Audio/video sync test
Shows whether the sound in the auditorium runs ahead of or behind the picture, and by how many milliseconds. Countdown, offset finder from -100 to +100 ms, clapperboard scene and an L, C, R and LFE channel check. Flat container, 1:03.
Files and SHA-256
- 2K · XCT-SyncTest_TST-1_F_EN-XX_INT-TD_71_2K_NULL_20261009_NUL_SMPTE_OV.zip · 1,35 GB
SHA-256 49337bdca4ddde7481ad1a4ef3b9b54d0f73fa529eb267d5e447f0766bc51a3a
What the package contains
- 1/4 Countdown
- A countdown from 10 to 0: flash and 1 kHz beep are on the same frame. If you see and hear them apart, sync is off.
- 2/4 Offset finder
- Eleven flashes, each beep shifted by a known number of milliseconds, from -100 to +100 ms in 20 ms steps. The screen shows how early or late the sound is for each one.
- 3/4 Real life
- A ping-pong ball and a clapperboard: every bounce goes "tok", the clapper goes "clack". Confirms the sync on real-looking footage.
- 4/4 Channels
- The same click from L, C, R and LFE in turn. All four should land on their flash; a late LFE is a classic.
- Accuracy
- Sync measured inside the finished DCP (JPEG 2000 picture against PCM sound): worst error 0.13 ms across 26 events.
How to use the test
- 1
Sit at the reference seat: about two-thirds back, on the centre line. Sound travels about 3 ms per metre, so row 20 hears it 1.4 frames late. That is physics, not the system.
- 2
Countdown: flash and beep should come together. If you perceive them apart, go on to the offset finder.
- 3
Offset finder: of the eleven pairs, pick the one where flash and beep feel exactly together. The value above the circle tells how far the sound is shifted in that pair.
- 4
0 ms looks right: sync is fine. +40 looks right: your sound comes 40 ms early, add 40 ms of audio delay in the processor. -40 looks right: your sound comes 40 ms late, remove 40 ms of delay.
- 5
After adjusting, play the test again and confirm on the ball and clapper scene and on the channel check, where a late LFE often shows up.
- 6
The preview on this page is only a demonstration. Measure in the auditorium: the browser, computer speakers and wireless headphones add their own delay.
Technical parameters
- Container
- Flat 1.85:1 · 1998 × 1080
- Duration
- 1:03 at 24 fps (1506 frames)
- Bit rate
- 200 Mbit/s
- Audio
- 16-channel PCM 24 bit / 48 kHz, 7.1 DS layout (ISDCF Doc4), SMPTE MCA labels
- Accuracy
- worst error 0.13 ms across 26 events, measured in the finished DCP
- Authored with
- DCP-o-matic 2.19 from a 4K ProRes 4444 master
- Standard
- SMPTE DCP, unencrypted (no KDM), content kind TST
- Picture
- JPEG 2000 in MXF, X'Y'Z' 12 bit, gamma 2.6 (SMPTE ST 428-1)
- Package
- ASSETMAP, CPL, PKL, VOLINDEX and MXF in a folder ending in _SMPTE_OV, zipped
Versions
- v1.0 (2026-10-09)
- First release: countdown with flash and 1 kHz beep on the same frame, offset finder from -100 to +100 ms in 20 ms steps, ping-pong ball and clapperboard scene and an L, C, R and LFE channel check.
- Sync measured inside the finished DCP: worst error 0.13 ms across 26 events.
All releases: GitHub Releases · Free for projector alignment, commercial use included, under the licence CC BY-ND 4.0
Frequently asked questions
What sync tolerance does DCI allow?
The DCI Digital Cinema System Specification (v1.3, section 7.5.4.2.1) requires the server to play sound to picture frame accurately, i.e. within 41.7 ms at 24 fps. Audio delay must be adjustable by ±5 frames in 10 ms steps (section 7.4.1.8). The rest of the chain, processor, amplifiers and speaker distance, is trimmed with exactly that setting.
When does the audience notice an offset?
The broadcast recommendation ITU-R BT.1359 puts the detection threshold at about 45 ms when sound leads and about 125 ms when it lags. Sound ahead of the picture is worse, because in nature sound always arrives after the picture. The 20 ms steps of the offset finder catch that threshold reliably.
Why does it matter where I sit?
Sound travels about 3 ms per metre, so row 20 hears it about 1.4 frames late. That is physics, not the system. The test is therefore judged from the reference seat, about two-thirds back, on the centre line.
Can I measure sync from the web preview?
No. The browser, computer speakers and above all wireless headphones add their own delay. The preview shows what the test looks like; only the DCP played in the auditorium can be measured.
More test DCPs
Something not working?