Label Studio 1.20.0: Spectrograms, Time Series Sync, Playground 2.0, and JSONL Support

Label Studio 1.20.0 is here with powerful upgrades for audio, video, and time series workflows, plus smarter data imports, better review experiences, and an upgraded Playground that now works as a standalone app. Whether you're building training data from scratch or refining evaluation sets, this release brings more precision, flexibility, and insight to your annotation stack.

Sync Time Series with Audio and Video

You can now align time series data with audio and video using the new sync parameter. Each frame is mapped to its corresponding timestamp, so you can track events more accurately across modalities.

Learn more about Time Series Labeling with Audio and Video Synchronization here.

Spectrogram Support for Audio Analysis

Visualize frequency and amplitude over time using spectrograms for audio tasks. Customize settings like dBs, mel bands, and color schemes to surface subtle patterns that waveforms might miss. Ideal for voice, music, and any fine-grained audio signal analysis

Learn more here.

New Multichannel Time Series View

Group multiple time series streams into a single view with the new Multichannel tag. You can use this tag to combine and view multiple time series channels simultaneously on a single channel, with synchronized interactions.The Multichannel tag significantly improves the usability and correlation of time series data, making it easier for users to analyze and pinpoint relationships across different signals.

Learn more here: How to Label Multichannel Time Series Data for AI Models.

Playground 2.0: Now Standalone, Audio-Ready

The Label Studio Playground has been rebuilt and now supports more data types (including audio) and is available as a standalone app. Paste your config, preview tasks instantly, and tweak setups without leaving the browser.

Enhancements

• Interactive View All: Play audio, scrub video, and highlight regions side-by-side when reviewing annotations.
• JSONL Support: Import JSONL files from cloud storage, useful for OpenAI fine-tuning, HuggingFace, or Sagemaker pipelines.
• Grid View Upgrades:You can now adjust image size and choose how many images appear per row in the Data Manager. This gives you more flexibility when scanning and reviewing visual tasks at scale.
• Selective Annotation Deletion:Previously, you could only delete all annotations for all selected tasks. Now, you can bulk-delete annotations by a specific user across selected tasks for more precise cleanup.
• UX Polish: Improved empty states, import steps, and panel guidance to reduce confusion and onboarding friction.

Ready to explore? Check out the full Label Studio 1.20.0 changelog.