Gender-Affirming Voice Training for Trans Women: Acoustic Outcomes and Their Associations With Listener Perceptions Related to Gender.

OBJECTIVES: To investigate acoustic outcomes of gender-affirming voice training for trans women wanting to develop a female sounding voice and to describe what happens acoustically when male sounding voices become more female sounding. STUDY DESIGN: Prospective treatment study with repeated measures. METHODS: N = 74 trans women completed a voice training program of 8-12 sessions and had their voices audio recorded twice before and twice after training. Reference data were obtained from N = 40 cisgender speakers. Fundamental frequency (fo), formant frequencies (F1-F4), sound pressure level (Leq), and level difference between first and second harmonic (L1-L2) were extracted from a reading passage and spontaneous speech. N = 79 naive listeners provided gender-related ratings of participants' audio recordings. A linear mixed-effects model was used to estimate average training effects. Individual level analyses determined how changes in acoustic data were related to listeners' ratings. RESULTS: Group data showed substantial training effects on fo (average, minimum, and maximum) and formant frequencies. Individual data demonstrated that many participants also increased Leq and some increased L1-L2. Measures that most strongly predicted listener ratings of a female sounding voice were: fo, average formant frequency, and Leq. CONCLUSIONS: This is the largest prospective study reporting on acoustic outcomes of gender-affirming voice training for trans women. We confirm findings from previous smaller scale studies by demonstrating that listener perceptions of male and female sounding voices are related to acoustic voice features, and that voice training for trans women wanting to sound female is associated with desirable acoustic changes, indicating training effectiveness. Although acoustic measures can be a valuable indicator of training effectiveness, particularly from the perspective of clinicians and researchers, we contend that a combination of outcome measures, including client perspectives, are needed to provide comprehensive evaluation of gender-affirming voice training that is relevant for all stakeholders.

Fundamental frequency disturbances in female and male singers' pitch glides through long tube with varied resistances.

Source-filter interaction can disturb vocal fold vibration frequency. Resonance frequency/bandwidth ratios (Q-values) may affect such interaction. Occurrences of fundamental frequency (fo) disturbances were measured in ascending pitch glides produced by four female and five male singers phonating into a 70 cm long tube. Pitch glides were produced with varied resonance Q-values of the vocal tract + tube compound (VT + tube): (i) tube end open, (ii) tube end open with nasalization, and (iii) with a piece of cotton wool in the tube end (conditions Op, Ns, and Ct, respectively). Disturbances of fo were identified by calculating the derivative of the low-pass filtered fo curve. Resonance frequencies of the compound VT+tube system were determined from ringings and glottal aspiration noise observed in narrowband spectrograms. Disturbances of fo tended to occur when a partial was close to a resonance of the compound VT+tube system. The number of such disturbances was significantly lower when the resonance Q-values were reduced (conditions Ns and Ct), particularly for the males. In some participants, resonance Q-values seemed less influential, suggesting little effect of source-filter interaction. The study sheds light on factors affecting source-filter interaction and fo control and is, therefore, relevant to voice pedagogy and theory of voice production.

Long-Term Average Spectrum Characteristics of Portuguese Fado-Canção from Coimbra.

Descriptions of acoustical characteristics of Fado, a Portuguese urban style sung in Lisbon and Oporto, are scarce, particularly concerning Fado-Canção, a related style sung in Coimbra. The present study aims at describing long-term average spectrum (LTAS) parameters of 16 professional singers while singing and reading the lyrics of a typical Fado-Canção. LTAS parameters were investigated in terms of: (1) equivalent sound level (Leq); (2) spectral differences between 3 frequency bands 0-2, 2-5, and 5-8 kHz; and (3) quantification of spectral prominence between 2 and 4 kHz, calculated as the level difference between the peak in this frequency region and a reference trendline between 1 and 5 kHz, henceforth Formant Cluster Prominence (FCP). Given that Fado-Canção, besides Fado and traditional styles, originated also from classical singing, and that previous studies on Fado suggest the absence of a singer's formant cluster, the averaged LTAS for all Fado-Canção singers was further compared to the LTAS of two world-touring opera baritones singing an operatic aria and a lied. Results show that Fado-Canção is commonly sung with a Leq of 86.4 dB and a FCP of about 10 dB, values significantly higher when compared to reading. The FCP in Fado-Canção, although smaller than for the two classical opera singers' examples (14.8 and 20 dB, respectively), suggests that the style preserved some of its original lyrical influence. However, because younger singers present higher energy in the 5-8 kHz region relative to the remaining frequency bands as compared to older singers, it seems that Fado-Canção may be drifting towards non-classical vocal practices. FCP seems to be a promising straightforward method to quantify the degree of formant clustering around the region of the singer's formant in LTAS, allowing comparisons between different singers and singing styles.

Female Voice-Related Sexual Attractiveness to Males: Does it Vary With Different Degrees of Conception Likelihood?

Previous investigations have found that female voice-related attractiveness to males increases when both conception likelihood (CL) and voice fundamental frequency (fo) are elevated. To test this hypothesis, we conducted a perceptual experiment where 78 heterosexual males rated sexual attractiveness of 9 female voice samples, recorded at menstrual, follicular and luteal phases of the menstrual cycle under two double-blinded randomly allocated conditions: a natural menstrual cycle (placebo condition) and when using an oral contraceptive pill (OCP condition). The voice samples yielded a total of 54 stimuli that were visually sorted and rated using Visor software. Concentrations of estrogens, progesterone and testosterone were analyzed, and measurements of speaking fundamental frequency (sfo) and its standard deviation (sfoSD), fo derivative (dfo) and fo slope were made. A multilevel ordinal logistic regression model nested in listeners and in females, and adjusted by phase and condition, was carried out to assess the association between ratings and: (1) phases and conditions; (2) sex steroid hormonal concentrations; and (3) voice parameters. A high probability of obtaining high ratings of voice sexual attractiveness was found for: (1) menstrual phase of placebo use and follicular phase of OCP use; (2) for low estradiol to progesterone ratio and testosterone concentrations; and (3) for low dfo. The latter showed a moderate statistical association with ratings of high attractiveness, as compared with the small association found for the remaining variables. It seems that the voice is a weak cue for female CL. Female sexual attraction to males may be a consequence of what females do in order to regulate their extended sexuality across the menstrual cycle rather than of estrus cues, the use of paralinguistic speech patterns being an example.

The Impact of Maxillary Advancement on Consonant Proficiency in Patients With Cleft Lip and Palate, Lay Listeners' Opinion, and Patients' Satisfaction With Speech.

OBJECTIVE: This study investigates the impact of maxillary advancement (Le Fort I osteotomy) on consonant proficiency in patients with cleft lip and palate (CLP) and explores how these patients and lay people perceive their speech 1 year post Le Fort I osteotomy. DESIGN: Retrospective group study before and after treatment. PARTICIPANTS: All patients with CLP who had undergone Le Fort I osteotomy for maxillary retrognathia between 2007 and 2010 at Karolinska University Hospital, Sweden (n = 21). Six patients were excluded due to additional malformations and missing data. Two experienced speech and language pathologists assessed consonant proficiency, and speech accuracy was determined by lay listeners from pre- and postoperative standardized audio recordings. The patients' satisfaction with speech postoperatively was collected from medical records. MAIN OUTCOME MEASURES: Percentage of oral consonants correct and acoustic analysis of /s/, lay listeners' opinion, and patients' satisfaction with speech. RESULTS: One year postoperation, 11 of the 15 patients had improved articulation, especially on the /s/-sound, without speech intervention. The mean percentage of oral consonants correct before treatment (82%) was significantly improved later (95%; P > .01). This assessment was supported by the patients' satisfaction with speech. However, lay listeners' opinion on accuracy was inconsistent. Length of maxillary advancement or change in occlusion did not correlate with change in articulation. CONCLUSION: Maxillary advancement performed to normalize occlusion and facial profile improved consonant proficiency in patients with CLP 1 year postoperation. Lay listeners' and patients' perceptions of speech need further exploration.

Reliability of Hypernasality Rating: Comparison of 3 Different Methods for Perceptual Assessment.

OBJECTIVE: To compare reliability in auditory-perceptual assessment of hypernasality for 3 different methods and to explore the influence of language background. DESIGN: Comparative methodological study. PARTICIPANTS AND MATERIALS: Audio recordings of 5-year-old Swedish-speaking children with repaired cleft lip and palate consisting of 73 stimuli of 9 nonnasal single-word strings in 3 different randomized orders. Four experienced speech-language pathologists (2 native speakers of Brazilian-Portuguese and 2 native speakers of Swedish) participated as listeners. After individual training, each listener performed the hypernasality rating task. Each order of stimuli was analyzed individually using the 2-step, VISOR and Borg centiMax scale methods. MAIN OUTCOME MEASURES: Comparison of intra- and inter-rater reliability, and consistency  for each method within language of the listener and between listener languages (Swedish and Brazilian-Portuguese). RESULTS: Good to excellent intra-rater reliability was found within each listener for all methods, 2-step: κ = 0.59-0.93; VISOR: intraclass correlation coefficient (ICC) = 0.80-0.99; Borg centiMax (cM) scale: ICC = 0.80-1.00. The highest inter-rater reliability was demonstrated for VISOR (ICC = 0.60-0.90) and Borg cM-scale (ICC = 0.40-0.80). High consistency within each method was found with the highest for the Borg cM scale (ICC = 0.89-0.91). There was a significant difference in the ratings between the Swedish and the Brazilian listeners for all methods. CONCLUSIONS: The category-ratio scale Borg cM was considered most reliable in the assessment of hypernasality. Language background of Brazilian-Portuguese listeners influenced the perceptual ratings of hypernasality in Swedish speech samples, despite their experience in perceptual assessment of cleft palate speech disorders.

Tutorial and Guidelines on Measurement of Sound Pressure Level in Voice and Speech.

Purpose: Sound pressure level (SPL) measurement of voice and speech is often considered a trivial matter, but the measured levels are often reported incorrectly or incompletely, making them difficult to compare among various studies. This article aims at explaining the fundamental principles behind these measurements and providing guidelines to improve their accuracy and reproducibility. Method: Basic information is put together from standards, technical, voice and speech literature, and practical experience of the authors and is explained for nontechnical readers. Results: Variation of SPL with distance, sound level meters and their accuracy, frequency and time weightings, and background noise topics are reviewed. Several calibration procedures for SPL measurements are described for stand-mounted and head-mounted microphones. Conclusions: SPL of voice and speech should be reported together with the mouth-to-microphone distance so that the levels can be related to vocal power. Sound level measurement settings (i.e., frequency weighting and time weighting/averaging) should always be specified. Classified sound level meters should be used to assure measurement accuracy. Head-mounted microphones placed at the proximity of the mouth improve signal-to-noise ratio and can be taken advantage of for voice SPL measurements when calibrated. Background noise levels should be reported besides the sound levels of voice and speech.

Vocal Behavior in Environmental Noise: Comparisons Between Work and Leisure Conditions in Women With Work-related Voice Disorders and Matched Controls.

OBJECTIVES: This study aimed to assess vocal behavior in women with voice-intensive occupations to investigate differences between patients and controls and between work and leisure conditions with environmental noise level as an experimental factor. METHODS: Patients with work-related voice disorders, 10 with phonasthenia and 10 with vocal nodules, were matched regarding age, profession, and workplace with 20 vocally healthy colleagues. The sound pressure level of environmental noise and the speakers' voice, fundamental frequency, and phonation ratio were registered from morning to night during 1 week with a voice accumulator. Voice data were assessed in low (≤55 dBA), moderate, and high (>70 dBA) environmental noise levels. RESULTS: The average environmental noise level was significantly higher during the work condition for patients with vocal nodules (73.9 dBA) and their controls (73.0 dBA) compared with patients with phonasthenia (68.3 dBA) and their controls (67.1 dBA). The average voice level and the fundamental frequency were also significantly higher during work for the patients with vocal nodules and their controls. During the leisure condition, there were no significant differences in average noise and voice level nor fundamental frequency between the groups. The patients with vocal nodules and their controls spent significantly more time and used their voices significantly more in high-environmental noise levels. CONCLUSIONS: High noise levels during work and demands from the occupation impact vocal behavior. Thus, assessment of voice ergonomics should be part of the work environmental management. To reduce environmental noise levels is important to improve voice ergonomic conditions in communication-intensive and vocally demanding workplaces.

Resonance Tube Phonation in Water-the Effect of Tube Diameter and Water Depth on Back Pressure and Bubble Characteristics at Different Airflows.

OBJECTIVES: Resonance tube phonation with tube end in water is a voice therapy method in which the patient phonates through a glass tube, keeping the free end of the tube submerged in water, creating bubbles. The purpose of this experimental study was to determine flow-pressure relationship, flow thresholds between bubble types, and bubble frequency as a function of flow and back volume. METHODS: A flow-driven vocal tract simulator was used for recording the back pressure produced by resonance tubes with inner diameters of 8 and 9 mm submerged at water depths of 0-7 cm. Visual inspection of bubble types through video recording was also performed. RESULTS: The static back pressure was largely determined by the water depth. The narrower tube provided a slightly higher back pressure for a given flow and depth. The amplitude of the pressure oscillations increased with flow and depth. Depending on flow, the bubbles were emitted from the tube in three distinct types with increasing flow: one by one, pairwise, and in a chaotic manner. The bubble frequency was slightly higher for the narrower tube. An increase in back volume led to a decrease in bubble frequency. CONCLUSIONS: This study provides data on the physical properties of resonance tube phonation with the tube end in water. This information will be useful in future research when looking into the possible effects of this type of voice training.

Real-Time Visual Feedback of Airflow in Voice Training: Aerodynamic Properties of Two Flow Ball Devices.

OBJECTIVES: Flow ball devices have been used as teaching tools to provide visual real-time feedback of airflow during singing. This study aims at exploring static back pressure and ball height as function of flow for two devices, marketed as flow ball and floating ball game. STUDY DESIGN: This is a comparative descriptive study. METHODS: A flow-driven vocal tract simulator was used to investigate the aerodynamic properties of these two devices, testing them for four different ball sizes. The flow range investigated was between 0 and 0.5 L/s. Audio, flow, pressure, and ball height were recorded. RESULTS: The flow pressure profiles for both tested devices were similar to those observed in previous studies on narrow tubes. For lifting the ball, both devices had a flow and a pressure threshold. The tested floating ball game required considerably higher back pressure for a given flow as compared with the flow ball. CONCLUSIONS: Both tested devices have similar effects on back pressure as straws of 3.7 and 3.0 mm in diameter for the flow ball and the floating ball game, respectively. One might argue that both devices could be used as tools for practicing semi-occluded vocal tract exercises, with the additional benefit of providing real-time visual feedback of airflow during phonation. The flow threshold, combined with the flow feedback, would increase awareness of flow, rather than of pressure, during exercises using a flow ball device.

Long-Term Follow-Up of Patients with Spasmodic Dysphonia and Improved Voice despite Discontinuation of Treatment.

OBJECTIVE: To evaluate voice function in patients with adductor spasmodic dysphonia (AdSD) who discontinued botulinum toxin (BTX) treatment because they felt that their voice had improved sufficiently. PATIENTS AND METHODS: Twenty-eight patients quit treatment in 2004, of whom 20 fulfilled the inclusion criteria for the study, with 3 subsequently excluded because of return of symptoms, leaving 17 patients (11 males, 6 females) included in this follow-up study. A questionnaire concerning current voice function and the Voice Handicap Index were completed. Audio-perceptual voice assessments were done by 3 listeners. The inter- and intrarater reliabilities were r > 0.80. RESULTS: All patients had a subjectively good stable voice, but with differences in their audio-perceptual voice assessment scores. Based on the pre-/posttreatment auditory scores on the overall degree of AdSD, patients were divided into 2 subgroups showing more and less improvement, with 10 and 7 patients, respectively. The subgroup with more improvement had shorter duration from the onset of symptoms until the start of BTX treatment, and included 7 males compared to only 4 males in the subgroup with less improvement. CONCLUSION: It seems plausible that the symptoms of spasmodic dysphonia may decrease over time. Early intervention and male gender seem to be important factors for long-term reduction of the voice symptoms of AdSD.

The Flow and Pressure Relationships in Different Tubes Commonly Used for Semi-occluded Vocal Tract Exercises.

This experimental study investigated the back pressure (P(back)) versus flow (U) relationship for 10 different tubes commonly used for semi-occluded vocal tract exercises, that is, eight straws of different lengths and diameters, a resonance tube, and a silicone tube similar to a Lax Vox tube. All tubes were assessed with the free end in air. The resonance tube and silicone tube were further assessed with the free end under water at the depths from 1 to 7 cm in steps of 1 cm. The results showed that relative changes in the diameter of straws affect P(back) considerably more compared with the same amount of relative change in length. Additionally, once tubes are submerged into water, P(back) needs to overcome the pressure generated by the water depth before flow can start. Under this condition, only a small increase in P(back) was observed as the flow was increased. Therefore, the wider tubes submerged into water produced an almost constant P(back) determined by the water depth, whereas the thinner straws in air produced relatively large changes to P(back) as flow was changed. These differences may be taken advantage of when customizing exercises for different users and diagnoses and optimizing the therapy outcome.

The softest sound levels of the human voice in normal subjects.

Accurate measurement of the softest sound levels of phonation presents technical and methodological challenges. This study aimed at (1) reliably obtaining normative data on sustained softest sound levels for the vowel [a:] at comfortable pitch; (2) comparing the results for different frequency and time weighting methods; and (3) refining the Union of European Phoniatricians' recommendation on allowed background noise levels for scientific and equipment manufacturers' purposes. Eighty healthy untrained participants (40 females, 40 males) were investigated in quiet rooms using a head-mounted microphone and a sound level meter at 30 cm distance. The one-second-equivalent sound levels were more stable and more representative for evaluating the softest sustained phonations than the fast-time-weighted levels. At 30 cm, these levels were in the range of 48-61 dB(C)/41-53 dB(A) for females and 49 - 64 dB(C)/35-53 dB(A) for males (5% to 95% quantile range). These ranges may serve as reference data in evaluating vocal normality. In order to reach a signal-to-noise ratio of at least 10 dB for more than 95% of the normal population, the background noise should be below 25 dB(A) and 38 dB(C), respectively, for the softest phonation measurements at 30 cm distance. For the A-weighting, this is 15 dB lower than the previously recommended value.

Resonance tube phonation in water: High-speed imaging, electroglottographic and oral pressure observations of vocal fold vibrations--a pilot study.

Phonation into glass tubes ('resonance tubes'), keeping the free end of the tube in water, has been a frequently used voice therapy method in Finland and more recently also in other countries. The purpose of this exploratory study was to investigate what effects tube phonation with and without water has on the larynx. Two participants were included in the study. The methods used were high-speed imaging, electroglottographic observations of vocal fold vibrations, and measurements of oral pressure during tube phonation. Results showed that the fluctuation in the back pressure during tube phonation in water altered the vocal fold vibrations. In the high-speed imaging, effects were found in the open quotient and amplitude variation of the glottal opening. The open quotient increased with increasing water depth (from 2 cm to 6 cm). A modulation effect by the water bubbles on the vocal fold vibrations was seen both in the high-speed glottal area tracings and in the electroglottography signal. A second experiment revealed that the increased average oral pressure was largely determined by the water depth. The increased open quotient can possibly be explained by an increased abduction of the vocal folds and/or a reduced transglottal pressure. The back pressure of the bubbles also modulates glottal vibrations with a possible 'massage' effect on the vocal folds. This effect and the well-defined average pressure increase due to the known water depth are different from those of other methods using a semi-occluded vocal tract.

Subglottal pressure oscillations accompanying phonation.

Acoustic and aerodynamic properties of the voice source and vocal tract have been extensively analyzed during the last half century. Corresponding investigations of the subglottal system are rare but can be assumed to be relevant to voice production. In the present exploratory study, subglottal pressure was recorded in a male adult subject by means of tracheal puncture. Also recorded were the oral airflow and audio signals. Effects of vowel, phonation type, and vocal register shifts on the subglottal pressure waveform were examined. The moment of maximum flow declination rate was synchronous with the main positive peak of the subglottal pressure waveform. The three lowest subglottal resonance frequencies, determined by inverse filtering and long-term average spectra of the subglottal pressure during speech, were found to be about 500, 1220, and 2000Hz, irrespective of supraglottal variations and phonation type. However, the subglottal pressure waveform was affected by the supraglottal formants, whereas the radiated vowel spectra did not show clear influence by the subglottal resonances. The fundamental frequency immediately preceding and immediately following a register break in pitch glides did not show systematic relationships with formants or with the lowest subglottal resonance.

Comparing closed quotient in children singers' voices as measured by high-speed-imaging, electroglottography, and inverse filtering.

The closed quotient, i.e., the ratio between the closed phase and the period, is commonly studied in voice research. However, the term may refer to measures derived from different methods, such as inverse filtering, electroglottography or high-speed digital imaging (HSDI). This investigation compares closed quotient data measured by these three methods in two boy singers. Each singer produced sustained tones on two different pitches and a glissando. Audio, electroglottographic signal (EGG), and HSDI were recorded simultaneously. The audio signal was inverse filtered by means of the decap program; the closed phase was defined as the flat minimum portion of the flow glottogram. Glottal area was automatically measured in the high speed images by the built-in camera software, and the closed phase was defined as the flat minimum portion of the area-signal. The EGG-signal was analyzed in four different ways using the matlab open quotient interface. The closed quotient data taken from the EGG were found to be considerably higher than those obtained from inverse filtering. Also, substantial differences were found between the closed quotient derived from HSDI and those derived from inverse filtering. The findings illustrate the importance of distinguishing between these quotients.

Guidelines for selecting microphones for human voice production research.

PURPOSE: This tutorial addresses fundamental characteristics of microphones (frequency response, frequency range, dynamic range, and directionality), which are important for accurate measurements of voice and speech. METHOD: Technical and voice literature was reviewed and analyzed. The following recommendations on desirable microphone characteristics were formulated: The frequency response of microphones should be flat (i.e., variation of less than 2 dB) within the frequency range between the lowest expected fundamental frequency of voice and the highest spectral component of interest. The equivalent noise level of the microphones is recommended to be at least 15 dB lower than the sound level of the softest phonations. The upper limit of the dynamic range of the microphone should be above the sound level of the loudest phonations. Directional microphones should be placed at the distance that corresponds to their maximally flat frequency response, to avoid the proximity effect; otherwise, they will be unsuitable for spectral and level measurements. Numerical values for these recommendations were derived for the microphone distances of 30 cm and 5 cm. CONCLUSIONS: The recommendations, while preliminary and in need of further numerical justification, should provide the basis for better accuracy and repeatability of studies on voice and speech production in the future.