Why slow the music at all
New choreography is learned in pieces, and the pieces are learned slowly first. A slower mix gives athletes time to hear the count, place the skill on it and fix their timing before the speed hides mistakes. It also lets a coach watch the count land: at full speed a stunt group that hits on the 6 instead of the 5 looks nearly right; at 75 percent it is plainly late.
What each speed does
The tempo and the length change together. For a 144 BPM mix and a 1:30 routine:
| Speed | Tempo heard | One 8-count lasts | Routine lasts |
|---|---|---|---|
| 100 percent | 144 BPM | 3.33 s | 1:30 |
| 90 percent | 129.6 BPM | 3.7 s | 1:40 |
| 85 percent | 122.4 BPM | 3.92 s | 1:46 |
| 75 percent | 108 BPM | 4.44 s | 2:00 |
| 60 percent | 86.4 BPM | 5.56 s | 2:30 |
| 50 percent | 72 BPM | 6.67 s | 3:00 |
How the pitch is kept
A record player slowed down sounds lower because each sound wave is stretched. To slow the music but keep the pitch, the sound is cut into short overlapping slices (a few hundredths of a second each), and the slices are laid down further apart than they were taken. Each slice keeps its own pitch; the joins are chosen where the waveforms line up best so they do not click. This family of methods is called overlap-add with waveform similarity. In the browser the player uses the built-in version: the pitch is kept by default when the playback speed changes, and MDN lists that as supported across current browsers since December 2023.
Drums smear slightly at very low speeds because each hit is repeated or stretched between slices. That is why Cheersheet stops at 50 percent: below it the count is hard to hear.
Worked example: a practice track
Worked example. A coach saves a practice track of a 1:30 routine at 75 percent. The file is exactly 1/0.75 times as long as the mix: 2:00, or 5,292,000 samples at 44.1 kHz. Every count gets a click at its slowed time, highest on each 1 and higher on each 5, and a 4 count lead-in (5, 6, 7, 8) before the first 1 when the intro leaves room. The same routine at 50 percent gives a 3:00 file with 7,938,000 samples.
The track is rendered by Cheersheet's own code, sample by sample, so the same routine at the same speed gives the same file every time from the same mix in the same browser. Another browser decodes the mix with its own decoder, so its file can differ slightly.
A practice plan by speed
- 75 percent: teach the section, counts called out loud, clicks on.
- 85 percent: run the section, clicks on, the coach watches the hits.
- 90 percent: run the section, clicks off, athletes count in their heads.
- 100 percent: run it full out, then drop back to 85 for anything that fell apart.
Common mistakes
- Slowing with an app that lowers the pitch. The music sounds wrong and athletes learn the wrong feel.
- Going too slow. Below about 60 percent the count stops sounding like music.
- Never practising at full speed. Timing learned only at 75 percent falls apart at 100.
- Counting a slowed mix at the full speed tempo. The counts move with the music; only the time between them grows.
The count sheet editor plays your mix from 50 to 100 percent with clicked or spoken counts, and loops any section. For a whole practice plan, see teaching counts to a team.
Sources
Checked on 27 September 2026. Count sheets have no single governing standard; what is written here is common practice in our own words, and every number in a worked example is computed by Cheersheet's own engine when the page is built. Where your gym, choreographer or ruleset does it differently, follow them.
- Cheersheet's own engine and its tests. Every worked number on these pages (tempos, 8-count lengths, how long a routine takes at a practice speed, what the detector found on the test tracks) is computed by the same code the sheet runs, when the page is built.
- MDN Web Docs, HTMLMediaElement: preservesPitch. Pitch is kept by default when the playback rate changes; listed as Baseline, widely available since December 2023. Opened 27 September 2026.
- Verhelst, W. and Roelands, M., An overlap-add technique based on waveform similarity (WSOLA), ICASSP 1993. The family of time stretch methods the practice track uses to slow music while keeping its pitch.