Video & LED calculator
Video Signal Chain Latency Calculator
Glass-to-glass latency is the sum of every stage between lens and screen, and each stage costs a fraction of a frame to several frames. At 59.94 fps one frame is 16.7 ms; a typical IMAG chain of camera, switcher, scaler, and LED processing lands between 3 and 5 frames. Enter your chain to get the total in frames and milliseconds.
Sensor readout + internal processing; broadcast cameras run about 1 frame.
Receiving card plus panel scan; projectors and broadcast monitors differ.
Wireless links, fiber codecs, or anything specified in milliseconds.
Formulas
Frames to milliseconds
ms = frames / fps × 1000- fps:
- system frame rate (16.7 ms per frame at 59.94)
Chain total
latency = Σ stage delays (frames converted at the system rate)How it works
Every processing stage buffers at least part of a frame: cameras read the sensor and process, switchers align sources to their own timing, format converters and scalers hold a frame to resize it, and LED processors buffer for the receiving cards before the panels scan. None of them are optional once they are in the path, so latency management is chain design: remove stages, do not chase settings.
The audience notices IMAG latency as lip-sync against the live person on stage. Small delays read as tight; somewhere in the region of a few frames the disconnect becomes visible on close shots, which is why many engineers try to keep camera-to-screen under roughly 3 to 4 frames and why low-latency camera and processor modes exist. Distance helps: audiences far from the stage hear the PA late too, which masks screen delay.
Genlocking sources to the switcher removes the frame-sync buffer many switchers add to wild sources, often the single cheapest frame saved in the whole chain. Scalers are the other big spender: feeding the LED processor its native resolution and rate can bypass an entire conversion stage.
Worked example: Corporate IMAG at 59.94: camera 1F, switcher 1F, scaler 0.5F, LED processing 2F
- 1.Total frames: 1 + 1 + 0.5 + 2 = 4.5 frames.
- 2.Milliseconds: 4.5 / 59.94 × 1000 = 75.1 ms.
- 3.Cutting the scaler (native format into the processor) saves 8.3 ms and drops the total to 4 frames.
About 75 ms glass to glass; removing one conversion stage gets under 70 ms.
Typical per-stage delays (verify per model)
| Stage | Typical delay |
|---|---|
| Broadcast camera (sensor to SDI) | 0.5-1.5 frames |
| Production switcher (genlocked source) | 0.5-1.5 frames |
| Frame sync on a wild source | +1 frame |
| Scaler / format converter | 0.5-2 frames |
| LED processor + panels | 1-3 frames |
| Wireless video link | 1-3 frames (model dependent) |
Field notes
- Measure the truth: shoot a running timecode display through the chain and photograph the display and screen together; the frame difference is your real latency.
- Audio delay to the screen PA should match the video path, not fight it; delaying audio to the latest screen is standard practice on big IMAG shows.
Frequently asked questions
How many frames of delay is acceptable for IMAG?
There is no standard number; many engineers aim to keep the full chain in the 2 to 4 frame range at 59.94 because close-up lip sync starts reading as detached beyond that. Screens far from the stage tolerate more because PA arrival delays the audio too.
What adds the most latency in a video chain?
Format conversion and LED processing are the usual heavyweights, one to several frames each. Unsynchronized sources force switchers to add a frame of sync as well; genlock removes it.
How many milliseconds is one frame of video?
1000 divided by the frame rate: 16.7 ms at 59.94/60 fps, 20 ms at 50, 33.4 ms at 29.97, 41.7 ms at 24.