Speech Evaluation System

Overview

This repository contains a Node.js application for evaluating spoken audio files based on various linguistic and sentiment analysis metrics. The system transcribes audio to text using Google Cloud Speech-to-Text API, evaluates the transcribed text against predefined criteria, and provides a comprehensive evaluation score.

Features

1. Audio-to-Text Conversion: Utilizes Google Cloud's Speech-to-Text API to convert audio files (in .wav format) to textual transcripts.

2. Evaluation Metrics:

   • Accuracy: Measures how closely the transcribed text matches the predefined text.
   • Hesitation Rate: Calculates the frequency of pauses and filler words, indicative of speaker confidence.
   • Pronunciation Accuracy: Assesses pronunciation fidelity using phonetic encoding (Soundex algorithm).
   • Expression Score: Evaluates sentiment and emotional tone using the AFINN-based sentiment analysis library.
   • Weighted Average Score: Combines all metrics to provide an overall evaluation score.
   • Grade Assignment: Assigns a grade (e.g., Excellent, Very Good) based on the weighted average score.

3. Technologies Used:

   • Node.js: Backend server environment.
   • Express.js: Web framework for routing and handling HTTP requests.
   • Multer: Middleware for handling file uploads.
   • Google Cloud Speech-to-Text API: Converts audio data to text.
   • Sentiment: Node.js library for sentiment analysis.
   • MongoDB (as future scope): For storing and querying evaluation results.

4. Future Scope:

   • Machine Learning Implementation: Enhance accuracy through training models on diverse speech patterns and improving sentiment analysis.
   • Database Integration: Store evaluation results for historical analysis and reporting.
   • Real-time Evaluation: Implement streaming audio processing for immediate feedback during speech.
   • User Interface: Develop a frontend interface for easy file uploads, real-time evaluation, and result visualization.

VS code backend:

Postman API testing:

Setup Instructions

1. Clone the repository:

   git clone <repository-url>

2. Install dependencies:

   cd <repository-folder>
   npm install

3. Set up Google Cloud credentials:

   • Obtain a service account key for Speech-to-Text API and set the GOOGLE_APPLICATION_CREDENTIALS environment variable to its path.

4. Start the server:

   npm start

   The server will run on http://localhost:3000.

Usage

• Uploading Audio: Use /evaluate endpoint (POST request with audio file) to upload an audio file for evaluation.
• Local File Evaluation: Uncomment and modify evaluateLocalAudioFile function to evaluate a local .wav file.

Example

const localAudioFilePath = 'path/to/local/audio/sample.wav';
evaluateLocalAudioFile(localAudioFilePath);

Contributors

• Add your name here if contributing to this project.

Contact

For questions or issues, please contact vedxnt2912@gmail.com.

---

Feel free to expand on each section with more detailed explanations or add additional functionalities as needed. This readme aims to provide a comprehensive overview and setup guide for anyone interested in using or contributing to the project.