Talking to Your Dog with Ember
summary
Prototype deciphers barks into alert / greeting / playful / distress by analyzing audio with the Web Audio API and Ember, using an OfflineAudioContext and AnalyzerNode to extract time/frequency data, filtering to 250–4000 Hz, detecting spikes as barks and mapping patterns (one high = alert, one low = distress, two = greeting, >2 mid = playful). Demo at wolf.plus accepts uploads and mic input, code is open source, works for broad buckets but needs more labels and finer pitch bins.
AI-written summary · about
chapters (13)
- 0:00Start
- 1:49Motivation behind this Talk
- 2:32Demo
- 6:56The Analyzer Node
- 7:34Get by Frequency Data and Git by Time Domain Data
- 10:30Offline Audio Context
- 11:16Ember File Upload
- 12:24Audio Uploader
- 13:16Analyze Audio Function
- 13:58Figuring Out Dog Bark Science
- 18:43Progressive Web Apps
- 20:08Manifest Js
- 22:20More Bark Types
related talks
36mFrom El Jardín to Ember: The Aittendee Experience
20mExploring the hidden potential of sound data
44mAdventures in Puppy Training with A.I. and a Raspberry Pi
01h 14mExploring the hidden potential of sound data
27mDeciphering Brainwaves with the Web Audio API
19mSounds Aware: A Mobile App for Raising Awareness of Environmental Sound
43mDeciphering animal emotions using machine learning and vocalisations
25mLooking at Sound