Milliseconds to BPM Calculator

A duration on its own does not identify a tempo. 375 ms is a quarter note at 160 BPM, a dotted eighth at 120 BPM, and an eighth note at 80 BPM. All three are true, and which one you want depends entirely on what that duration represents musically.

That is why the note value is an input rather than an assumption. Once you say which subdivision the time stands for, the arithmetic is direct:

where $f$ is how many quarter notes the chosen value spans. A 500 ms duration read as a quarter note gives 120 BPM. The same 500 ms read as a dotted eighth points at 180 BPM instead.

The most common reason to run this backwards is a preset or a reference track. A delay patch that sounds right at 428 ms was probably built at a specific tempo, and knowing which one tells you whether it will translate to your session. A reference mix whose delay you can hear and measure gives you its tempo without needing the project file.

The tool also shows the resulting quarter-note length, so you can sanity-check the answer against something familiar. If the quarter note comes out at 500 ms you know immediately that the tempo is 120 BPM, because that pair is worth memorising.

Sample packs are full of audio with no tempo in the filename. Loop mode solves that from two pieces of information: how long the file is, and how many bars it contains.

The length is exact and the software already knows it. The bar count is the only thing you supply, and it is the easier of the two to get right because you count it by ear rather than measure it. Play the loop, count along, and note how many times you reach a downbeat before it repeats.

From there the arithmetic is straightforward:

A 4-bar loop in 4/4 lasting 7.5 seconds comes out at 128 BPM. A 2-bar loop of 4 seconds is 120 BPM.

Counting bars rather than beats is deliberate. Bars are much harder to miscount than individual beats, because each one starts with something audible: a kick, a chord change, the top of the phrase. Miscounting beats by one gives a wrong answer that looks plausible, while miscounting bars by one usually produces a tempo obviously too fast or too slow to be real.

If the result looks wrong, the bar count is nearly always the reason. Halving or doubling it and rerunning is faster than second-guessing the arithmetic, and the tool shows the total beat count so you can check whether it matches what you hear.

Every tempo is also its own half and its own double. A loop read as 4 bars at 90 BPM is equally 2 bars at 180 BPM: the audio is identical, the counting is different, and neither reading is incorrect.

This is why the tool shows half-time and double-time alongside the main result rather than picking for you. The arithmetic cannot resolve the ambiguity because it is not an arithmetic question.

What resolves it is the drum pattern. Find the backbeat, the snare that answers the kick. If that snare falls on beats two and four of your count, the count is right. If it falls only on beat three, you are counting twice as fast as the music feels and the half-time reading is the musical one. If the snare appears twice per beat, you are counting too slowly.

Genre conventions help as a sanity check. House and techno sit around 120 to 130 BPM, hip-hop commonly 80 to 95, drum and bass near 174, trap often notated at 140 while feeling like 70. A calculated 87 BPM in a folder of drum and bass is almost certainly 174 counted at half speed.

Trap is the useful illustration of why both numbers matter. Producers write it at 140 BPM because the hi-hats subdivide fast, while the kick and snare pattern feels like 70. Both numbers are correct and each is the right one for a different decision: 140 for the hi-hat programming, 70 for the groove.

A loop that computes to 119.98 BPM is a 120 BPM loop. The fractional part is almost always an artefact of the file rather than the music.

Trailing silence is the usual cause. A loop exported with a few milliseconds of empty audio at the end is slightly longer than the musical content, which drags the calculated tempo down. A file trimmed a few milliseconds short does the opposite and reads slightly fast. Either way, the deviation is small and the nearest round number is the real tempo.

The practical rule is to round when the result is within about 0.1 BPM of a whole number, and to look closer when it is not. A result of 118.6 BPM is not a rounding artefact and suggests the bar count is wrong or the source was never on a grid.

Some sources genuinely were never on a grid. Music recorded to tape, transferred from vinyl, or played without a click drifts continuously, and any single measurement is a local average rather than a fixed tempo. Live recordings from before the click-track era commonly wander by a BPM or two across a song, and no calculation will produce one true number because there is not one.

Sample rate conversion introduces its own small offsets, as does any pitch or time processing applied to the loop before you received it.

When the number matters, the fix is the same in every case: trim the loop precisely to its musical length at a zero crossing, then measure again.

Leaving the note value at quarter note when the time is something else is the error that produces the most confidently wrong answers. A dotted eighth delay entered as a quarter note reports a tempo two thirds of the truth, and the result is plausible enough to go unquestioned.

Miscounting the bars is the loop-mode equivalent. Counting a 2-bar loop as 4 bars halves the reported tempo exactly, which is the single most common reason a result comes out at an implausible speed.

Mixing seconds and milliseconds is easy to do when moving between the two modes. Delay mode takes milliseconds, loop mode takes seconds, and a loop length typed as 7,500 instead of 7.5 produces a nonsensical answer immediately, which at least makes it self-correcting.

Assuming the DAW's own detection is right is worth resisting. Automatic tempo detection is generally reliable on clean four-on-the-floor material and unreliable on sparse, swung or heavily processed audio. Where they disagree, counting bars by hand is the tiebreaker.

Forgetting that time signature is an input catches people working outside 4/4. A 4-bar loop in 3/4 has 12 beats, not 16, and using the wrong beats-per-bar scales the answer by the ratio between them.

Finally, do not average across a section with a tempo change. Measure the part you intend to use, because a value computed across a transition describes neither side of it.