Web Application ยท 2023
SPEECH TRANSCRIPTION
A versatile speech-to-text web application built with Flask. Supports real-time microphone input, audio file uploads, and video file transcription using Google Web Speech API.
The Problem & User Needs
Accessibility and documentation generation often require transcribing audio from various sources. Users needed a unified tool that could handle real-time dictation as well as pre-recorded audio and video files without requiring complex software installations.
Role & Contributions
Backend & Frontend Developer
Integrating Google Speech API, building the Flask backend, handling audio/video file conversions, and creating the web interface.
Research Methods
Investigated API limitations and constraints regarding audio formats and file sizes. Researched Python libraries for robust media processing (MoviePy, PyDub) to ensure seamless handling of diverse user uploads.
Design Process
Designed a processing pipeline that normalizes various media inputs into a standard WAV format for the Speech API.
Created a minimalist, straightforward HTML/CSS interface allowing users to easily select between the three input modes.
Measurable Outcomes
Delivered a functional tool capable of handling MP3, WAV, and video files. The automatic format detection and conversion pipeline reduced transcription failure rates during testing. Successfully demonstrated real-time API integration.