Free · no login · nothing uploaded
Transition timer.
When to press play on the next record so its phrase lands on this one's, how long the blend takes in seconds, and exactly how many bars an unmatched tempo survives before it flams. Drop both tracks and it plans the mix against their real grids.
Playing out
Drop the track
Or type the tempo below. Nothing is uploaded.
Coming in
Drop the track
Or type the tempo below. Nothing is uploaded.
Blend length
Tempo
A 32-bar blend takes
60.0s
At 128.00 BPM. Drop both records in above and this becomes a plan: where to press play, where the new drop lands, and how much of the outgoing track is left underneath it.
Matching the tempo
+1.59%
on the incoming track's pitch fader takes 126.00 to 128.00. That is +0.27 semitones — 27 cents. Type the incoming key above and it will tell you where the pitch puts it.
Leaving them free-running
29.8 ms a bar
Audible as a flam within 3 beats and a trainwreck by bar 2. Over the whole 32-bar blend the two grids separate by 952 ms — 2.0 beats.
How long an unmatched tempo survives
At 128 BPM in 4/4, with both decks free-running. A flam is two transients you can hear as separate rather than as one, and 20 ms is the conservative line.
| Out by | Slip per bar | Flamming after | Wrecked after |
|---|---|---|---|
| 0.1 BPM | 1.5 ms | 13.7 bars | 41.0 bars |
| 0.25 BPM | 3.7 ms | 5.5 bars | 16.4 bars |
| 0.5 BPM | 7.3 ms | 2.7 bars | 8.2 bars |
| 1 BPM | 14.5 ms | 1.4 bars | 4.1 bars |
| 2 BPM | 28.8 ms | 0.7 bars | 2.1 bars |
| 4 BPM | 56.8 ms | 0.4 bars | 1.1 bars |
The number nobody publishes
Half a BPM is three bars.
Two decks at different tempos separate linearly, and the rate is much faster than it feels.
Beat k of a record at 128 BPM happens at k × 60/128 seconds. Beat k of a record at 127.5 happens at k × 60/127.5. Subtract one from the other and you have the whole of beatmatching: the two grids separate by a fixed amount every beat, forever, and nothing about the music slows it down.
At 128 BPM, half a BPM out is 7.3 milliseconds of slip per bar. Two transients about 20 milliseconds apart stop sounding like one drum and start sounding like a flam, so half a BPM is audible after 2.7 bars — before you have finished the first phrase — and a trainwreck by bar 8.
A tenth of a BPM is closer than a pitch fader can be set by hand; one pixel of travel on a ±8% fader is worth more than that. It still only buys 14 bars. This is not an argument against beatmatching by ear — it is a description of what beatmatching by ear actually is, which is a correction every few bars for the length of the blend.
| Tempos differ by | Slip per bar | Flamming after | Wrecked after |
|---|---|---|---|
| 0.1 BPM | 1.5 ms | 13.7 bars | 41.0 bars |
| 0.25 BPM | 3.7 ms | 5.5 bars | 16.4 bars |
| 0.5 BPM | 7.3 ms | 2.7 bars | 8.2 bars |
| 1 BPM | 14.5 ms | 1.4 bars | 4.1 bars |
| 2 BPM | 28.8 ms | 0.7 bars | 2.1 bars |
| 4 BPM | 56.8 ms | 0.4 bars | 1.1 bars |
Closed form, not a simulation — every row is 60 × (1/bpmB − 1/bpmA) × 4, and you can check any of them on a calculator. 20 ms is the conservative line for two transients fusing into one percept; kicks are more forgiving than that and hats are less.
The length
How long a blend actually takes
Bars, not seconds — because the seconds change with the tempo and the phrasing does not.
Nobody brings a record in for “forty seconds”. They bring it in over sixteen or thirty-two bars, and how long that takes depends entirely on the tempo. Thirty-two bars is a minute at 128 and 44 seconds at 174, which is why a blend that feels generous in house feels endless in drum and bass.
The thing to keep whole is the phrase count, not the seconds. A 24-bar blend is not “a bit shorter than 32” — it is three phrases against four, so the two records are half a phrase apart by the end and every change in one lands in the wrong place in the other.
| Bars | 120 BPM | 128 BPM | 140 BPM | 150 BPM | 174 BPM |
|---|---|---|---|---|---|
| 8 bars | 16.0s | 15.0s | 13.7s | 12.8s | 11.0s |
| 16 bars | 32.0s | 30.0s | 27.4s | 25.6s | 22.1s |
| 32 bars | 64.0s | 60.0s | 54.9s | 51.2s | 44.1s |
| 64 bars | 128.0s | 120.0s | 109.7s | 102.4s | 88.3s |
What it costs
Matching the tempo is a key change.
Pulling a record onto another one's tempo moves its key, and the wheel does not move the way you expect.
With key lock off, a pitch fader is a speed control and the key goes with it. Twelve semitones is a doubling of frequency, so the shift is 12 × log₂ of the tempo ratio: a 126 record pulled up to 128 moves about +0.27 of a semitone, which is under a quarter tone and leaves the Camelot code alone. Around 6% you are a full semitone out.
And a semitone is not one step round the Camelot wheel. The wheel is the circle of fifths, so one semitone up is seven positions clockwise — 8A up a semitone is 3A, not 9A. That is the single most counter-intuitive thing about the notation and the reason a pitched record never lands where a DJ expects. The key transposer is the page for that on its own.
Key lock turns the trade around rather than removing it: the key stays put and the artefacts arrive instead. The key and tempo changer measures exactly what that costs, and the answer is that the mode with no artefacts at all is the plain resample nobody markets.
What it does not know
It plans the mix, it does not pick the records
Everything here is timing. It has no opinion about whether these two records belong together — whether the keys sit right, whether the energy goes the way you want it to, whether the second one is the obvious follow to the first. That is a different question and score my transition answers it against the catalog.
It also does not know about the room, which is the variable that decides how long a blend should really be. A 64-bar blend is a beautiful thing at eleven o'clock and a mistake at two.
Frequently asked
Common questions
How long should a DJ transition be?+
Thirty-two bars is the club default and it is 60 seconds at 128 BPM. Sixteen is the working length for faster styles — 22 seconds at 174, which is already a long time to hold two drum and bass records together. The number that matters is not the seconds, it is that the length is a whole number of phrases, because that is what decides whether the two records change together or against each other. A 24-bar blend is not "a bit shorter than 32", it is three phrases against four.
How close do two BPMs have to be to mix without sync?+
Much closer than people think, and it is arithmetic rather than opinion. At 128 BPM, half a BPM out slips 7.3 milliseconds every bar, so the two records are audibly flamming after about 3 bars and unusable after 8. Even a tenth of a BPM — closer than you can set a pitch fader by hand — only buys about 14 bars. That is why beatmatching by ear is a continuous correction rather than a thing you do once, and why a beat grid is worth having.
When should I press play on the next track?+
On a downbeat that starts a phrase in the record already playing, with the incoming track cued to its own first downbeat. Those two conditions are what make the phrases line up; everything else is taste. Drop both records here and it names the moment — the timecode and the bar number — and shows you where the incoming drop then lands relative to the outgoing one.
What does matching the tempo do to the key?+
With key lock off, a pitch change is a speed change, so the key moves with it. Twelve semitones is a doubling, so the shift is 12 × log2 of the tempo ratio — pulling a 126 record up to 128 is about +0.27 of a semitone, which is under a quarter tone and leaves the Camelot code where it was. Get to about 6% and you are a full semitone out, which moves the key seven positions round the Camelot wheel rather than one. This page shows both numbers.
Is it cheating to use sync?+
The arithmetic on this page is the honest answer to that argument: two free-running decks separate at a fixed rate and it is a fast one, so a DJ mixing by ear is making a correction every few bars for the whole blend. Sync removes that particular chore and nothing else — it does not choose the record, phrase the mix, work the EQ or read the room. What it is genuinely bad at is a record with a wandering tempo or a badly drawn grid, where it will confidently hold the wrong thing.
Why does it want both files?+
Because the useful question is not "how long is 32 bars", it is "where does the new drop land in the record that is already playing". That needs the tempo, the first downbeat and the structure of both, which means reading both. With no files at all it still answers the blend length, the drift and the pitch — those are the parts that only need two numbers.
Does it upload anything?+
No. Both files are decoded and analysed in this browser tab. There is no upload endpoint behind the page and no network request carries your audio.
Explore
Beatmatch trainer · DJ ear training · Phrase counter · DJ metronome · Cue point calculator · BPM compatibility · Pitch & tempo · Score my transition · Key compatibility · Blend · Beatmatching · All tools
See which tracks actually fit this one
DJ terms on this page: Blend · Transition · Beatmatching · Sync · Pitch fader · Key lock · Beatgrid drift · Phrase · Full glossary