Klaske van Sluis

72 4.2. Methods with the help of a substitute voice. The preferred substitute voice method in the Netherlands is tracheo-esophageal speech, with help of a voice prosthesis. Voice is generated by vibrations of the pharyngo-esophageal segment. Tracheo- esophageal speech is usually perceived as rough, irregular and reduced in dy- namic range [1]. Several studies have described the acoustic characteristics of tracheo-esophageal speech [2–4]. Weak periodicity and a high noise component are characteristics of tracheo-esophageal speech [5]. Tracheo-esophageal speak- ers have higher values for jitter and shimmer compared to laryngeal voices [1]. This is a result of using the PE-segment as voicing source, as the tissue is less suited to sustain stable vibrations compared to the vocal folds. Jitter and shim- mer will not be discussed in this paper. The focus is on fundamental frequency (F0), harmonics-to-noise ratio (HNR), percentage voiced (%V), and maximum voicing duration (MVD). The main difficulties for tracheo-esophageal speakers regarding intelligibil- ity are in the production of word-initial consonants, rather than consonants in word-final position [6]. Tracheo-esophageal speakers also encounter difficulties distinguishing between voiced and voiceless consonants, voicing vowels, main- taining pitch, phrasing, and producing /h/ and /k/, the consonants that are produced in areas most affected by the operation [6]. According to Jongmans et al. [6], listeners often confuse laryngectomees’ ini- tial plosives. In Dutch, word-initial voiced plosives are produced with prevoicing (a negative voice onset time), while voiceless plosives are produced without as- piration (zero voice onset time) [7]. Van Alphen [7] suggests that presence of prevoicing is the most reliable cue to voicing distinction for listeners. In the study of Jongmans et al. [6] prevoicing is not taken into account. Jongmans et al. [6]suggest that in tracheo-esophageal speech, the speaker’s effort to create a voicing contrast affects the duration of the burst of the plosive as well as the duration of the following vowel. The aim of this research was to investigate how the acoustic features (prevoicing, burst duration and vowel duration) of initial plosives /t/ and /d/ change in Dutch speaking individuals before and after total laryngectomy. 4.2 Methods 4.2.1 Participants Seventeen subjects who underwent total laryngectomy were included in the study (2 female, mean age 65, range 47-78). Participants were included in this study when a pre- and post-treatment recording was present. Subjects in pallia- tive setting were excluded. All participants did receive speech language therapy sessions, until satisfactory speech was reached. Post-treatment recordings were made after on average 6 months after total laryngectomy (range 3-35). This study has been approved by the Institutes Research Board (registration num- ber IRBd18005).