Isochronic Tones The pulse is in the audio

An isochronic tone is one tone switched on and off at a fixed rate. Ten pulses a second is a 10 Hz isochronic tone; two a second is 2 Hz. The name is literal — equal time between pulses. The pitch of the tone and the rate of the pulsing are two separate settings. Conflate them and none of the rest makes sense.

How abruptly it switches is a third question, and it decides more than the rate does.

The modulation is a physical property of the recording. Put the file on an oscilloscope and the level rises and falls; the sound pressure at your eardrum genuinely changes. A binaural beat is the opposite case — neither channel is modulated, and the pulse appears only after the two ears have been combined centrally, which is why it needs headphones.

Amplitude survives everything a playback chain does to audio. A mono downmix, lossy compression, a single small driver, a Bluetooth hop. Nothing about a pulsed envelope depends on two channels staying in step with each other, so there is nothing here for the chain to lose.

The pulse rate gets all the attention. The envelope decides whether the file is listenable.

Gating a tone means multiplying it by something. Multiply a 200 Hz sine by a switch that flips ten times a second and you have not produced a 200 Hz tone with a rhythm laid over it. You have produced a new signal with a new spectrum. Smooth sinusoidal modulation is the tidy case: a carrier plus two sidebands, one either side, spaced at the modulation rate. Three components and nothing else.

A hard on/off switch is the untidy case. A square gate carries an endless harmonic series, and every harmonic puts another pair of sidebands into the spectrum, so energy ends up scattered well away from the carrier. Worse, each switch is an instantaneous jump in the waveform, and an instantaneous jump is a click — broadband, sharp, and repeated at exactly the rate you were hoping to relax to.

That is the whole explanation for why so much isochronic audio sounds cheap. It is not the pulse rate. It is the edges. A few milliseconds of fade at each end of every pulse removes the click completely and leaves the rhythm perfectly obvious.

Duty cycle is the fraction of each period the tone is on. Half is the usual choice, and the setting is audible when it moves. At ten per cent each pulse is a tick with almost no energy behind it, and the file reads as a click track with a tone attached. At ninety per cent the gaps have closed up and what is left is a wobble inside one continuous note. Neither is wrong. Both are a different piece of audio from the same nominal rate.

You can audit a file in about thirty seconds. Play it quietly through one small speaker and listen to a single edge instead of to the rhythm. A gated file ticks on every switch, and the tick sits on top of the tone rather than belonging to it. A file that was modulated has no attack at all — the level swells and falls and there is nothing to catch on.

Turning it down does not help. The click scales with the carrier, so a harsh file stays harsh at every volume: that ratio was fixed when the audio was generated and no playback setting reaches it. Any waveform editor settles the question visually. Zoom in far enough to see one pulse, and the shape of its leading edge is the whole answer.

At 40 Hz with a half-on duty cycle the tone sounds for 12.5 milliseconds and the gap after it is 12.5 milliseconds. A gap that short is not heard as a gap. There is nothing to count and no silence to notice, so what reaches you is a change in the tone's texture — roughness — instead of a rhythm.

The boundary sits somewhere between 15 and 20 Hz. It is a property of hearing, so nothing you change about the carrier, the envelope or the duty cycle will move it.

It cuts across the brainwave bands rather than lining up with them. Delta at 0.5 to 4 Hz, theta at 4 to 8 and alpha at 8 to 13 all sit below it and are heard as rhythms. Beta runs 13 to 30, so a beta pulse crosses the line partway up its own band: at 15 Hz the pulses are still separate events, at 25 Hz they have fused. Gamma, 30 to 100 Hz, is entirely past it.

The carrier is the pitch being gated, and it is an independent choice from the rate. A 2 Hz pulse can ride on 80 Hz or on 800 Hz. Both are 2 Hz isochronic tones and they are nothing alike to listen to.

Very low carriers go wrong for an arithmetic reason. One cycle of a 100 Hz tone lasts 10 milliseconds. At a 10 Hz pulse rate with the tone on for half of each period, one pulse contains five cycles, and five cycles is not enough for the ear to settle on a pitch. The result is a run of soft thumps. You hear the pulsing and never quite hear a note. Move the carrier to 400 Hz and the same 50 millisecond window holds twenty cycles, comfortably enough.

Low carriers also go missing on small hardware. A phone speaker or a laptop driver has essentially no output at 100 Hz, so what reaches you is not a quieter version of the tone. It is the modulation artefacts without the tone underneath.

High carriers fail from the other direction. Hearing sensitivity peaks somewhere in the 2 to 5 kHz region, so a 3 kHz carrier at the same nominal level is far louder than a 300 Hz one, and a piercing tone switching on and off gets unbearable inside a few minutes.

Which leaves roughly 100 to 500 Hz.

Audiology has a routine test for whether an ear is hearing a modulated sound, and it predates the wellness framing by decades. Present amplitude-modulated audio, record from the scalp, look for activity locked to the modulation rate. The result is called an auditory steady-state response. Because it depends on the sound being heard rather than on the patient reporting anything, it can estimate hearing thresholds in a sleeping infant.

So the base claim, that the auditory system follows an acoustic envelope, is not in dispute. It is measured in clinics every day.

An auditory steady-state response is a sensory following response in the hearing pathway. Brainwave entrainment as it is usually sold claims something much bigger: that an internal cortical rhythm is pulled into step, and that this changes attention, mood or sleep. Those are separate claims, and the gap between them is where the evidence runs out.

The best-organised evidence concerns binaural beats. A 2023 PLOS ONE review by Ingendoh, Posny and Heine assembled fourteen of those studies and found the balance against entrainment, with its authors unwilling to close the question. Not one of the fourteen tested an isochronic tone. That is the best-organised evidence any of these formats has, and it is where all of them have to start.

Schwarz and Taylor, in Clinical Neurophysiology in 2005, recorded steady-state responses to 40 Hz binaural and monaural beats. The binaural response appeared only at low stimulus frequencies, around 400 Hz, and became undetectable above 3 kHz. The monaural beat — two tones summed, so the envelope rises and falls in the air as an isochronic pulse does, though the spectrum differs — produced a larger response, at both low and high frequencies. It is easy to read that as proof that acoustically modulated audio is the stronger approach. It is not. What was measured is the size of a cortical response, which is not a benefit to anybody, and nothing about isochronic tones was tested.

One 2025 Scientific Reports experiment, by Melnichuk, Cooper and Hawk, varied the parameters instead of assuming them: eighty participants across sixteen conditions, crossing beat frequency, carrier tone at 340 or 400 Hz, onset timing and the presence of a masking noise. Sixty-four of the eighty made it past attrition and a performance exclusion into the analysis. Its headline was a carrier effect: measured against a matched pure tone, the gamma beats scored better on percent hits when they rode 340 Hz (p = 0.002) and not when they rode 400. A second and separate contrast found higher scores wherever white noise was running behind the tones (p = 0.033). Neither arrangement slowed the decline in vigilance across the task. The EEG followed the beats, but the carrier made no difference to that and the noise weakened it, so the two measures pointed different ways. All of it was binaural audio, which does not transfer to a gated tone, and it is one experiment. None of this establishes that any of these formats treats any condition.

Isochronic is the largest category in Arkhitec's frequency library: 68 presets across all five bands, from delta pulses for bedtime up through gamma. None of them need headphones.

The pulse is modulated, not gated. A sine oscillator drives the gain of the carrier, so the level glides between silence and full depth and there is no switching edge anywhere in the signal path. The click described above cannot occur, because there is no edge for it to occur on.

Carriers run from 200 Hz to 350 Hz, so the whole set sits inside the window above — nothing down where a phone speaker gives up, nothing up where the tone turns piercing. The general series rises with the band: 200 Hz under the delta rates, 250 Hz under theta, 300 Hz from alpha upward.

Layering is the part worth setting up. Put a delta, theta or alpha preset under rain, ocean or singing bowls in the 7-layer mixer, let the Smart Mix Engine auto-balance the levels, and set the sleep timer to end it. The gamma presets are the exception: at 40 Hz there is no rhythm for a soundscape to sit over, so run those bare and quiet.

You can hear an isochronic tone running. Everything good and everything annoying about the format comes from that.

A 14 Hz pulse under a work session is a metronome, and a metronome is something you can lock onto deliberately. A binaural beat gives you nothing audible to lock onto at all. At bedtime the same property turns on you: an audible pulse every half-second is exactly the sort of detail that keeps attention switched on.

A bedside speaker is the compensation. You are not sleeping in earbuds, which is a real claim on bedtime that binaural audio cannot make, and the audible pulse is what you pay for it. If the pulse keeps you awake, a monaural preset at the same rate is the gentler option. Its envelope is one continuous swell rather than something with a top and a bottom to it, it plays from the same speaker, and it is easier to lie in the dark with.

At 10 Hz the tone is only sounding for 50 milliseconds. For a gated pulse that caps how long the fades can be, so slow delta and theta rates have room for a soft envelope while a 20 Hz beta pulse is necessarily crisper — the table above has the numbers. A sinusoidal envelope sidesteps the constraint entirely, because its rise is a quarter of the period and stays smooth at any rate.

Rain will not hide a 4 Hz pulse.

It does not need to. A delta, theta or alpha pulse still reads as a rhythm under broadband sound, so bury it and let it sit under the mix rather than on top: a pulse you have to listen slightly harder for is still entirely there in the waveform. A 40 Hz gamma pulse has no rhythm to preserve, and covering it hides the only cue that it is running.