Speech-to-text visual feedback using interim symbols
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current speech-to-text technologies in computing devices often cause users to pause mid-sentence due to transcription errors, leading to more errors and inefficient processing, as they display near-real-time text conversions which can be distracting and disrupt natural speech patterns.
Innovation Solution
The method involves receiving audio input, determining a series of words, and initially outputting substituted symbols on the screen instead of alphanumeric characters, which correspond to the words, and only replacing these symbols with text once the audio input has completed, allowing for natural speech and reducing transcription errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If near-real-time text conversion is displayed on the screen during speech input, then the user receives immediate feedback, but the user becomes distracted and pauses mid-sentence which causes more transcription errors
Solution Approach 1:
The patent introduces substituted symbols (e.g., asterisks or bullet points) as an intermediary representation between the audio input and the final transcription text. These symbols provide visual feedback that speech is being processed without displaying the actual transcribed words, thereby maintaining natural speech flow while still offering confirmation to the user.
Solution Approach 2:
The system performs preliminary speech-to-text conversion during the speech input phase and stores the intermediate results. The actual display of the transcribed text is delayed until after the speech input completes, allowing the user to speak naturally without real-time visual distraction while still enabling post-processing and correction.
2Loss of information
If real-time transcription feedback is provided to the user, then the user can monitor the transcription process, but the user pauses to correct errors which disrupts natural speech patterns and increases overall errors
Solution Approach 1:
The system performs the speech-to-text conversion in advance during the speech input phase, storing intermediate results. The display of transcription text is delayed until after speech completion, allowing error detection to occur post-speech rather than interrupting the speech flow, thereby maintaining productivity while still enabling error detection.
Solution Approach 2:
The transcription process is segmented into distinct phases: (1) speech input with substituted symbol display, (2) post-speech text display, and (3) optional correction phase. This segmentation allows the system to provide error detection capability without requiring continuous real-time feedback that would interrupt speech flow.
3Loss of time
If the computing device displays transcribed words immediately during speech input, then the user can see the conversion in progress, but the user becomes distracted and may pause speaking which causes additional transcription errors
Solution Approach 1:
Substituted symbols serve as an intermediary visual feedback mechanism that confirms speech is being processed without revealing the actual transcribed content. This maintains rapid feedback response time (principle of immediate acknowledgment) while preventing the distraction that would compromise transcription accuracy.
Solution Approach 2:
The system prepares and stores intermediate transcription results during speech input but delays their display until after speech completion. This preliminary processing ensures rapid response readiness while the delayed display prevents accuracy-compromising distractions.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A computing device receives an audio input from a user. The computing device determines a series of words from the audio input. The computing device outputs, for display, one or more substituted symbols. The one or more substituted symbols correspond to at least a portion of the series of words. In response to determining that receipt of the audio input has completed, the computing device outputs, for display, alphanumeric characters comprising the series of words in place of the one or more substituted symbols.