BPM to MS Delay Time Calculator

A delay that lands exactly on the eighth note disappears into the groove. Set the same delay to 280 ms on a 120 BPM track and it fights the drums, smearing the backbeat instead of reinforcing it. That single difference is why most producers reach for a conversion table before they touch a delay plugin.

Tempo sync works because rhythmic music is built on subdivisions of one pulse. When an effect repeats at a subdivision of that pulse, every repeat coincides with something already in the arrangement: a hi-hat, a snare, the next chord stab. The ear reads those coincidences as intention. When the repeats fall between the grid points, the ear reads them as blur.

The same logic applies well beyond delay. Reverb pre-delay decides whether a vocal sits in front of its own tail or inside it. A compressor release timed to a subdivision lets the gain recover exactly as the next hit arrives, so the pumping breathes with the track instead of against it. Tremolo and auto-pan sweep in time with the bar rather than drifting across it.

Free-running timing is still a legitimate choice. Slap-back echoes on vocals often sound more human at 90 ms regardless of tempo, and a slowly drifting chorus is the point of a chorus. The difference is that unsynced timing should be a decision, not an accident. Knowing the grid value first tells you how far off the grid you have moved.

One minute holds 60,000 milliseconds. Divide that by the number of beats in a minute and you get the length of one beat, which in common time is the quarter note:

At 120 BPM the quarter note is 500 ms. Everything else on the table follows from halving and doubling. An eighth note is half a quarter, so 250 ms. A sixteenth is half again, 125 ms. A whole note is four quarters, 2,000 ms. The chain of powers of two is why the numbers look so clean at round tempos and so awkward at tempos like 137 BPM.

Two modifiers sit on top of that chain. A dotted note adds half its own length, so it lasts $1.5$ times the straight value. A triplet squeezes three notes into the space of two, so each one lasts two thirds of the straight value. At 128 BPM the straight eighth is 234.38 ms, its dotted version 351.56 ms, and its triplet 156.25 ms.

The last column converts duration into frequency, which matters for anything that oscillates rather than repeats:

A 500 ms quarter note is 2 Hz. That is the number a synth LFO wants when you need it to complete exactly one cycle per beat.

Delay choice is mostly a question of how much space the arrangement has. A dotted eighth is the workhorse of modern pop and electronic music because it repeats slightly off the obvious pulse, filling gaps without doubling anything. A straight eighth reinforces the existing rhythm and thickens it. A quarter-note delay is spacious and works when the track is sparse, since each repeat gets room to decay.

For vocals, short is usually safer. A 1 / 32 or 1 / 64 value reads as thickening rather than echo, which is exactly what reverb pre-delay is for. Setting pre-delay to the sixty-fourth note keeps the dry vocal in front for a few milliseconds before the tail arrives, so intelligibility survives a long reverb.

Compressor timing uses the same table in reverse. Pick a release that lands near a subdivision the drums actually play. On a 90 BPM hip-hop beat the quarter is 666.67 ms, so a release around 333 ms recovers in time for the next eighth. Too fast and the compressor chatters; too slow and it never lets go.

Two habits make the table pay off faster. Write the quarter-note value on a sticky note when you start a session, because every other number is a halving or doubling away. And check whether the plugin is already synced: if it shows note values rather than milliseconds, the conversion is happening for you and typing a millisecond value will override it.

Practising against a click while you set these values makes the relationships obvious. A metronome at the session tempo gives you the pulse to listen against.

Many synths and effects express modulation speed in Hertz rather than note values. Hertz counts cycles per second, so converting is a matter of asking how many times per second the chosen note length fits.

At 120 BPM a quarter note is 500 ms, which is half a second, so an LFO at 2 Hz completes one full cycle per beat. An eighth note at the same tempo is 250 ms, giving 4 Hz. Slower modulation runs the other way: a whole note at 2,000 ms is 0.5 Hz, one cycle every two seconds.

This matters most for tremolo, auto-pan and filter wobbles, where the movement is meant to be felt as rhythm. It matters less for vibrato and chorus, which usually sit between 4 and 7 Hz for musical reasons that have nothing to do with the tempo.

Watch the waveform shape as well as the rate. A triangle LFO at 2 Hz peaks twice per beat, once up and once down, so it can read as an eighth-note pulse even though the rate says quarter note. A sawtooth resets once per cycle and reads as the rate you set. When a modulation feels twice as fast as expected, the waveform is usually the reason rather than the arithmetic.

For rhythms that divide the bar unevenly, such as three against four, the plain subdivision table stops being enough and a polyrhythm tool becomes the better reference.

The most frequent error is using the bar instead of the beat. At 120 BPM in common time the bar is 2,000 ms, not 500 ms. Typing the bar length into a delay gives repeats a quarter of the intended speed, which sounds like a mistake in the arrangement rather than a mistake in the setting.

Doubling when you meant to halve is the second. Going from a quarter to an eighth makes the value smaller, because the note is shorter. It is easy to reverse in a hurry, and the result is always musical enough to be confusing rather than obviously wrong.

Then there is the plugin that is already synced. If a delay shows 1/8 D rather than a millisecond figure, it is reading the host tempo and will follow tempo changes on its own. Typing a millisecond value there often switches it to free mode, so it stops following when the tempo automation moves.

Small display differences are usually rounding, not error. A DAW showing 351.6 ms where the table says 351.56 ms has simply truncated the readout. The underlying arithmetic is identical, and the difference is far below anything audible.

Finally, tempo changes invalidate manual values. A section that moves from 120 to 128 BPM needs new numbers for every hand-typed delay, which is the main argument for using a plugin's own sync when it has one and reserving manual entry for the effects that do not.