Utterance Processing

It helps to understand how Utterly Voice processes each utterance.

When the microphone is on, Utterly Voice is monitoring the microphone signal levels. When the signal goes above the configured microphone threshold (see Setup), it starts streaming the microphone data to the recognizer. Once the signal goes below the threshold for a duration of the configured utterance gap milliseconds value (see Setup), it stops streaming microphone data and requests a transcript for the completed utterance.

The completed utterance is processed by the recognizer, which returns a list of possible transcripts that match the audio. Based on your settings, the most likely match is selected. Utterly Voice provides a default recognizer, but you can configure other third party recognizers.

Next, the selected text phrase is sent to the interpreter, which is a component of the Utterly Voice application. The interpreter uses your settings to determine if the phrase contains any commands or whether the text should be typed as-is.

Finally, commands within the utterance are executed and text is typed as dictated by the utterance. The utterance is added to the user interface history area, including results from both the recognizer and interpreter.