In acoustic well logging, for each inversion depths of a well at which logging of data occurs, acoustic log signals representative of waveforms received at acoustic receivers are processed in a frequency domain to derive field dispersion curve(s). A neural net is operated to generate formation shear slowness value(s) from the curve(s), and resulting signal(s) indicative of shear slowness values are saved, transmitted, plotted, printed or processed. An apparatus for carrying out the method includes a logging tool having at least one activatable acoustic wave source; spaced and acoustically isolated therefrom in the logging tool an array of acoustic detectors that on the detection of acoustic wave energy generate electrical or electronic log signal(s) characteristic of acoustic energy waves detected by the acoustic detector(s); and at least one processing device associated with or forming part of the logging tool for processing the log signal(s).

A method for acoustic logging a wellbore includes making sonic logging measurements while rotating a logging while drilling tool in a wellbore, the sonic logging measurements including low frequency measurements and high frequency measurements; estimating a low frequency slowness of the subterranean formation from the low frequency measurements; estimating a high frequency slowness of the subterranean formation from the high frequency measurements; and classifying the subterranean formation as homogeneous when a difference between the low frequency slowness and the high frequency slowness is less than a threshold and heterogeneous when the difference is greater than the threshold.