This project focuses on classifying one-second audio clips into four genres: classical, pop, rock, and blues using machine learning techniques. By processing MFCCs and mel-spectrograms, we combine traditional audio processing techniques with modern deep learning methods to achieve accurate predictions.
- Audio Representations:
- MFCCs: Extracted features (mean and standard deviation) for each clip.
- Mel-Spectrograms: Time-frequency representations of audio.
- Deep Learning Models:
- Feedforward Neural Networks (FNNs) for MFCC-based classification.
- Convolutional Neural Networks (CNNs) for mel-spectrogram-based classification.
Overall, the combination of traditional audio processing techniques with advanced neural network architectures achieves reliable and scalable results in the domain of music genre classification.