I've been playting around with sox and using the trim command it should be fairly simple to split the whole audio into n parts (with a fixed length per part).
However as I intend to split spoken recordings it might happen that a simple splitting will split in the middle of a word.
Is there a way to prevent that and make sure that parts contain "whole words"?
Take a look at the sox silence command on the (archived) sox webpage for manual and Sound eXchange on sourceforge.
sox original.wav new.wav silence 1 0.5 2% 1 2.0 2% : newfile : restart
original.wav - the audio file to be spliced.
new.wav - will be the name of the new audio files with numbers appended to each slice (new1.wav, new2.wav, new3.wav...).
silence - name of the effect.
1 0.5 2% - above_periods, duration, threshold.
1 2.0 2% - below_periods, duration, threshold.