Vowel height corresponds quite closely to the center frequency of the lowest resonance in the vowel token, also called the first formant or F1. The F1 ranges between 200 and 800 Hz for a male voice, where low F1 values indicate high (close) vowels, and high F1 corresponds with low (open) vowels. The center frequency of the second lowest resonance, the second formant or F2 is a good estimate of the length of the mouth cavity, i.e., it codes the constriction place of the vowel along the front-back dimension. Since the length of the mouth (front cavity) is also affected by lip posture (lip rounding adds a centimeter of length to the front cavity, lip spreading shortens the front cavity), the F2 codes constriction place and lip rounding simultaneously. The F2 ranges between 600 Hz for back rounded vowels ([u]) to 2400 Hz for front spread vowels ([i]). The values mentioned here are representative of the resonances of an average male vocal tract. Since female vocal tracts are about 15% shorter than those of males, the formant values for female voices are higher than those of males by 15%. By plotting vowels in a plane defined by the F1 value (from top to bottom) and the F2 value (from right to left) the vowels are arranged in a manner that reflects their position in a traditional IPA vowel diagram. Since the human ear is less sensitive to formant differences at higher frequencies than at lower frequencies, the hertz values are rescaled to Bark units so as to reflect these auditory sensitivity differences.

Acoustic data have been published on the vowels of Persian (e.g., Ansarin, 2004; Salehi et al., 2009; Mohamadi et al., 2011; Aronov et al., 2017; Ghaffarvand Mokari et al., 2017; Jones, 2019) and of Azerbaijani (Peivasti, 2012; Ghaffarvand Mokari, Werner, 2016). Acoustic data on the vowels of British English produced by monolingual Persian-speaking EFL learners can be found in Pillai and Delavari (2012). Acoustic analyses of the English vowels produced by monolingual Azerbaijani EFL learners were reported by Ghaffarvand Mokari et al. (2013). No data are available in the literature on the production of the vowels of American English by matched groups of monolingual Persian and early bilingual Azerbaijani/Persian EFL learners. This is what the present study aims to provide.