A sound processing process, device and software are disclosed which seek to improve pitch perception, with particular application to auditory prostheses. After input sound signals are processed into channels, an algorithm is applied to selectively increase the modulation depth of the envelope signals. In certain embodiments, the channelized signals are adjusted in timing so as to align the phase of modulated envelope signals in different channels. This results in provision of synchronous (phase aligned) modulation periodicity across channels and hence less pitch ambiguity for listeners to the processed signal, or for application of the signal to hearing prostheses such as cochlear implants. In some embodiments, a broadband envelope signal is used to modulate the level of the narrow band channel signals, so that voicing frequency modulation information in the broadband envelope signal is provided in all narrow band channel signals and the phase of modulated signals in the channels are aligned. A preferred form uses an envelope signal which is modified to increase its modulation depth, and is normalized so as to allow for modulation of the channel signals.