A gait is defined by two numbers per limb: its duty factor β (the fraction of the stride cycle a foot spends on the ground) and its phase φ relative to a reference limb. This is the diagram method R. McNeill Hildebrand introduced in 1965 and still used to classify animal gaits today.
stride period T = 1 / stride frequency
foot phase(t) = ((t/T) + φ_leg) mod 1
stance if foot phase < β, swing otherwise
Symmetrical gaits use evenly-spaced phases (trot: diagonal pairs 0.5 apart, β≈0.5). Asymmetrical gaits (canter, gallop) use uneven spacing and a lower duty factor, which is what produces a moment of full suspension — all four feet off the ground.
Whether an animal should walk, trot or gallop at a given speed is predicted well by the dimensionless Froude number (R. McNeill Alexander, 1984), which compares forward speed to the pendulum-like natural frequency of the leg:
Fr = v² / (g·L)
v = forward speed, g = 9.8 m/s², L = leg length
Fr ≈ 0.5 → walk / trot transition
Fr ≈ 2.5 → trot / gallop transition
- Gait buttons — set the duty factor and phase offsets for that gait's footfall pattern.
- Stride frequency — how many strides per second; combined with the gait's stride length this sets forward speed.
- Leg length — rescales the model (like comparing a cat to a horse) and changes the Froude number at a given speed, since Fr depends on L.
- Footfall diagram — one row per leg; a filled bar means that foot is on the ground. The moving line is the current instant, matched to the animated model.