Describir: Employment of Spectral Voicing Information for Speech and Speaker Recognition in Noisy Conditions