Skip to content

Tape saturation: level against reference, distortion and listening

A simple model of tape saturation: a soft limiter with unit gain at low level, tied to a reference level. You set where the gain falls by 1 dB (take it from the tape's data sheet), see the distortion curve against level, and can listen to a file of your own before and after. It is a sketch of the transfer function, not a model of a particular tape.

The model

1. Parameters

dB above reference

From the tape's data sheet. 8 dB is only a starting value chosen by the project, not any tape's.

dBFS (peak)

The tape's 0 dB level (the reference flux) in your file. A signal at this level passes unchanged.

Adds even harmonics (second, fourth). 0 = odd only.

dB

Raises the highs before the limiter and lowers them after, so they saturate sooner. A simplified model.

4× oversampling is built in: no alias products.

Distortion

2. The distortion curve at 315 Hz

Harmonics 2–5 against the fundamental, in percent, as a function of level above the reference. The usual “maximum output level” criterion is 3 % third harmonic at 315 Hz; it is a convention, not a limit of your tape.

Computing…

Listen

3. Listen

A test crescendo or a file of your own (the first 30 s, stays in the browser). Compare original and processed; you can save the result.

The ideas behind the controls

Fix the reference level
Everything is measured against it: set it to the tape's 0 dB level before drawing conclusions.
Compare original and processed
The effect is heard first as peak compression, only then as distortion.
Highs saturate sooner
With high-frequency emphasis, harsh sounds round off before the bass: an approximation of real behaviour.

Common mistakes

  • Believing 8 dB is your tape's value: it is only a start; take it from the data sheet.
  • Applying saturation to a file already at the peak (0 dBFS): you will hear digital clipping, not tape.
  • Taking this model for the emulation of a particular tape.

What this tool cannot tell you

  • It is a static (tanh) transfer function with a few adjustments; real tape has hysteresis, bias and frequency-dependent behaviour.
  • The high-frequency emphasis roughly approximates earlier saturation at high frequencies; it is not measured on tape.
  • You give the reference level and the 1 dB point; the tool holds no tape data.
  • The 3 % third harmonic at 315 Hz definition is a convention for maximum output level, not a standard checked here.
  • File processing uses the first 30 s and works at 4× the sample rate.

Where the formulas come from

  • The tape's data sheet — The source for the level where gain falls and for the reference flux; the tool holds neither.
  • The tanh transfer function — A usual soft-limiter model; the asymmetry and emphasis adjustments are the project's.

Frequently asked questions

Why a reference level?

Tape is aligned to a reference flux; a signal's level matters only against it.

Why oversampling?

A nonlinearity makes harmonics that, above half the sample rate, fold back into the audio band (aliasing). Processing at 4× filters them out.

What are odd and even harmonics?

A symmetric limiter gives only odd harmonics (third, fifth); an asymmetry adds even ones (second, fourth).

Tape saturation: level against reference, distortion and listening · Strunin