Page tree
Skip to end of metadata
Go to start of metadata

The Speech transport handler allows the recognition of speech as an input stream. To do so, we use the sphinx4 speech recognition toolkit. There are two transport handler available: Speech which recognize speech as text and NumericSpeech which recognize only numbers and represents the input stream as integer values. The sphinx4 toolkit can use guided grammar or a language model (N-Gram Language Model). Thus, you can define the language your transport should understand.

Options

  • config: The path to the sphinx4 configuration file. See cmusphinx.sourceforge.net/wiki/ for more information about configuring sphinx. (Default: a configuration with a small vocabulary) (optional)

Example

PQL

Speech Transport Handler
input = ACCESS({source='Source',
wrapper='GenericPush',
transport='Speech',
protocol='Document',
dataHandler='Tuple',
schema=[
['value', 'String']]
})

input = ACCESS({source='Source',
wrapper='GenericPush',
transport='NumericSpeech',
protocol='Document',
dataHandler='Tuple',
schema=[
['value', 'Integer']]
})

CQL

Speech Transport Handler
CREATE STREAM source (VALUE STRING)
    WRAPPER 'GenericPush'
    PROTOCOL 'Document'
    TRANSPORT 'Speech'
    DATAHANDLER 'Tuple'

CREATE STREAM source (VALUE INTEGER)
    WRAPPER 'GenericPush'
    PROTOCOL 'Document'
    TRANSPORT 'NumericSpeech'
    DATAHANDLER 'Tuple'
  • No labels