AudioLens: listen to audio and inspect waveforms and spectrograms in VS Code

AudioLens lets you listen to audio, inspect waveforms, and analyze spectrograms directly in VS Code. Open a recording beside the code, transcript, manifest, or model output you are working on, then select a region to hear and inspect it more closely.

I maintain AudioLens, a free, open-source extension under Apache-2.0. The extension was developed primarily with AI-generated code, and this post was prepared by an AI agent; DEV's AI disclosure is enabled.

Listen to the part you want to inspect

Open an audio file with AudioLens and press Space to play or pause. Seek along the timeline, or drag a time selection to play that region.

For multi-channel recordings, each channel has its own track. Use Mute and Solo to isolate what you hear, with per-channel gain and pan controls. This is useful when a recording sounds wrong but the problem belongs to only one channel.

See the waveform and spectrogram together

Each channel can show a waveform, a spectrogram, or both, with a shared timeline. Zoom into a transient, compare channels, or follow a tonal component across time while listening.

The spectrogram uses STFT analysis and exposes controls for FFT size, window function, frequency scale, displayed frequency range, and color palette. Frequency scales include linear, logarithmic, Mel, Bark, and ERB. You can choose the view that helps with the signal you are inspecting.

Drag a region to inspect the active channel's measurements, including RMS, sample peak, clipping ratio, dominant frequency, and spectral centroid. These give you numerical context alongside what you see and hear.

Keep audio beside your development workflow

AudioLens also supports workflows that come up in speech and audio development:

• Remote SSH: preview and inspect audio in a remote workspace without manually copying a dataset into a separate local audio application.
• Raw PCM: specify sample rate, channel count, encoding, byte order, and byte offset explicitly.
• Manifest and log files: hover over a supported audio path and choose Open in AudioLens.
• Kaldi WAV Ark: inspect archives containing RIFF/WAVE entries. This is specific to WAV entries, not every Kaldi archive type.

Try it

Install AudioLens from the VS Code Marketplace, or run:

code --install-extension simzhou.audiolens

Open an audio file. If VS Code chooses another viewer, use Reopen Editor With… → AudioLens. Try selecting a short region, listening to it, and switching between waveform and spectrogram views.

For a reproducible example, this companion Gist uses only Python's standard library to generate three deliberate problems: a silent channel interval, hard clipping, and a raw PCM sample-rate mismatch. It includes reference measurements checked against the saved samples, and the generated files work with other audio tools too. The Gist's editor walkthrough has not been independently executed as a UI test. The GIF above is a general product demo, separate from that fixture.

Source code, documentation, and more GIF demos

What would make audio inspection easier in your VS Code workflow? I'd welcome feedback about playback, spectrogram controls, multi-channel inspection, or remote files. Reproducible problems can also go in the project's GitHub issues; please avoid attaching private recordings.