Mechanism for Providing User Guidance and Latency Concealment for Automatic Speech Recognition Systems
AssigneePromptu Systems Corporation
InventorLaurent Charriere
Life of this patent
Estimating the term…
Abstract
Audio input to a user device is captured in a buffer and played back to the user while being sent to and recognized by an automatic speech recognition (ASR) system. Overlapping the playback with the speech recognition processing masks a portion of the true latency of the ASR system thus improving the user's perception of the ASR system's responsiveness. Further, upon hearing the playback, the user is intuitively guided to self-correct for any defects in the captured audio.
Claims16
An automatic speech recognition method comprising:
- receiving from a user of a mobile phone a signal to start capturing audio;
- begin capturing audio in a buffer of the mobile phone in response to the received signal to start capturing audio;
- receiving from the user of the mobile phone a signal to stop capturing audio;
- end capturing audio in the buffer of the mobile phone in response to the received signal to stop capturing audio;
- and performing the following steps in sequence:
- in a first step, begin sending the captured audio from the buffer of the mobile phone to an automatic speech recognition system of a server located across a network from the mobile phone;
- in a second step, begin playing back the captured audio from the buffer of the mobile phone;
- in a third step, receiving at the mobile phone a recognized version of the captured audio from the automatic speech recognition system of the server located across the network from the mobile phone;
- and
- in a fourth step, rendering at the mobile phone the, recognized version of the captured audio.
An apparatus comprising:
- a microphone configured to receive audio from a user of the mobile phone;
- a buffer configured to store the received audio;
- an utterance gating control configured to start and stop the buffer storing the received audio;
- a loudspeaker;
- a processor configured to perform the following steps in sequence:
- in a first step, begin sending the stored audio across a network to a automatic speech recognition system,
- in a second step, after waiting a predefined period of time after the utterance gating control has stopped the buffer storing the received audio, control play back of the stored audio through the loudspeaker;
- in a third step, receive across the network from the automatic speech recognition system a recognized version of the sent audio, and
- in a fourth step, render at the mobile phone the received recognized version of the sent audio.
A computer readable storage medium having embodied thereon a program, the program executable by a processor to perform a method for automatic speech recognition, the method comprising:
- receiving from a user of a mobile phone a signal to start capturing audio;
- begin capturing audio in a buffer of the mobile phone in response to the received signal to start capturing audio;
- receiving from the user of the mobile phone a signal to stop capturing audio;
- end capturing audio in the buffer of the mobile phone in response to the received signal to stop capturing audio;
- and performing the following steps in sequence:
- in a first step, begin sending the captured audio from the buffer of the mobile phone to an automatic speech recognition system of a server located across a network from the mobile phone;
- in a second step, begin playing back the captured audio from the buffer of the mobile phone;
- in a third step, receiving at the mobile phone a recognized version of the captured audio from the automatic speech recognition system of the server located across the network from the mobile phone;
- and
- in a fourth step, rendering at the mobile phone the recognized version of the captured audio.
Dependencies are inferred from the claim text.