Show HN: Entropic: information-driven variable-rate media playback

Category: library

Tags: media-playback, variable-speed, llm, audio-processing, speech-enhancement

Score: 6.8/10 (Innovation: 7, Technical: 7, Documentation: 7, Utility: 6)

Entropic is a Python tool that uses a local LLM to analyze speech and dynamically adjust audio/video playback speed based on information content, slowing down for unfamiliar words and accelerating through predictable segments. It offers two modes (VBR and skiplow) and supports both audio and video input, making it a novel approach to time-efficient media consumption. The project is interesting for its innovative combination of NLP surprisal metrics with real-time audio processing to address a common listener pain point.

Target audience: backend devs, data engineers

Repository: https://github.com/patrickxia/entropic · Python · GPL-2.0

View on Hacker News