WEEK 4: RESEARCH FINDINGS

This week we were tasked to provide an update as our midterm assignment. This post will cover the main points of my presentation.

MAIN POINT OF CHANGE:

• Pre-recorded voice & script -> AI Generative Conversations & Voice Recognition

• Purchased the rotary phone

DO I USE A RASPBERRY PI OR ARDUINO?

• Raspberry Pi!

  • Accurate audio processing and API integration

  • Can handle speech recognition & text to speech

  • USB audio support

  • I am more proficient in Python

  • Can multitask

UPDATED TIMELINE

• WINTER

  • WEEK 5: Purchase equipment/hardware, begin coding

    WEEK 6: Continue coding

    • Focus on text-to-speech & API integration

  • WEEK 7: Coding complete, begin hardware assembly

  • WEEK 8: Assembled, begin testing

  • WEEK 9: Final presentation

• SPRING

  • WEEK 11: PROGRAM API UNDER CERTAIN PERSONALITIES, FIND/CREATE VOICE BANKS

  • WEEK 12-13: BEGIN SCULPTURE 

    • BUY MATERIALS, CREATE SIMPLE MOLD

  • WEEK 14: ASSEMBLE EVERYTHING

  • WEEK 15: MIDTERM

  • WEEK 16-18: USER TESTING/DEBUGGING

HARDWARE

Raspberry Pi 4 Model B (4GB RAM)	Main processing unit
Vintage Rotary Phone Shell	Thrifted or vintage phone body
USB Audio Adapter	Enables microphone and speaker connectivity (e.g., UGREEN USB Audio Adapter)
Mini Speaker (3W-5W, 4Ω or 8Ω)	Fits inside the phone's original speaker location
Electret Microphone Module (MAX9814 or similar)	Captures user speech, placed in the original microphone location
SPDT Micro Switch	Detects handset lift and placement
5V 3A Power Supply with USB-C Connector	Powers the Raspberry Pi (e.g., LitStar 5V 3A USB Type C Adapter)
3.5mm Audio Jack Breakout Board	Facilitates clean audio connections inside the phone shell

SOFTWARE

• Raspberry Pi OS (Lite or Full)

  • 16GB+ micro SD card

• Python (3.x)

• Speech-to-Text API (e.g., Google Cloud Speech-to-Text API)

• Text-to-Speech API (e.g., ElevenLabs API / Google Text-to-Speech)

CORE INTERACTIONS

• Handset Detection

• User Speech Input

• AI-Generated Response

• Audio Playback

• Session Termination

ASSEMBLY STEPS

Step 1: Disassemble the Rotary Phone

Step 2: Install the Raspberry Pi

Step 3: Set Up Handset Detection

Step 4: Position the Microphone and Speaker

Step 5: Connect Audio Components

Step 6: Develop and Deploy Software

CODE LINEUP

• Configure API key

• Audio settings for output

• Detect when phone is picked up

• Functions

  • AI prompts user with dial options

  • Dial

  • Choose AI depending on dial choice

  • Prompt introduction

  • Record audio

  • Transcribe audio

  • Generate response

  • Text-to-Speech

  • Call interaction

• Loop until interaction ends