VOICE Summit 2019 | VED Talks | Melissa Hammersley | July 23
summary
A company builds an app that listens continuously to a reader and layers music, sound effects and character voices timed to printed books, moving from failing keyword spotting to on-device speech recognition for low-latency responses. A web-based drag-and-drop soundscape editor lets creators align cues to text, and the system deploys to mobile, web and OTT. Results show increased child engagement and natural conversation; designers are urged to design beyond call-and-response voice assistants.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 0:29How Melissa got into voice
- 1:59How the app works
- 3:22The journey so far
- 4:44New products
- 6:09Building products
- 7:29Mixed reality storytelling
- 10:02Software
- 10:46Outro
related talks
36mVoice Enabling Mobile Apps and Custom Hardware Devices
38mCreate the next generation of user experiences with text generation
22mBuilding next-gen speech synthesis for Bookmate audiobook service
21mVeo 3 for Developers
03h 52mLivestream 1
24mFail fast, fail cheap, making voice apps the guerrilla way
35mHaving a Voice on Android: How hard can it be?
06mSensing to Hear: Speech Enhancement for Mobile Devices Using Acoustic Signals