Spectrogram generator

Access the spectrogram generator in the Dataset tab Import > Create with spectrogram generator.

Afterwards the following dialog will appear allowing you creating images from audio files or csv data.

Converts audio or CSV files into spectrogram images for audio/time-series classification tasks.

Option	Description
Target split	Train, test, or validation directory
Sampling rate	Must match the source data's sampling rate

Synthetic Training Data

When real training data is scarce (edge cases, rare objects, privacy constraints), synthetic data generation via Rendered.ai provides a scalable alternative.