition requests via RecognitionContext inside RecognitionConfig.

A list of tokens (e.g “name”, “airport” etc.) that serve has placeholders in the model where a client provided list of phrases or words may be used to aid speech recognition and produce the exact desired recognition output.

The actual phrase or word.

This is an optional field. The boost value is a positive number which is used to increase the probability of the phrase or word appearing in the output. This setting can be used to differentiate between similar sounding words, with the desired word given a bigger boost value.

By default, all phrases or words are given an equal probability of 1/N (where N = total number of phrases or words). If a boost value is provided, the new probability is (boost + 1) * 1/N. We normalize the boosted probabilities for all the phrases or words so that they sum to one. This means that the boost value only has an effect if there are relative differences in the values for different phrases or words. That is, if all phrases or words have the same boost value, after normalization they will all still have the same probability. This also means that the boost value can be any positive value, but it is best to stick between 0 to 20.

Negative values are not supported and will be treated as 0 values.

List of models available for use that match the request.

Unique identifier of the model. This identifier is used to choose the model that should be used for recognition, and is specified in the RecognitionConfig message.

Model name. This is a concise name describing the model, and maybe presented to the end-user, for example, to help choose which model to use for their recognition task.

Model attributes

Audio sample rate supported by the model

Attributes specifc to supporting recognition context.

Text representing the transcription of the words that the user spoke.

The transcript will be formatted according to the servers formatting configuration. If you want the raw transcript, please see the field raw_transcript. If the server is configured to not use any formatting, then this field will contain the raw transcript.

As an example, if the spoken utterance was “four people”, and the server was configured to format numbers, this field would be set to “4 people”.

Text representing the transcription of the words that the user spoke, without any formatting. This field will be populated only the config RecognitionConfig.enable_raw_transcript is set to true. Otherwise this field will be an empty string. If you want the formatted transcript, please see the field transcript.

As an example, if the spoken utterance was here are four words, this field would be set to “HERE ARE FOUR WORDS”.

Confidence estimate between 0 and 1. A higher number represents a higher likelihood of the output being correct.

A list of word-specific information for each recognized word in the transcript field. This is available only if enable_word_confidence or enable_word_time_offsets was set to true in the RecognitionConfig.

A list of word-specific information for each recognized word in the raw_transcript field. This is available only if enable_word_confidence or enable_word_time_offsets was set to true and enable_raw_transcript is also set to true in the RecognitionConfig.

Time offset relative to the beginning of audio received by the recognizer and corresponding to the start of this utterance.

Duration of the current utterance in the spoken audio.

Unique identifier of the model to use, as obtained from a Model message.

Encoding of audio data sent/streamed through the RecognitionAudio messages. For encodings like WAV/MP3 that have headers, the headers are expected to be sent at the beginning of the stream, not in every RecognitionAudio message.

If not specified, the default encoding is RAW_LINEAR16.

Depending on how they are configured, server instances of this service may not support all the encodings enumerated above. They are always required to accept RAW_LINEAR16. If any other Encoding is specified, and it is not available on the server being used, the recognition request will result in an appropriate error message.

Idle Timeout of the created Recognizer. If no audio data is received by the recognizer for this duration, ongoing rpc calls will result in an error, the recognizer will be destroyed and thus more audio may not be sent to the same recognizer. The server may impose a limit on the maximum idle timeout that can be specified, and if the value in this message exceeds that serverside value, creating of the recognizer will fail with an error.

This is an optional field. If this is set to true, each result will include a list of words and the start time offset (timestamp) and the duration for each of those words. If set to false, no word-level timestamps will be returned. The default is false.

This is an optional field. If this is set to true, each result will include a list of words and the confidence for those words. If false, no word-level confidence information is returned. The default is false.

This is an optional field. If this is set to true, the field RecognitionAlternative.raw_transcript will be populated with the raw transcripts output from the recognizer will be exposed without any formatting rules applied. If this is set to false, that field will not be set in the results. The RecognitionAlternative.transcript will always be populated with text formatted according to the server’s settings.

This is an optional field. If this is set to true, the results will include a confusion network. If set to false, no confusion network will be returned. The default is false. If the model being used does not support a confusion network, results may be returned without a confusion network available. If this field is set to true, then enable_raw_transcript is also forced to be true.

This is an optional field. If the audio has multiple channels, this field should be configured with the list of channel indices that should be transcribed. Channels are 0-indexed.

Example: [0] for a mono file, [0, 1] for a stereo file.

If this field is not set, a mono file will be assumed by default and only channel-0 will be transcribed even if the file actually has additional channels.

Channels that are present in the audio may be omitted, but it is an error to include a channel index in this field that is not present in the audio. Channels may be listed in any order but the same index may not be repeated in this list.

BAD: [0, 2] for a stereo file; BAD: [0, 0] for a mono file.

This is an optional field. If there is any metadata associated with the audio being sent, use this field to provide it to cubic. The server may record this metadata when processing the request. The server does not use this field for any other purpose.

This is an optional field for providing any additional context information that may aid speech recognition. This can also be used to add out-of-vocabulary words to the model or boost recognition of specific proper names or commands. Context information must be pre-compiled via the CompileContext() method.

List of compiled context information, with each entry being compiled from a list of words or phrases using the CompileContext method.

Any custom metadata that the client wants to associate with the recording. This could be a simple string (e.g. a tracing ID) or structured data (e.g. JSON)

An n-best list of recognition hypotheses alternatives

If this is set to true, it denotes that the result is an interim partial result, and could change after more audio is processed. If unset, or set to false, it denotes that this is a final result and will not change.

Servers are not required to implement support for returning partial results, and clients should generally not depend on their availability.

If enable_confusion_network was set to true in the RecognitionConfig, and if the model supports it, a confusion network will be available in the results.

Channel of the audio file that this result was transcribed from. For a mono file, or RAW_LINEAR16 input, this will be set to 0.

Provides configuration to create the recognizer.

The audio data to be recognized

version of the cubic library handling the recognition

version of the server handling these requests

The actual word in the text

Confidence estimate between 0 and 1. A higher number represents a higher likelihood that the word was correctly recognized.

Time offset relative to the beginning of audio received by the recognizer and corresponding to the start of this spoken word.

Duration of the current word in the spoken audio.