Acoustic-phonetic description of infant speech samples: coding reliability and related methodological issues