Voice Interfaces in Flutter
summary
Shows a working Flutter assistant using platform APIs (Android SpeechRecognizer, iOS SFSpeechRecognizer) and common plugins, with a live demo that recognizes speech, calls a simple handler, and speaks back via platform TTS — TTS works, on-device STT is noticeably worse. Covers signal details: 16 kHz sampling, 10–25 ms frames, FFT then 13 MFCC coefficients fed to an acoustic model and a language model; compares on-device vs cloud models, notes device-specific language models, early cutoffs, unreliable partial results, and mitigation options.
AI-written summary · about
chapters (12)
- 0:00Voice interfaces matter
- 1:49Voice interface components
- 3:37Microphones capture sound
- 6:28Speech recognition pipeline
- 16:06Text to speech pipeline
- 19:01Flutter for voice interfaces
- 20:47Speech service differences
- 21:51Live coding voice assistant
- 30:45Speech recognition parameters
- 33:17On-device vs cloud models
- 34:41Text to speech demos
- 36:53Summary
related talks
32mHow to build an Auracast assistant app with Flutter
47mAgentive Experiences With Flutter: What, How, and Why Bother?
32mHands on Flutter
49mFlutter In Practice – Tips and tricks for writing Flutter apps, from Start to Finish
33mBuilding Cross-Platform, Multi-Modal Agents with Flutter
42mMore Flutter and Dart and AI, Oh My!
39mBuild voice bots for mobile with Dialogflow and Flutter
43mIntroduction to Flutter