Tiny App: Baby Heartbeat Listener

What it does

Tiny is an iOS app designed to create an intimate bonding experience for expecting parents. Using your iPhone"""s microphone and AirPods, parents can listen to sounds from the womb while watching a beautiful, reactive "Orb" visualization that responds to audio input. The app includes an interactive pregnancy timeline and the ability to record and privately share these special moments with family through secure, code-protected rooms.

Note: This app is for emotional bonding and entertainment purposes only, not medical monitoring. iPhones cannot reliably detect fetal heartbeats—always consult your healthcare provider for medical concerns.

How I built it

Built the app in SwiftUI with a focus on creating a tactile, high-end experience that makes parents feel connected to their baby.

• The "Orb" Visualization: Created a custom Canvas-based component that transforms audio input into fluid, organic animations. The OrbLiveListenViewModel maps real-time sound amplitude to dynamic shape changes, giving the interface a "living" quality that reacts to what parents hear.

• Audio Processing: Used AVFoundation and AudioKit to process microphone input through a Parametric EQ and High Shelf Filter, emphasizing lower frequencies around 200Hz to highlight rhythmic sounds while preserving the natural ambient audio of the experience.

• Signal Analysis: Built a custom audio analyzer using the Accelerate framework with FFT processing. A dynamic peak buffer detects rhythmic patterns in the audio without aggressive filtering, keeping the soundscape natural and immersive rather than artificially silent.

• AirPods Integration: Implemented seamless audio routing for AirPods using AVAudioSession, ensuring low-latency monitoring while simultaneously recording high-quality audio to Firebase Storage for safekeeping.

Challenges

The biggest challenge was creating an immersive audio experience that felt authentic and emotionally resonant.

• Solution: Instead of using harsh noise gates that create unnatural silence, I implemented a dynamic peak buffer that adapts to the ambient noise floor. This preserves the natural "whooshing" and environmental sounds that make the experience feel real, while still identifying rhythmic patterns to drive the Orb"""s visualization. The result is an interface that feels alive and responsive without feeling clinical or artificial.