Subtitling Block Segmentation for Live Captioning Modes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current live subtitling technologies lack user flexibility in presentation modes, often presenting subtitiles word-by-word, which can be unpleasant and of poor quality, and do not allow users to choose between modes such as block-wise, line-by-line, or sentence-by-sentence display.
Innovation Solution
The method allows users to control the presentation mode of subtitiles by inserting end-of-block conditions into the subtitiles stream or file, enabling multiple presentation modes such as sentence-by-sentence or block-wise display, based on predefined characters or punctuation, and supports various standards like MPEG-TS and W3C TTML.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If sub-titles are transmitted word by word as soon as available, then latency is minimized, but readability and coherence deteriorate
Solution Approach 1:
The patent segments sub-title content into meaningful blocks (sentences, phrases, or thematic groups) rather than transmitting individual words. This segmentation allows the system to maintain low latency by transmitting blocks as soon as they are formed, while improving readability by presenting coherent linguistic units to the user.
Solution Approach 2:
The system performs preliminary grouping of words into coherent blocks before transmission. By organizing sub-title content into meaningful units in advance (even in real-time), the system ensures that users receive readable, coherent text without significant delay, resolving the contradiction between speed and quality.
2Productivity
If sub-titles are presented line by line with automatic shift up, then continuous display is maintained, but user tracking difficulty increases
Solution Approach 1:
The patent introduces a block-based segmentation approach that groups multiple lines into a single stable display block. This prevents the traditional line-by-line shift-up effect by maintaining the entire block in view, allowing users to track content without constant repositioning while still achieving continuous display through block progression.
Solution Approach 2:
The system dynamically adjusts display behavior based on block completion rather than line completion. This dynamic approach allows the display to remain stable during block presentation, improving user tracking, while still progressing continuously through the sub-title content by transitioning between blocks.
3Adaptability or versatility
If multiple presentation modes are provided, then user preference satisfaction improves, but system complexity increases
Solution Approach 1:
The patent implements a universal block-based presentation framework that can accommodate multiple display modes (word-by-word, line-by-line, sentence-by-sentence, or custom block configurations) through a single coherent architecture. This multi-functional approach allows the system to provide diverse presentation options without proportionally increasing complexity, as the underlying block segmentation mechanism serves all modes.
Solution Approach 2:
The system achieves multiple presentation modes by changing parameters of the block-based display system rather than implementing separate processing pipelines. By adjusting block size, segmentation criteria, and display timing parameters, the system can switch between different presentation modes while maintaining the same core architecture, thereby improving versatility without excessive complexity increase.
Data Source
AI summary
Creating a sub-titles stream or file composed of sub-titles elements. For each sub-titles element in said sub-titles elements, a sub-titles element is inserted into the sub-titles stream or file, at least one end-of-block condition related to a mode of presentation of sub-titles is verified, and a datum representative of an end of a block according to the mode of presentation of sub-titles is inserted into the sub-titles stream or file upon satisfying said at least one end-of-block condition. A sub-titles stream or file may be presented on a terminal of a user by selecting on the terminal a mode of presentation of sub-titles, reading a sub-titles block associated with the mode of presentation in the stream or file, and presenting on the terminal the at least one sub-titles block according to the mode of presentation.