A comparative and acoustic analysis of phoneme /g/ in spontaneous speech in Basque Country, Andalusia and Madrid

In this work, we perform the acoustic analysis of the features of harmonicity, sonority, duration and intensity of each /g/ sound extracted from a sample of spontaneous speech from a corpus collected from recordings of local television programs in Andalusia, Madrid and the Basque Country. This corpus consists of 1292 utterances, in which the velar appears on 151 occasions in different contexts. These sounds are described and categorized thanks to the Praat spectrographic analysis program and a comparison of their features among the three dialectal communities that the corpus contains. The most relevant results allow us to conclude that undescribed realizations of [ɣ] occur in spontaneous speech; that the intensity curves of these approximant sounds yield significant figures for their characterization; and that the frequency of production of the approximant [ɣ] is greater in Andalusia than in Madrid or Euskadi.