Background in performance. Most often, the western lyrical singing technique involves the use of three singing voice registers: The chest register for the low sounds, the male-falsetto or female-head register for the high ones, and the mixed or medium register between them. According to the authors, the so-called mixed voice covers different realities: It is either a full register ranging from B3 to F5, or a narrow area around primo passagio, from A3 to F4. The purpose of practicing voix mixte is to homogenize chest and head voices in their overlapping area. The true nature of this mixing needs to be clarified. Background in acoustics. When producing a succession of sounds, from the lowest to the highest one, two main different laryngeal mechanisms, M1 and M2, are used one after the other (Henrich, 2003). In the median part of the tessitura where the so-called voix mixte is found (from 200 Hz to 400 Hz), these two laryngeal mechanisms can be used, but their dynamical possibilities and the resulting sound spectral properties are very different. It has been shown that voix mixte register is not produced by an intermediate laryngeal mechanism (Chuberre, 2000) and that two different registers can be considered: The one produced in M1 (mx1) and the other produced in M2 (mx2) (Castellengo, 2004). It is therefore necessary to identify the phonatory adjustments which allow the singers to produce, inside a given mechanism, sounds for which timber is closer to the one obtained when using another mechanism. Aims. The aim is to identify the mechanism in use for mixed voice productions and to compare on same-pitched sounds the spectral characteristics of mx1 and mx2 productions with the ones commonly found using laryngeal mechanisms M1 and M2. Main contribution. The phonetogram shows that the dynamic in the overlap area is larger for sounds produced in M1 than in M2, and shifted to higher vocal intensities. Acoustics analyses confirm the spectral similarities between mixed sounds produced in a given laryngeal mechanism and the sounds commonly found in the opposite mechanism. The imitation of M2-like voice quality for mx1 is achieved by a decrease of vocal intensity, which goes with a decrease of high-harmonics spectral energy. The imitation of M1-like voice quality for mx2 requires an increase of vocal intensity, and timbre enrichment. Implications. Literature about singing voice registers is very confused (Henrich, 2006). However, except for very few authors who defend the hypertrophy of one unique register, typical of a pedagogy historically linked to the XXth century, singing teachers recognize that managing transitions by the use of voix mixte is one of the most important part of a singer's education. Practicing voix mixte is very subtle and requires laryngeal and resonantial synergies to produce a desired sound quality while extending the physiological limits. Our investigation aims to clarify the terminology to reconcile acousticians and singing teachers in explaining the different voix mixtes by clearly distinguishing notions of laryngeal mechanisms and registers.