IBM Speech to Text

v0.3.0

IBM Speech to Text v0.3.0 IBMSpeechToText.Message.Start View Source

Message sent over WebSocket before audio to be recognized

Allows to configure recognition process. The options are documented in IBM API docs

If an option is set to nil it won't be included in a message and thus the default value for recognition will be used.

Link to this section Summary

Types

channels()
content_type()
sample_rate()
t()

Link to this section Types

Link to this type

channels() View Source 
channels() :: pos_integer()

Link to this type

content_type() View Source 
content_type() ::
  {:alaw, sample_rate()}
  | {:basic, sample_rate()}
  | :flac
  | :g729
  | {:l16, sample_rate()}
  | {:l16, sample_rate(), channels()}
  | :mp3
  | :mpeg
  | :mulaw
  | :ogg
  | {:ogg, :opus}
  | {:ogg, :vorbis}
  | :wav
  | :webm
  | {:webm, :ogg}
  | {:webm, :vorbis}

Link to this type

sample_rate() View Source 
sample_rate() :: pos_integer()

Link to this type

t() View Source 
t() :: %IBMSpeechToText.Message.Start{
  content_type: content_type() | nil,
  customization_weight: term(),
  grammar_name: term(),
  inactivity_timeout: integer() | nil,
  interim_results: boolean() | nil,
  keywords: [String.t()] | nil,
  keywords_threshold: float() | nil,
  max_alternatives: non_neg_integer() | nil,
  profanity_filter: boolean() | nil,
  redaction: boolean() | nil,
  smart_formatting: boolean() | nil,
  speaker_labels: boolean() | nil,
  timestamps: boolean() | nil,
  word_alternatives_threshold: float() | nil,
  word_confidence: boolean() | nil
}