The use of speckle decorrelation techniques to calculate the displacement of a moving transducer has demonstrated promise. We introduce a technique to estimate distance between image planes without assuming a constant transducer velocity. We developed a theoretical analysis of the uncertainty in estimated plane spacing as a function of speckle size, patch size and the number of planes used for normalization. The best estimates of plane spacing are obtained when the distance between acquired image planes is of the order of half the speckle size. In this region, the uncertainty in estimated plane spacing was < 15% for an 8.1mm (axial) x 9.1mm (lateral) patch, increasing to 33% for an 8.1mm x 1.5mm patch. Patch size is limited to regions of fully developed speckle, and also by brightness gradients in the image. Although brightness gradients seem insignificant on intensity images, they can cause a large bias in plane spacing estimates made using linearized data. Another major source of bias in plane spacing estimates was the number of planes, Nz, used to calculate the normalization factors (averages of brightness and squared brightness). Optimum Nz was found to be 5 to 10 planes, depending on plane spacing.