● Covered by 1 source · 1 reportLow impact1 neutral
Overtone Tool Transcribes and Cuts Captions for Multiple Screen Shapes from One File
Key points
- Overtone stores words, times, and speakers once.
- Generates separate caption cuts for different aspect ratios.
- CLI tool for transcription or subtitle recutting.
- Video.js plugin integrates captions into players.
- Supports offline processing with whisper.cpp and Ollama.
Unified Caption Management
Overtone introduces a method for managing captions where each spoken word, its timestamp, and speaker information are stored once in a single .overtone.json file. This central repository allows the tool to generate distinct caption cuts tailored for various screen aspect ratios, such as 16:9 and 9:16, from this unified source. This approach contrasts with traditional methods that create a single caption file, which often results in poor readability when adapted to different screen sizes.
Addressing Caption Readability Issues
Traditional captioning often involves filling lines to a length limit, leading to awkward line breaks and shared captions between speakers when displayed on different screen shapes. Overtone's system ensures that captions adhere to reading rules, preventing lines from ending on weak words, keeping names and numbers together, and assigning individual captions to each speaker. It also uses pauses to create appropriate line or caption breaks, improving comprehension for viewers.
Command-Line Interface and Integrations
The tool includes a Command-Line Interface (CLI) that enables users to create the .overtone.json file from existing subtitles (SRT) or directly from audio and video files. For transcription, it supports provider API keys like OpenAI and Azure AI Speech, and also offers an offline mode using whisper.cpp and Ollama. Additionally, Overtone provides a video.js plugin, allowing for the display of the appropriate caption cut within a video player based on its aspect ratio.
Technical Implementation
Unlike formats such as WebVTT or TTML, which store finished caption cues, .overtone.json stores individual words and their metadata, along with pointers to these words for each specific cut. This design allows for flexible generation of standard formats like WebVTT or SRT per cut when needed. The recutting functionality works offline without API keys or ffmpeg, requiring Node.js 22.12 or newer, while transcription from audio/video may require ffmpeg and API keys depending on the chosen method.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
Primary sources
GitHub cmc_internal/api GitHub github/collect GitHub lonthq/overtone-captions GitHub _private/browser GitHub get-started/accessibility GitHub en/copilotReporting from
Overtone is a new tool that generates captions for various screen aspect ratios from a single source file, storing each spoken word with its time and speaker. It addresses issues with traditional captioning by creating separate cuts for different screen shapes, ensuring readability and adherence to captioning rules. The tool offers CLI functionality for transcribing audio/video or recutting existing subtitles, and a video.js plugin for display.