Empirical Evaluation of the Role of Vocal Fold Collision on Relative Fundamental Frequency in Voicing Offset.

OBJECTIVES: Relative fundamental frequency (RFF) is an acoustic measure of changes in fundamental frequency during voicing transitions. The physiological mechanisms underlying RFF remain unclear. Recent modeling suggests that changes in RFF during voicing offset are due to decreases in overall system stiffness as a direct result of the cessation of vocal fold collision. To evaluate this finding empirically, here we examined whether variable timing between the end of vocal fold collision and the final voicing cycle used to calculate RFF explained the variability in RFF across individual voicing offset utterances. METHODS: RFF during voicing offset was calculated from /ifi/ utterances produced by 35 participants under endoscopy, with and without vocal effort. RFF was calculated via two methods, in which utterances were aligned by (1) the end of vocal fold collision, or (2) the end of voicing. Analyses of variance were used to determine the effects of vocal effort and RFF method on the mean and standard deviation of RFF. RESULTS: Aligning by vocal fold collision resulted in statistically significantly lower standard deviations. RFF means were statistically higher using the collision method; however, the degree of vocal effort was statistically significant regardless of the method. CONCLUSIONS: These results provide empirical evidence to support that decreases in RFF during voicing offset are a result of decreases in system stiffness due to termination of vocal fold collision.

Resynthesis of Transmasculine Voices to Assess Gender Perception as a Function of Testosterone Therapy.

PURPOSE: The goal of this study was to use speech resynthesis to investigate the effects of changes to individual acoustic features on speech-based gender perception of transmasculine voice samples following the onset of hormone replacement therapy (HRT) with exogenous testosterone. We hypothesized that mean fundamental frequency (f o) would have the largest effect on gender perception of any single acoustic feature. METHOD: Mean f o, f o contour, and formant frequencies were calculated for three pairs of transmasculine speech samples before and after HRT onset. Sixteen speech samples with unique combinations of these acoustic features from each pair of speech samples were resynthesized. Twenty young adult listeners evaluated each synthesized speech sample for gender perception and synthetic quality. Two analyses of variance were used to investigate the effects of acoustic features on gender perception and synthetic quality. RESULTS: Of the three acoustic features, mean f o was the only single feature that had a statistically significant effect on gender perception. Differences between the speech samples before and after HRT onset that were not captured by changes in f o and formant frequencies also had a statistically significant effect on gender perception. CONCLUSION: In these transmasculine voice samples, mean f o was the most important acoustic feature for voice masculinization as a result of HRT; future investigations in a larger number of transmasculine speakers and on the effects of behavioral therapy-based changes in concert with HRT is warranted.

Automated Relative Fundamental Frequency Algorithms for Use With Neck-Surface Accelerometer Signals.

OBJECTIVE: Relative fundamental frequency (RFF) has been suggested as a potential acoustic measure of vocal effort. However, current clinical standards for RFF measures require time-consuming manual markings. Previous semi-automated algorithms have been developed to calculate RFF from microphone signals. The current study aimed to develop fully automated algorithms to calculate RFF from neck-surface accelerometer signals for ecological momentary assessment and ambulatory monitoring of voice. METHODS: Training a set of 2646 /vowel-fricative-vowel/ utterances from 317 unique speakers, with and without voice disorders, was used to develop automated algorithms to calculate RFF values from neck-surface accelerometer signals. The algorithms first rejected utterances with poor vowel-to-noise ratios, then identified fricative locations, then used signal features to determine voicing boundary cycles, and finally calculated corresponding RFF values. These automated RFF values were compared to the clinical gold-standard of manual RFF calculated from simultaneously collected microphone signals in a novel test set of 639 utterances from 77 unique speakers. RESULTS: Automated accelerometer-based RFF values resulted in an average mean bias error (MBE) across all cycles of 0.027 ST, with an MBE of 0.152 ST and -0.252 ST in the offset and onset cycles closest to the fricative, respectively. CONCLUSION: All MBE values were smaller than the expected changes in RFF values following successful voice therapy, suggesting that the current algorithms could be used for ecological momentary assessment and ambulatory monitoring via neck-surface accelerometer signals.

Vocal fold kinematics and relative fundamental frequency as a function of obstruent type and speaker age.

The acoustic measure, relative fundamental frequency (RFF), has been proposed as an objective metric for assessing vocal hyperfunction; however, its underlying physiological mechanisms have not yet been fully characterized. This study aimed to characterize the relationship between RFF and vocal fold kinematics. Simultaneous acoustic and high-speed videoendoscopic (HSV) recordings were collected as younger and older speakers repeated the utterances /ifi/ and /iti/. RFF values at voicing offsets and onsets surrounding the obstruents were estimated from acoustic recordings, whereas glottal angles, durations of voicing offset and onset, and a kinematic estimate of laryngeal stiffness (KS) were obtained from HSV images. No differences were found between younger and older speakers for any measure. RFF did not differ between the two obstruents at voicing offset; however, fricatives necessitated larger glottal angles and longer durations to devoice. RFF values were lower and glottal angles were greater for stops relative to fricatives at voicing onset. KS values were greater in stops relative to fricatives. The less adducted vocal folds with greater KS and lower RFF at voicing onset for stops relative to fricatives in this study were in accordance with prior speculations that decreased vocal fold contact area and increased laryngeal stiffness may decrease RFF.

The Relationship Between Voice Onset Time and Increase in Vocal Effort and Fundamental Frequency.

Purpose Prior work suggests that voice onset time (VOT) may be impacted by laryngeal tension: VOT means decrease when individuals with typical voices increase their fundamental frequency (f o) and VOT variability is increased in individuals with vocal hyperfunction, a voice disorder characterized by increased laryngeal tension. This study further explored the relationship between VOT and laryngeal tension during increased f o, vocal effort, and vocal strain. Method Sixteen typical speakers of American English were instructed to produce VOT utterances under four conditions: baseline, high pitch, effort, and strain. Repeated-measures analysis of variance models were used to analyze the effects of condition on VOT means and standard deviations (SDs); pairwise comparisons were used to determine significant differences between conditions. Results Voicing, condition, and their interaction significantly affected VOT means. Voiceless VOT means significantly decreased for high pitch (p < .001) relative to baseline; however, no changes in voiceless VOT means were found for effort or strain relative to baseline. Although condition had a significant effect on VOT SDs, there were no significant differences between effort, strain, and high pitch conditions relative to baseline. Conclusions Speakers with typical voices likely engage different musculature to increase pitch than to increase vocal effort and strain. The increased VOT variability present with vocal hyperfunction is not seen in individuals with typical voices using increased effort and strain, supporting the assertion that this feature of vocal hyperfunction may be related to disordered vocal motor control rather than resulting from effortful voice production.

Testosterone therapy masculinizes speech and gender presentation in transgender men.

Voice is one of the most noticeably dimorphic traits in humans and plays a central role in gender presentation. Transgender males seeking to align internal identity and external gender expression frequently undergo testosterone (T) therapy to masculinize their voices and other traits. We aimed to determine the importance of changes in vocal masculinity for transgender men and to determine the effectiveness of T therapy at masculinizing three speech parameters: fundamental frequency (i.e., pitch) mean and variation (fo and fo-SD) and estimated vocal tract length (VTL) derived from formant frequencies. Thirty transgender men aged 20 to 40 rated their satisfaction with traits prior to and after T therapy and contributed speech samples and salivary T. Similar-aged cisgender men and women contributed speech samples for comparison. We show that transmen viewed voice change as critical to transition success compared to other masculine traits. However, T therapy may not be sufficient to fully masculinize speech: while fo and fo-SD were largely indistinguishable from cismen, VTL was intermediate between cismen and ciswomen. fo was correlated with salivary T, and VTL associated with T therapy duration. This argues for additional approaches, such as behavior therapy and/or longer duration of hormone therapy, to improve speech transition.

Acoustic Identification of the Voicing Boundary during Intervocalic Offsets and Onsets based on Vocal Fold Vibratory Measures.

Methods for automating relative fundamental frequency (RFF)-an acoustic estimate of laryngeal tension-rely on manual identification of voiced/unvoiced boundaries from acoustic signals. This study determined the effect of incorporating features derived from vocal fold vibratory transitions for acoustic boundary detection. Simultaneous microphone and flexible nasendoscope recordings were collected from adults with typical voices (N=69) and with voices characterized by excessive laryngeal tension (N=53) producing voiced-unvoiced-voiced utterances. Acoustic features that coincided with vocal fold vibratory transitions were identified and incorporated into an automated RFF algorithm ("aRFF-APH"). Voiced/unvoiced boundary detection accuracy was compared between the aRFF-APH algorithm, a recently published version of the automated RFF algorithm ("aRFF-AP"), and gold-standard, manual RFF estimation. Chi-square tests were performed to characterize differences in boundary cycle identification accuracy among the three RFF estimation methods. Voiced/unvoiced boundary detection accuracy significantly differed by RFF estimation method for voicing offsets and onsets. Of 7721 productions, 76.0% of boundaries were accurately identified via the aRFF-APH algorithm, compared to 70.3% with the aRFF-AP algorithm and 20.4% with manual estimation. Incorporating acoustic features that corresponded with voiced/unvoiced boundaries led to improvements in boundary detection accuracy that surpassed the gold-standard method for calculating RFF.

Cursor Click Modality in an Accelerometer-Based Computer Access Device.

The purpose of this study is to investigate the effects of different cursor click modalities in an alternative computer access device using accelerometry from head tilt to control cursor movement. Eighteen healthy adults performed a target acquisition task using the device with five different cursor click modalities, while maintaining cursor movement control via accelerometry. Three dwell-based click modalities with dwell times of 0.5 s, 1.0 s, and 1.5 s were tested. Two surface electromyography-based click modalities - with the sensor placed next to the eye for a blink and above the eyebrow for a brow raise - were tested. Performance was evaluated using metrics of target selection accuracy, path efficiency, target selection time, and user effort. Surface electromyography-based click modalities were as fast as the shortest dwell time and as accurate as the longest dwell time, and also minimized user effort. Three of the four performance metrics were not affected by sensor location. Future studies will investigate if these results are similar in individuals with neuromuscular disorders.

Formant-Estimated Vocal Tract Length and Extrinsic Laryngeal Muscle Activation During Modulation of Vocal Effort in Healthy Speakers.

Purpose The goal of this study was to explore the relationships among vocal effort, extrinsic laryngeal muscle activity, and vocal tract length (VTL) within healthy speakers. We hypothesized that increased vocal effort would result in increased suprahyoid muscle activation and decreased VTL, as previously observed in individuals with vocal hyperfunction. Method Twenty-eight healthy speakers of American English produced vowel-consonant-vowel utterances under varying levels of vocal effort. VTL was estimated from the vowel formants. Three surface electromyography sensors measured the activation of the suprahyoid and infrahyoid muscle groups. A general linear model was used to investigate the effects of vocal effort level and surface electromyography on VTL. Two additional general linear models were used to investigate the effects of vocal effort on suprahyoid and infrahyoid muscle activities. Results Neither vocal effort nor extrinsic muscle activity showed significant effects on VTL; however, the degree of extrinsic muscle activity of both suprahyoid and infrahyoid muscle groups increased with increases in vocal effort. Conclusion Increasing vocal effort resulted in increased activation of both suprahyoid and infrahyoid musculature in healthy adults, with no change to VTL.