Audio Data Augmentation and Pipelines

Apply spectral gating and statistical Wiener filtering to denoise speech signals.

Perform voice activity detection (VAD) to segment speech from non-speech regions in an audio stream.

Implement time-domain audio augmentation techniques, including speed perturbation and pitch shifting.

Implement frequency-domain audio augmentation by applying SpecAugment to mel-spectrograms.

Design and build a robust audio data loading and batching pipeline in PyTorch using torchaudio.

Integrate on-the-fly feature extraction (e.g., mel-spectrogram generation) into a PyTorch data pipeline.

Explain the importance of forced alignment and apply a pre-trained model to generate word-level timestamps.

Implement and interpret standard ASR evaluation metrics, specifically Word Error Rate (WER) and Character Error Rate (CER).