deriva tape machine

At first glance it is simple: nine controls, a response graph, two reels that spool. Underneath is a physical model of magnetic tape, written from published research papers.¹

Press the tape and drag it. It is a performance, not a setting.

Put it on a mix bus for glue, on drums for weight, or run a whole track through it at the slowest speed when you want it to sound like it has lived a life.

Input and drive set the record level and how hard the tape is magnetized, and bias runs from a badly aligned machine to a clean one. Wow & flutter, wear and hiss are a well-kept deck until halfway, and past that the transport stumbles, the tape breathes and drops out, and the noise floor keeps going well past what a real machine would give you.

The speed lever changes everything at once: losses, head bump, flutter rates, hiss. Slow tape is dark tape. Tone is the desk's EQ, for when you need to dial back some of those high frequencies.

Every control is a drawing that shows its own value, and one line explains whatever you touch. The colours are yours, chassis, display and signal, each with hue, saturation and brightness. Seven presets, from a serviced deck to a seasick one, and room for your own.

ON THE IPAD

deriva is an Audio Unit (AUv3) for any host that loads them: Logic Pro, GarageBand, AUM, Cubasis, Loopy Pro, Drambo and the rest. Open it on a track, or on the mix, the way you would any effect.

The app itself is the same machine, full screen, with the iPad's audio input running through it. Plug in headphones or an audio interface and play through it: a guitar, a synth, a voice. On the built-in speaker alone the input is muted, because the speaker would feed the microphone, and the app says so.

Requires iPadOS 15 or later. The desktop version for macOS, Windows and Linux is a separate purchase at eikremaudio.com.

¹ Jatin Chowdhury, "Real-Time Physical Modelling for Analog Tape Machines", Proc. DAFx-19, Birmingham, 2019 (CC-BY 3.0). Otto Mikkonen, Alec Wright, Eloi Moliner and Vesa Välimäki, "Neural Modeling of Magnetic Tape Recorders", Proc. DAFx-23, Copenhagen, 2023. deriva is an independent implementation written from the published papers.