Looking for feedbacks on FluffyAudio Fourier — Spectral Audio Editor for Apple Silicon

[paoling]

Fourier Manual

Hello everyone,

Together with my colleague Olmo, I've been working on something different:

Fourier, a spectral audio editor for Apple Silicon Macs. We're now looking for people who'd be interested in trying it out and giving us their honest feedback.

There were a few reasons this project mattered to me. There are plenty of openly available, state-of-the-art audio models and algorithms out there, yet a lot of companies charge a hefty price for solutions that can, in many cases, be reproduced in a Colab notebook or with a handful of Python scripts.

So the idea was simple: bring those models together inside a complete, feature-rich audio editor that anyone can actually use, while taking full advantage of the compute power of the new Apple Silicon chips (M1 through M5). Because the whole editor is GPU-accelerated, the experience stays fast and fluid, unlike a lot of other software that unfortunately still relies on interfaces and frameworks (often QT) designed many years ago.

For now, Fourier runs only on macOS 14–15 and on Apple Silicon, but down the road we may look into a way to bring it to Windows too.

The editor is essentially complete. Some highlights:

- 45+ modules covering a wide range of audio correction and restoration tasks.
- A text mode that transcribes speech in virtually any language, lets you edit the transcript as easily as a Word document, and then regenerates the audio in the same voice seamlessly.
- A pitch-correction system whose quality has little to envy from far more expensive software.
- A modular architecture we plan to keep expanding with new features and modules over time.

Honestly, I think it's one of the best audio editors on the market right now, though, being a brand-new product with a feature list that already rivals tools that have been around for years, there may still be a few small things to polish here and there. As it stands, it's an extremely fast, feature-packed editor that puts advanced generative and restoration tools, the kind usually reserved for developers or locked behind subscriptions, directly in your hands. And crucially, all processing happens locally, entirely on your machine.

We're looking for people who'd like to try Fourier (in exchange for a complete copy of the software) and share feedback, or even suggestions, so we can keep improving it and deliver a better and better piece of software.

Fourier will go on sale very soon (somewhere between $100 and $150), with regular updates planned. So if anyone would like a copy in exchange for a bit of feedback, I'd love to hear from you.

Requirements
- Apple Silicon Mac (M1–M5)
- macOS Sonoma (14) or later, up to Tahoe
- At least 16 GB of RAM (more RAM makes running some of the heavier models easier)

---

- Platform: Native macOS app, built exclusively for Apple Silicon (M1–M5). Not a cross-platform wrapper — written in Swift with a modern SwiftUI/AppKit interface.
- Processing: 100% on-device. Nothing is uploaded to the cloud; all analysis, restoration, and AI generation run locally, GPU/Neural-Engine accelerated via Metal and Core ML.
- Editing core: Full waveform + spectrogram editing, spectral repair (heal / attenuate / boost), and time-frequency selection tools.
- Restoration suite: De-hum, de-click, de-clip, de-ess, de-reverb, broadband denoise, and more.
- AI suite (all local): Speech transcription (any language), voice re-generation from edited text, note/pitch detection and correction, source separation, and generative audio.
- Mastering & stereo: Reference match, stereo width control, mono↔stereo, and related tools.
- Architecture: Modular — new modules and features are added on a rolling basis.
- System requirements: Apple Silicon Mac, macOS 14+ (Sonoma → Tahoe), 16 GB RAM minimum.

Fourier Manual






Please write here or PM me if you are interested in having a copy and exchange feedbacks!

[soundbylaura]

Hi Paolo,

That looks AMAZING! I'd love to help out with testing and feedback.

[wst3]

Hi Paolo,

Not quite enough to make me jump ship, but it is now on the list of reasons to consider a Mac. The pain of figuring out which tools are cross platform, and the threat of the need to learn new tools prevents that at the moment.

So may I please beg for a Windows version?

Based solely on the quality of your libraries I'd be willing to jump in with both feet. And if that wasn't enough, the manual seems well written. And the feature list is impressive.

I hope you have great success with Fourier! I'll be here waiting for the Windows port

aside - I hope this won't prevent you from creating new libraries!!!

Bill

[Luke]

Ciao Paolo, this looks amazing and you are definitely onto something. Like others, I’ve been perennially searching for the right combo of features/UI/price point. Does this or will it eventually do multi-track audio?

Either way, I would love to help test it out. I’ll be back in the studio Monday.

[paoling]

Hello guys, thank you for your nice words!

I’ll try to show some examples of the nicest features in Fourier soon, so you can get an idea of what the software does. Since it’s developed entirely in Swift, Apple’s programming language, it would be quite hard to port it to Windows, but we’ll definitely give some thought to what we can do.

A few things:

How come a sample library developer has suddenly jumped into coding and AI?

The reality is that FluffyAudio is not my only company. In fact, in 2019 I also co-founded Zoundream, a company that uses AI to analyze and interpret patterns in baby cries. For me, it was a wonderful opportunity to learn more about a technology that I knew would become relevant in our industry. During these years, I’ve also experimented with training generative models on appropriately licensed and carefully selected material for sample generation and audio manipulation. Eventually, I realized that we couldn’t compete with heavily funded companies operating with very different approaches to training data and copyright.

That didn’t deter us from experimenting with new technologies to help us in our work, and in more recent libraries I’ve sped up the scripting process in Kontakt with AI-assisted coding. During these years, I’ve also developed a couple of plugins, such as “FluffyAudio Spots,” which was a kind of granular synth influenced by physics. I may resurrect it one day. It was cool! I’ve also worked on other interesting experiments. Moreover, we’ve developed quite a few internal tools that have helped us dramatically speed up the editing of our libraries.

The sample library market has become somewhat saturated, and it is increasingly difficult for new products to offer something genuinely different. While we still have quite a few products in production, the reality is that a new product is not automatically better than anything that came before it. While that’s a good thing for the lifecycle of products that have always been good, like CSS or a few of our flagship libraries, it has made this a difficult niche in which to sustain a company.

I’m happy that FluffyAudio, a company founded in 2012, is still alive after fourteen years, despite the times. We’re taking advantage of the opportunities offered by some of these new technologies to discover new intersections between audio, music, and computers. Where will this lead us? I don’t know, but as long as we have ideas to explore, we’ll be super happy to continue sharing our work with the community.

I've written/writing PMs to soundbylaura and Luke, thanks for your kind interest!

[Ashermusic]

I would love to be helpful but I am not good with visuals, actually have something of a learning disability, so not for me.

[My name is Nobody]

Hi Paolo,

I would also be interested in trying out what this can do.

Please let us know what the actual conditions are that make feedback useful, as I am sure I am not acquainted with all that is in this software you built.

Thanks for considering

Best

[paoling]

Thank you so much! It’s an audio editor, so at the very least, it should work as one.

It allows for spectral editing, such as removing noise and other unwanted sounds. The idea is that if you encounter any issues while using it, you can simply tell us and we’ll try to fix them. Also, if there’s a feature you particularly like in another piece of software, let us know and we’ll try to add it to Fourier. :-)

Also, Jay, if you’d like a copy, I’ll be happy to send one to you! The editor works much like any other audio editor, and there’s also a waveform view, just in case. You can open a WAV file and perform various operations on it.

The only requirement is that you have a MacBook with Apple silicon, running macOS Sonoma or later, and at least 16 GB of RAM. It also works with 8 GB, but many of the AI features would be difficult to run.

[Thomas Mavian]

Hi Paolo,

I wish you all success with this and like Bill I'll keep an eye open for a Windows port.