Blog › Comparisons

Somuno vs Audacity Noise Reduction: free classic or AI cleanup?

Hongyang Jin · November 7, 2026 · 7 min read

Audacity is a free, open-source audio editor, and its Noise Reduction effect is the classic way to clean up a recording. You select a moment of pure noise, Audacity learns its “noise profile”, and then it reduces that noise across the track. It's built for steady sounds like hiss, hum and air conditioning. Somuno is a Mac app that uses an AI speech-enhancement model instead: there's no profile to capture, and it also removes room reverb and repairs damage such as clipping and missing high frequencies.

If your problem is steady hiss or hum, start with Audacity. It's free, it runs on Mac, Windows and Linux, and you may need nothing else. Choose Somuno if your noise isn't the steady kind Noise Reduction is built for, if the voice sounds echoey or damaged, or if you need to clean video files or whole folders at once – something Audacity 4.0 can't batch yet – and you have an Apple silicon Mac. We haven't run listening tests comparing the two, so try both on your own recording: both can be tried for free.

Somuno and Audacity at a glance

How Audacity's Noise Reduction works

Noise Reduction works in two passes, as described in Audacity's manual:

1. Select a stretch of audio that contains only the noise – a pause, or the room tone before someone speaks – and choose Get Noise Profile.
2. Select the audio you want to clean and apply the effect, with three controls: Noise reduction (0–48 dB, default 6), Sensitivity (default 6) and Frequency smoothing (0–12 bands, default 6).

You can preview what's being removed with the “Residue” option. It's transparent and controllable: you can see exactly what it does and turn it up or down. The manual is also honest about its limits. It says the effect “suits constant background noise; it will not remove one-off sounds like a door slam”. The older 3.x manual adds that it does poorly on “irregular background noise such as from traffic or an audience”, or when the noise is very variable or very loud, and that heavy settings cause “musical noise” – a warbly, watery sound.

So the classic approach depends on two things: the noise has to be steady, and you need a clean stretch of it. That's one reason to record a few seconds of room tone before you start talking, whatever tool you'll use.

How Somuno is different

Somuno doesn't need a noise profile: you just drop in files, and its AI speech-enhancement model handles several kinds of damage in one pass:

The trade-off is control. Audacity gives you three parameters and a preview of the removed noise. Somuno gives you two model choices – General Model, or Keep Laughter, which removes noise but keeps laughter – and one Strength slider from 0 to 100 % that mixes the original back in. There's no spectral editing and no per-band control. Somuno is also built for speech only: it removes or distorts music and other non-speech sounds in the background.

Echo and room reverb

Audacity's documentation describes Noise Reduction as an effect for steady background noise, not for room reverb. Removing echo is one of the main things Somuno is for. Here is Somuno's output on a talk recorded across a large, echoey room, with default settings. These spectrograms show Somuno's result only. We haven't processed this clip with Audacity.

Original After Somuno

More on this in how to remove echo from a voice recording.

Audacity 4.0: batches and AI plugins

Audacity 4.0 came out on September 3, 2026, and the current version is 4.0.2 (November 2, 2026). Two things haven't arrived in 4.0 yet:

• No batch processing for now. In Audacity 3.x, Macros can apply a chain of effects to many files. The 4.0 release notes say the Macro Manager and the scripting pipe aren't included and are planned for a future release.
• No AI noise plugins on 4.0 yet. Intel's free OpenVINO plugins add AI effects to Audacity, including Noise Suppression, and run locally. On Mac they're a beta build, and Audacity's download page warns they're “not yet compatible with the latest Audacity version”. An Intel contributor wrote in October 2026 that the Mac plugins work only with Audacity 3.7.x.

So, as of November 2026, if you want batches or the AI plugins in Audacity on a Mac, you'd stay on Audacity 3.7.x for now. Somuno's queue is in the free version: add files or whole folders, and each result is saved next to its original with “-enhanced” in the name. See how to clean up a whole folder of recordings.

Video files

Audacity works on audio. Its 3.x manual says that on Mac it can import the audio from unprotected MP4 and MOV files, and from most other video files with the optional FFmpeg library, which is a separate install. Its documentation describes no video export, so you'd put the cleaned audio back in a video editor. Somuno replaces the soundtrack of MP4, MOV and M4V files directly and copies the picture without re-encoding.

Which one should you choose?

How to compare them on your own recording

1. Pick a recording and export a one-minute WAV clip with the problem in it. Somuno's free version enhances the first 60 seconds of each file.
2. In Audacity, profile a second or two of pure noise, apply Noise Reduction at the defaults, then export.
3. In Somuno, process the same clip at the default settings.
4. On headphones, listen to the pauses (is the noise gone, or replaced by warbly “musical noise”?) and the consonants (do S, T and F stay crisp?).
5. If a result sounds over-processed, back off: lower the reduction in Audacity or Strength in Somuno.

Don't stack the two on the same file: two cleanup passes often make a voice sound thin. For more methods, read how to remove background noise from audio on Mac, or hear the before-and-after samples.

Questions