Ultrasound Frame Timestamps: Define the Event Before Synchronizing the Clock

ALEVA — California, USA

Engineering question: when an ultrasound frame is paired with a tracker, an ECG stream or another instrument, what event does its timestamp actually identify?

A shared start is not a complete timing specification

A March 2026 peer-reviewed Scientific Data paper combines ultrasound, motion capture, electromyography and accelerometry. Its synchronization analysis explicitly considers how an external trigger arrives relative to an ultrasound frame boundary. The transferable lesson is that starting two recordings together does not establish the exact acquisition time of every frame. The reported experimental performance belongs to that research setup, not to ALEVA hardware. [1]

For an OEM interface, specify whether a timestamp marks the first transmission, the end of acquisition, completion of reconstruction or arrival at the host. Also define the acquisition interval represented by a compounded frame. These are proposed interface requirements; a single timestamp cannot describe every event in that interval.

Separate offset, drift and delivery delay

A useful first-order clock model is t_reference = a × t_device + b. Here, b represents a time offset and a represents the relative clock rate. A start marker can help estimate an offset; repeated reference events are needed to evaluate whether that offset changes during a recording. The 2025 Syntalos paper describes continuous analysis and correction of device timestamps for multimodal experiments. It is supporting synchronization research, not an ultrasound product benchmark. [2]

Consider a deliberately simplified example: two clocks differ by a constant 20 parts per million, with no correction. After 600 seconds, the accumulated timing difference is 600 × 20 × 10⁻⁶ = 0.012 seconds, or 12 milliseconds. This is an analytical illustration, not a measured device result or an acceptance threshold.

Host arrival time also includes transport, buffering and scheduling delays. Recording arrival time is useful for diagnosing delivery behavior, but it should not silently replace an acquisition timestamp. Preserve both when the interface provides them, together with their clock domains and units.

Preserve timing through export

DICOM PS3.3, edition 2026d, distinguishes nominal Frame Time from Frame Time Vector, which describes time increments between frames. It separately defines Recommended Display Frame Rate. These fields serve different purposes; a smooth playback rate is not evidence that the acquisition intervals were uniform. Implement the attributes required by the selected information object and verify how the receiving software interprets them. [3]

A practical verification sequence

Use a reference event that the ultrasound acquisition and the companion instrument can both observe. Document where each observation occurs in its signal chain. Compare repeated events near the beginning, middle and end of the intended recording duration, under representative host load. Keep the original timestamps and frame identifiers, then examine residual timing errors after calibration.

Exercise interruptions as separate cases: pause/resume, reconnect, counter rollover and a changed acquisition configuration. A gap, duplicate frame or clock reset should remain identifiable in the exported record. Repeat the timing comparison after export and import; checking only the live display leaves the downstream measurement chain untested.

The OEM/ODM decision

Hardware synchronization can reduce uncertainty but adds interfaces and integration work. Software alignment can accommodate more devices, while depending on timestamp quality, reference observations and a clock model that remains appropriate. Interpolation can align samples mathematically; it cannot recover an event that was never measured.

Before selecting an ultrasound engine, agree on timestamp semantics, clock-domain mapping, interruption behavior and verification evidence. Choose tolerances from the intended measurement task. Treat the example above and the cited studies as engineering guidance, not a claim of clinical suitability or ALEVA product performance.

Sources

1. Pallarès-López et al. A multimodal biomechanics dataset with synchronized kinematics and internal tissue motions during reaching. Scientific Data 13, 709, 18 March 2026. Peer-reviewed data descriptor. DOI: 10.1038/s41597-026-07019-3.

2. Klumpp et al. Syntalos: a software for precise synchronization of simultaneous multi-modal data acquisition and closed-loop interventions. Nature Communications 16, 708, 15 January 2025. Peer-reviewed research. DOI: 10.1038/s41467-025-56081-9.

3. DICOM PS3.3 2026d, C.7.6.5, Cine Module. Official standard, accessed 25 October 2026. Official reference (edition displayed: 2026d).