Video Personalization for Strength Training
summary
Tonal claims a working system that composes a bespoke workout video on demand by assembling tagged clips with metadata and server-side logic so each user gets a different timeline. The stack includes a cloud AI infrastructure that picks clips, client-side mixing with a modified ExoPlayer, real-time triggers and loops via triggered sections, and a sizeable on-device cache (about 20 GB local cache at 2 Mbps Wi‑Fi). Rep-count triggers, form tips, and reusable clips all run in production.
AI-written summary · about
chapters (11)
- 0:00Intro
- 1:19PROCESS
- 3:08PARTNER WORKOUTS
- 7:00MEDIA SELECTION: AUDIO
- 7:27ADVANCED AUDIO
- 8:50SYNCHRONIZED FEATURES
- 9:34BASIC FEATURES
- 11:16TRIGGERED SECTIONS
- 11:51TRIGGERED AUDIO
- 12:33TRIGGERED LOOPS
- 13:01FULL TIMELINE
related talks
40mPoseDance Build a TikTok Trainer
13mVid2Coach: Transforming How-To Videos into Task Assistants
19mAI video is Already App-ening
46mMastering AI Video Creation: Simplifying Complex Products for Business Success
13mVisMimic: Integrating Motion Chain in Feedback Video Generation for Motor Coaching
29mFrom Quick Wins to Revolutionising Productivity & CX with GenAI
25mCreative video editing with Jetpack Media3 and Gemini
18mBuilding AI-powered video generation tools for web and mobile platforms