Show HN: Am-1 Arpeggio Machine – A Love Letter to Berlin School Music

It allows you to quickly get up and running with a three-part harmony, based on a chosen key/scale/degree. But once you have your arp running, that's when the fun really begins, because you can fulfill your own Klaus Schulze fantasies right in the browser (and eventually on hardware).

First, you can tweak arp voices, patterns, octaves, attack-release shape, as well as tempo divisions and per-part loop lengths (unequal lengths phase against each other, a la Steve Reich).

Second, record your own 16-step Sequence and apply it to any part, forwards or reversed, and record a chord progression that plays itself, moving the whole ensemble one degree per bar while the melody transposes to follow.

Third, add a Drone layer from one of five voice types, including Solina-like strings and a Mellotron-like choir.

Fourth, pass your arps through Effects, such as a tempo-synced delay, including dotted-8ths, four kinds of reverb, and a swirly phaser.

Fifth, auto-tweak everything in the Mod section: Cutoff, Resonance, Shape, Drive, Reverb, Pitch, Feedback, Delay, Volume, Phaser. There are five LFO shapes, from a simple Triangle to a Lorenz-attractor Chaos modulator, with cycle times down to 30 minutes.

Sixth, now that you have everything going, you have ten real-time performance controls in Master: Tempo, Pitch, Swing, Drive, Entropy (from precise to seasick), Cutoff, Resonance, Feedback (tuned to the key root, it will howl when pushed), Glide, Output.

Two unique features: the Anomaly button generates a random patch that's always in key, and the year selector swaps the whole circuit: 2026 is clinical, while 1983 gets pulse-width drift, drive with bias memory, and a CRT-effect over the screen.

One more thing to share, this synth operates by default in keys and scales, but it has also been designed to allow you to break free from that and go fully experimental and create noise. Almost all parameters go well beyond polite (check out the Gnarl, Scream preset for an example, but cover your ears).

Happy to answer any questions about the Web Audio engine, as of now the entire codebase lives in a single HTML file. I'm currently working on porting to the Daisy Seed platform, which has been both challenging and fun.

I've been messing around with DAWs and soft synths since college, as a hobbyist, and my immediate reaction is that I am struggling a little to see what AM-1 is doing that hasn't already existed in more mature tools for quite a while.

Have you tried something like Stochas, for example? There are a bunch of free and paid sequencers/arpeggiators that already go pretty deep into evolving patterns, probability, polyrhythms, variations, live manipulation, etc.

Even just chaining a few free plugins together in a DAW seems like it would give you quite a bit more flexibility than this right now.

I do think the hardware angle could make this much more interesting, especially if the interface is designed in an interesting way.

But looking at the web version on its own, I'm not really seeing the unique idea or workflow yet compared to what already exists.

That doesn't mean I think you shouldn't build it, obviously. Building exactly the instrument you personally want is a perfectly good reason to make something. I'm mostly curious what you see as the part of AM-1 that it does differently or better than the stuff that's already out there.