Moving Image Editing for Natural Pause Filler Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for deleting pause fillers from moving images often leave the voice intact, causing a sense of strangeness during playback and require significant user effort to identify and edit these sections naturally.
Innovation Solution
An information processing apparatus that analyzes sound information to identify meaningless utterances, detects the speaker, and edits the moving image to delete these utterances, either by muting the sound or cutting the video and sound in the identified sections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If pause fillers are deleted from telop text only, then telop text quality is improved, but the moving image remains unnatural due to intact voice
Solution Approach 1:
The patent combines telop text processing with voice track processing into a unified system. The same pause filler detection mechanism is applied to both telop text and voice audio, ensuring that when a pause filler is identified in the voice track, the corresponding telop text is also updated to reflect this deletion, maintaining synchronization between visual and audio elements.
Solution Approach 2:
The patent introduces a synchronization mechanism that acts as an intermediary between voice processing and telop text processing. This intermediary ensures that deletions in the voice track are properly reflected in the telop text through coordinated timing and position adjustments, preventing desynchronization between the two processing streams.
2Reliability
If manual identification and deletion of pause filler sections is performed, then naturalness of moving image is improved, but time and effort required increase significantly
Solution Approach 1:
The patent implements automatic pause filler detection using acoustic analysis algorithms that independently identify pause filler sections without human intervention. The system analyzes voice tracks, detects characteristic acoustic patterns of pause fillers, and automatically determines deletion sections, performing the editing task autonomously without requiring user input or manual analysis.
Solution Approach 2:
The patent performs preliminary acoustic analysis and pause filler identification before the actual deletion operation. The system first analyzes the voice track to identify all pause filler sections, marks these sections for deletion, and then executes the deletion in a subsequent step. This preliminary identification action eliminates the need for manual section identification during the editing process.
3Reliability
If voice is deleted in identified sections, then naturalness of moving image is improved, but complexity of editing process increases
Solution Approach 1:
The patent segments the moving image into distinct temporal sections based on detected pause filler positions. The voice track is divided into segments corresponding to pause filler occurrences, allowing targeted deletion only in these specific sections while preserving the rest of the audio content. This segmentation approach simplifies the editing process by limiting operations to identified sections rather than requiring complex global processing.
Data Source
AI summary
An information processing apparatus capable of easily deleting meaningless utterances from a moving image without making the moving image unnatural. A moving image including sound information is acquired, and contents of an utterance by a speaker, which is recorded in the moving image, are analyzed based on the sound information of the moving image. A section including the meaningless utterance, in which a meaningless utterance uttered by the speaker appears, is acquired from a timeline of the sound information based on a result of the analysis. The speaker on a screen of the moving image in the section including the meaningless utterance is detected. Editing for deleting the meaningless utterance from the moving image is executed with respect to the section including the meaningless utterance of the moving image according to a result of the detection.