nvidia-noise-removal

Eliminate Background Noise with NVIDIA Maxine AI

Transform noisy field recordings into professional studio-quality audio. This skill leverages the NVIDIA Maxine Background Noise Removal (BNR) NIM via gRPC to isolate human speech from distracting environmental sounds. It effectively eliminates high-frequency clicks, low-frequency hums, and chaotic background chatter that standard filters can't touch.

What it does

• Removes mechanical noise like keyboard typing, mouse clicks, and PC fans.
• Filters out environmental distractions including traffic, wind, barking dogs, and HVAC hum.
• Preserves the natural tone and clarity of the speaker's voice.
• Provides adjustable intensity ratios (0.0 to 1.0) to balance noise suppression with vocal preservation.

Why use this skill

Standard noise gates often cut off the beginning or end of words, while basic EQ filters leave the audio sounding "thin." This AI-driven approach uses deep learning to understand the difference between speech and noise. By automating the gRPC handshake and protobuf communication with NVIDIA's infrastructure, this skill saves developers from the complexity of manual gRPC implementation and environment setup.

Supported Environment

• Protocol: High-performance gRPC (TLS secured).
• Requirements: Python 3.10+, NVIDIA API Key (free), and grpcio.
• Input: WAV files (up to 35MB).