Acoustic wave velocities in standing trees or live softwood species were measured by the time-of-flight (TOF) method. Tree velocities were compared with acoustic velocities measured in corresponding butt logs through a resonance acoustic method. The experimental data showed a skewed relationship between tree and log acoustic measurements. For most trees tested, observed tree velocities were significantly higher than log velocities. The results indicate that time-of-flight measurement in standing trees is likely dominated by dilatational or quasi-dilatational waves rather than one-dimensional plane waves. To make appropriate adjustments of observed tree velocities, two analytical models were developed for the species evaluated. Both the multivariate regression model and dilatational wave model were effective in eliminating deviation between tree and log velocity and reducing variability in velocity prediction.