@audio/denoise

@audio/denoise is a free browser-based tool for cleaning recordings and listening to or saving the result. It uses classical signal processing rather than AI noise removal, and the page says processing happens on the device, without uploading audio. In automatic mode, a classifier measures a recording, selects a specialized method, and shows its plan. Users can also select methods manually, including dehum, declick, declip, and gate. The tool describes thirteen open-source restoration methods plus the classifier. It lists steady background noise, mains hum, clicks, pops, vinyl crackle, clipping, wind rumble, plosives, sibilance, room reverb, and noise between phrases as targets. For music, it can treat hum, clicks, crackle, and clipping; broadband denoising should be used lightly and compared with the original. Machine-learning models may do better for heavy speech-over-noise. Automatic mode does not choose dereverb, so that method must be selected directly. The associated package is a single-pass noise reduction tool; developers can install it with npm and process mono Float32Array audio.

@audio/denoise offers several classical restoration methods and keeps processing in the browser. Consider its stated limits for heavy speech noise and automatic dereverb selection when deciding whether it fits a recording.

This tool suits people who want to clean recordings in a browser without uploading audio. Developers can also use its npm package to process mono Float32Array audio.

Overview

@audio/denoise is a free browser tool for cleaning recordings and saving the result. It suits people who want local, targeted audio repair, especially for hum, clicks, crackle or clipping. Its range of methods is useful for focused fixes, but heavy speech noise calls for a different approach.

Key features

Thirteen open-source restoration methods are paired with a classifier that measures a recording, recommends a specialized method and displays its plan. That gives newcomers a starting point while leaving experienced users free to choose methods such as dehum, declick, declip and gate manually. Automatic selection does not include dereverb, so users needing room-reverb treatment must select it themselves.

The target list covers steady background noise, mains hum, clicks, pops, vinyl crackle, clipping, wind rumble, plosives, sibilance, room reverb and noise between phrases. This breadth makes it relevant to spoken recordings and music repair, though it is not a universal speech-cleanup solution: the processing uses classical signal processing, and machine-learning models can perform better on speech buried in heavy noise.

For music, hum, clicks, crackle and clipping are sensible repair targets. Broadband denoising deserves more restraint; comparing the result with the original helps guard against reducing wanted material along with noise.

Processing runs in the browser, and the page states that audio is neither uploaded nor sent off the device. That is a strong fit for users who want local processing. Developers can also install the package with npm; its documented input and output are mono Float32Array audio, making it a more specialized option than a general-purpose editing workflow.

Pricing

The Free plan costs nothing, runs in the browser, requires no account and does not upload audio. It includes noise reduction, click and crackle removal, hum removal, declip repair and spectral repair. There is no batch processing, so users cleaning many recordings will need to work without that convenience. The plan supports both workflow formats and MIDI export, but not stem separation, chord detection or tablature output.

Platforms

@audio/denoise runs on the web. Browser processing keeps the audio on the device, while the documented npm package offers a route for developers working with mono Float32Array audio.

Who it's for

Choose it for individual recordings that need targeted classical restoration, especially music with hum, clicks, crackle or clipping, or when keeping audio on-device matters. Look elsewhere for batch cleanup or speech overwhelmed by heavy background noise. Users who need automatic dereverb selection should also expect to choose that method manually.

Pros and cons

• Thirteen specialized methods plus automatic recommendations: users can start with a classifier or choose a repair method themselves.
• On-device browser processing: the stated no-upload approach suits privacy-conscious cleanup.
• Useful music-repair targets: hum, clicks, crackle and clipping are explicitly covered.
• No batch processing: handling a collection of recordings may take more individual work.
• Limits on difficult speech: machine-learning models can do better with heavy speech-over-noise, and automatic mode does not select dereverb.

Alternatives

PD Cleaner is a free Windows 64-bit option with no account or trial, a fit for users who want a desktop alternative. Vinyl Restoration Suite is another free option; consider it if you want software across Linux, macOS or Windows. VinylRest offers Linux, macOS and web options, with VinylRest Live free in public beta but requiring a license for processing and allowing up to three active Mac installations. Wave Corrector is free for personal use on Linux and Windows, with no registration required for its Professional Edition. Cathar is a free alternative for Linux and macOS. CEDAR Cambridge is a paid modular hardware and software system for macOS and Windows; configuration varies, so contact its maker for information. Neiro is a free option for Linux, macOS, Windows, web or self-hosted use. Algorithmix Sound Laundry is a paid Windows alternative. For a broader comparison, see Audio Restoration Software or Music Transcription Software.

Verdict

@audio/denoise is a strong choice for free, on-device repairs when a recording has a specific, supported problem such as hum, clicks or clipping. Its specialized methods and manual control are the main reasons to choose it; the lack of batch processing and weaker fit for heavy speech noise are reasons to look elsewhere.