The invention generally relates a transducer apparatus in a device to obtain high signal-to-noise-ratio signals including speech in a noisy environment by a non-acoustic transducer or sensor adapted in two ways. One, adapted to sense free-field acoustical sounds and whose sensitivity is directive, and arranged to be most sensitive to a direction or axis according to the position or orientation of the device. Two, adapted to sense vibrations, movement or acceleration on the skin of the user of the device arising from the voice of the user. Embodiments and variations of the invention include where the two adaptions are combined, and with acoustical microphones. In the case of adaption two and with a microphone, a transducer apparatus resembling the characteristics of a close-talking microphone can be derived.