🎙️ Real-Time Gender Classification of Voice Signals Using DSP Techniques in MATLAB

🧠 Overview

This project demonstrates a real-time voice gender classification system developed entirely with Digital Signal Processing (DSP) techniques in MATLAB.
It uses pitch-based feature extraction to classify the speaker’s gender without relying on any machine learning or pre-trained models.

The goal is to highlight the power of classical DSP — filtering, energy analysis, and acoustic feature extraction — in solving real-world problems like speech-based classification.

---

⚙️ Features

• Real-time recording and analysis of speech input
• FIR band-pass filtering (80–300 Hz) for pitch isolation
• Silence removal using short-term energy thresholds
• Pitch extraction via MATLAB’s built-in pitch() function
• Threshold-based gender decision logic
• Interactive MATLAB GUI for recording and displaying results

---

🧩 Project Workflow

1. Audio Recording — Captures 5 seconds of voice input using MATLAB’s audiorecorder().
2. Filtering — Applies a FIR band-pass filter (80–300 Hz) to remove irrelevant frequencies.
3. Silence Removal — Uses frame-wise energy detection to exclude low-energy (silent) segments.
4. Feature Extraction — Extracts pitch using the Normalized Cross Correlation (NCF) method.
5. Classification — Determines gender based on average pitch:
   • Pitch < 190 Hz → Male
   • Pitch ≥ 190 Hz → Female
6. Visualization — GUI displays waveform, average pitch, and classification result.

---

🖥️ MATLAB GUI Overview

The GUI (built with App Designer) includes:

• Duration control (1–60 seconds)
• “Record & Classify” button with indicator lamp
• Live waveform plot
• Display of extracted pitch and classification output

---

🧪 Results

• The system successfully classified male and female voices in controlled environments.
• Pitch-based thresholds showed strong alignment with expected ranges:
  • Male: typically < 190 Hz
  • Female: typically > 190 Hz
• GUI provided immediate visual and textual feedback.

---

⚡ Challenges

• Background noise affecting pitch estimation
• Preserving soft speech segments during silence removal
• Avoiding phase distortion (solved via zero-phase filtering with filtfilt)

---

🧠 Learning Outcomes

• Implementation of classical DSP techniques in real-time applications
• Design and testing of FIR filters and signal preprocessing routines
• Practical understanding of acoustic feature extraction
• GUI development using MATLAB App Designer
• Improved understanding of speech processing pipelines

---

📂 Repository Contents

Folder	Description
/code	MATLAB source files (.m and .mlapp)
/docs	Full report (PDF) with explanation and system design
README.md	Documentation and usage overview
LICENSE	License information (MIT recommended)

---

🪪 License

This project is open-sourced under the MIT License, allowing use, modification, and distribution with attribution.

---

🌟 Star this repository if you find it helpful for your DSP, MATLAB, or signal processing projects!