Audio scenes | Treble Technologies

Audio scenes

Combine room impulse responses with audio content to build labeled datasets for audio ML training tasks such as speech recognition, sound event detection, source separation, and more. Covers constructing individual audio scenes, generating scene collections at scale from reusable rules, rendering scenes to noisy and clean audio pairs, and preparing audio datasets as source content.

📄️Audio scenes

Construct one audio scene manually from IRs, source tracks, and listener configuration.

📄️Scene collection

Generate a Scene Collection with many randomized audio scenes from reusable rules.

📄️Rendering a scene collection

Render scenes, extract target signals, and export metadata or transcripts.

📄️Track generator

Generate audio tracks from an audio dataset using rule-driven track types.

📄️Audio datasets

Prepare and load audio datasets for use with scene generation.