Voicemail Transcription Punctuation via Pause Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current voicemail transcription services fail to accurately represent pauses and hesitations in voicemail messages, resulting in unbroken text blocks that are difficult to read due to the treatment of longer pauses and meaningless sounds as short pauses, lacking appropriate punctuation.
Innovation Solution
A method is introduced to punctuate voicemail transcriptions by identifying hesitations such as 'um' and 'ahh' and replacing them with ellipses, or inserting ellipses at the longest pauses, to set a threshold of ellipses per message length, ensuring readable text.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If pauses of all lengths are treated the same in transcription, then the transcription process is simple, but the readability of the transcribed text deteriorates
Solution Approach 1:
The patent applies local quality by treating different pauses differently based on their duration. Short pauses are transcribed without special punctuation while longer pauses (exceeding a threshold duration) are marked with ellipses. This localized differentiation improves readability without complicating the overall transcription process.
2Productivity
If all pauses are transcribed without differentiation, then the transcription is fast and simple, but the meaning and interpretation of the message deteriorates
Solution Approach 1:
The patent changes the parameter of pause duration to determine transcription treatment. By comparing the duration of each pause against a predefined threshold, the system automatically applies different transcription markers. This parameter-based approach maintains high transcription speed while preserving meaningful pause information that aids message interpretation.
3Measurement precision
If ellipses are inserted at every pause, then all pauses are represented, but the text becomes overly punctuated and harder to read
Solution Approach 1:
The patent applies local quality by selectively inserting ellipses only at pauses that exceed a specific duration threshold. This selective approach ensures that only significant pauses are marked, maintaining accurate representation of meaningful pauses while avoiding over-punctuation that would hinder readability.
4Ease of operation
If hesitation sounds are replaced with ellipses, then the transcription becomes more readable, but the number of ellipses may exceed acceptable limits
Solution Approach 1:
The patent initially replaces all hesitation sounds with ellipses (excessive action) to maximize readability, then applies a post-processing step to limit the total number of ellipses to an acceptable threshold. This two-stage approach ensures readability improvement while controlling the quantity of punctuation to maintain text flow.
Data Source
AI summary
A system, method and software product punctuates voicemail transcription text. A transcription text of the voicemail message is generated and the pauses between words of the transcribed text are determined. Ellipses are inserted into the transcription text at the position of “er” and “ahh” type words and pauses between words of the transcribed text.