Audio Playback Latency Masking for Speech Recognition

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Automatic speech recognition systems face challenges in accuracy and responsiveness due to defective or corrupted audio inputs and latency issues, leading to user dissatisfaction.

Innovation Solution

A method and apparatus that capture audio on a mobile device, transmit it to a server for recognition, and simultaneously play back the captured audio locally, masking latency and providing user guidance on improving audio quality by highlighting defects.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If audio is transmitted to a server for recognition, then speech recognition accuracy is improved, but system responsiveness deteriorates due to network latency

Engineering Contradiction:

Improvespeech recognition accuracyVSAvoidsystem responsiveness

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system performs audio playback before the recognition result is available from the server. By playing back the captured audio immediately after capture (while still in the buffer), the system masks the perceived latency. The user hears the audio playback and then receives the recognition result without noticing the network transmission delay, as the playback has already occurred or is occurring concurrently.

Inventive Principle:

Principle #10Preliminary action

2Reliability

If audio capture is continuous, then complete speech is captured, but defective or corrupted audio inputs increase

Engineering Contradiction:

Improvecompleteness of speech captureVSAvoidaudio quality defects

Core Design Contradiction:

ReliabilityVSObject-affected harmful factors

Solution Approach 1:

The system provides feedback to the user by playing back the captured audio before recognition. This allows the user to hear the actual audio quality and understand what was captured. The playback serves as feedback about the audio input quality, enabling the user to adjust their speaking or re-capture if necessary, thereby improving the reliability of the final recognition result.

Inventive Principle:

Principle #23Feedback

3Speed

If audio processing is performed quickly, then system responsiveness is improved, but audio quality analysis becomes insufficient

Engineering Contradiction:

Improveprocessing speedVSAvoidaudio quality assessment

Core Design Contradiction:

SpeedVSMeasurement precision

Solution Approach 1:

The captured audio itself serves as an intermediary medium that provides information about audio quality. By playing back the audio, the system uses the audio signal as a mediator to convey quality information to the user without requiring complex automated quality analysis. The user's perception of the playback quality provides the assessment that would otherwise require sophisticated processing.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS8457963B2Mechanism for providing user guidance and latency concealment for automatic speech recognition systems

Publication Date: 2013.06.04 PROMPTU SYSTEMS CORP

AI summary

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.