The difference between motion that merely moves and motion that lets a driver read the car is decided in software. A short tour of the stack that decides it.
A motion platform is only as honest as the commands it receives. Speed, force, and travel all matter - but no actuator can recover a cue that arrived late, was filtered away, exaggerated, or built from incomplete data. The difference between motion that merely moves and motion that lets a driver read the car is decided in software - and software is the part of a motion simulator that almost nobody publishes. This is a short tour of ours.
Four stages, one owner
Every simulation reports the car differently: update rates from 60 Hz to 400 Hz; jitter, dropped frames, inconsistent channels, and the occasional invalid value. DDX Dynamo sanitizes and reconstructs that stream - cadence correction, discontinuity rejection, and a tuned blend of interpolation, extrapolation, and DDX PreActiv prediction - into one clean, consistent vehicle state, delivered to the rest of the stack 1,000 times per second.
From that state, Dynamo builds the cue. Surge, sway, heave, pitch, roll, and yaw are a poor cueing architecture; most motion software constructs them as independent channels. Real vehicle dynamics are coupled. Weight transfer through a banked turn is not "roll plus sway"; it is one 3D event, and Dynamo cues it natively in full 3D - slip angle, weight transfer, road detail, orientation, and grip as one coherent solution.
Dynamo then solves the actual platform geometry to determine what every actuator must do to produce it - together, at 1 kHz - and sends the result over a 1 kHz link to our custom controller, where in-house firmware coordinates actuator motion and haptics at an approximately 56 kHz integration tick. No fixed mechanical pivot defines how a DDX platform must feel; its virtual pivot points live in the mathematics, where they can change by vehicle, by display configuration, by application - or by software update.
Driven DynamiX owns every one of those stages. No third-party motion engine sits in the critical path - and there is no layer we cannot measure, attribute, and improve.
The latency we control, and the latency we inherit
Owning the stack makes the numbers attributable. DDX Precision and DDX Prime operate within an approximately 2.5-3.5 ms telemetry-to-motion latency envelope, measured from raw simulation telemetry arriving at Dynamo to accelerometer-detected physical motion onset. Making our own stack that fast solves only the latency we control; the rest lives upstream, in the simulation itself.
Measured at the median in iRacing: input, physics, and telemetry take about 21 ms, and the DDX software / hardware / electrical / mechanical stack adds roughly 3.5 ms on top of that, for a total of 24.5 ms from physical input to motion onset. Run that same telemetry through conventional interpolation smoothing and you add another 17 ms - 42 ms in total. The 3.5 ms is ours. The other 38.5 ms is the part every motion system inherits.
DDX PreActiv changes the problem from reaction to prediction: motion and force feedback respond to where the car is going, not where it was. It predicts the whole car - velocity and acceleration, orientation, per-wheel tire slip and suspension travel, steering-shaft torque - at multiple future time horizons, and Dynamo aligns the motion cues in time with the driver's inputs. DDX PreActiv delivers up to 2.7x the motion-cue fidelity of an optimized linear extrapolator, with a fraction of the harshness. Motion-cue fidelity here is how closely the delivered cue matches the car's real motion in shape and timing, measured in our own harness on held-out captures the predictor never trained on; the multiplier grows with how far ahead the cue is placed.
Hardware is fixed the day it ships. Software compounds.
Our hardware is excellent because the software demands it. A fast, detailed signal exposes flex, backlash, weak servo control, irregular command cadence, and thin safety monitoring - flaws that slower, more heavily filtered systems can conceal. But the machine's character is not frozen into its mechanics. Hardware provides force, stiffness, bandwidth, and control authority; software decides how they are used. With no fixed pivots and no motion behavior baked into the linkage, the same platform gains new cueing strategies, new simulations, new applications, and better prediction - with only a software update. Every improvement reaches simulators already in the field.
Read the full report. Software-Defined Motion walks the complete stack - telemetry ingest and reconstruction, cue construction, motion-controller command construction, and real-time firmware - as well as the measurement boundaries.
Download the white paper (PDF)
Or feel it for yourself: book a demo in Bellevue, WA or Austin, TX.
Study note: Latency figures are medians on one DDX Precision configuration, measured from physical button contact, or from raw telemetry arriving at Dynamo, to the first accelerometer-detected platform acceleration. The 42 ms figure describes a conventional interpolated pipeline, not DDX's measured performance, which is 24.5 ms at the median on that title. PreActiv fidelity is an internal, horizon-dependent comparison with no third-party benchmark; harshness is delivered jerk relative to a real-vehicle reference, not a comfort rating.