What problem does it solve?
This Skill helps production teams clean and enhance speech recordings affected by background noise, room reverberation, acoustic echo, poor microphones, narrowband audio, and competing speakers while preserving intelligibility, speaker identity, and performance quality.
Core Features & Use Cases
- Effect Selection: Choose the appropriate NVIDIA Maxine effect, including Background Noise Removal, combined denoise and dereverb, Acoustic Echo Cancellation, audio super-resolution, Studio Voice, Speaker Focus, and consent-controlled Voice Font.
- Runtime Planning: Design local AFX SDK, self-hosted BNR NIM, streaming, or transactional workflows with correct sample formats, model profiles, latency budgets, GPU requirements, and privacy controls.
- Production QA: Build A/B intensity tests, artifact ledgers, ASR checks, loudness validation, synchronization checks, and subjective listening reviews for podcasts, conferencing, broadcasts, advertisements, avatars, and social media.
- Use Case: Clean a remote podcast guest's noisy and reverberant recording with a combined denoise and dereverb workflow, then verify that speech remains natural before final mixing and delivery.
Quick Start
Use the NVIDIA Maxine audio effects skill to create a privacy-approved cleanup plan for the provided recording, select the correct runtime and effect, define processing parameters, and produce an A/B QA checklist.
