Voice and Communication Change for Gender Nonconforming Individuals: Giving Voice to the Person Inside

ABSTRACT In the seventh version of their Standards of Care, WPATH recognizes that, as each person is unique, so is the person's gender identity. The goal of speech-language therapists/pathologists is to help transgender people develop voice and communication that reflects their unique sense of gender. When outer expression is congruent with an inner sense of self, transgender people may find increased comfort, confidence, and improved function in everyday life. Transgender voice and communication is a relatively new area of practice within speech-language pathology/therapy and this document is intended to support clinicians and researchers working in this field. It begins with a review of the evidence-based literature in transgender voice and communication. The paper then discusses these clinical topics: trans-specific voice-and-communication assessment, voice feminization protocols and voice feminizing surgeries. There is also a section on speech and voice masculinization—an area that has received little previous attention. As minimal standards have yet to be established in this field the paper concludes with recommendations for good clinical care.

The goal of the Standing Committee on Voice and Speech is to encourage the highest standards of research and practice in the field of transgender voice and communication. In creating this document we wish to make current evidence available to speech-language therapists internationally and foster excellence in this field around the world. However, research and practice constantly evolve, and by the time this document is published, our knowledge will already have advanced beyond it. We hope simply to provide guidance for current practice and stimulate more and better guidelines and research in the future.

SCOPE
This document is intended for speech-language therapists who have already taken transgender sensitivity/awareness training or have experience working with the transgender community and are seeking more advanced guidance on how to be clinically effective in working with transgender-specific aspects of voice and communication. Gender nonconforming people who feel that their speech and voice are not congruent with their gender do not have voice disorders, and the protocols used to address this issue are not designed to treat voice disorders. However, it is possible to have both a voice disorder and gender incongruence in voice and speech. A discussion of voice disorders in transgender clients is found in the section "Vocal Health, Voice Habilitation and Voice Rehabilitation." It is our intention to provide a useful summary and interpretation of the research literature on voice, speech, and communication for transgender people. To do this we consulted the following databases: Medline, CINAHL, LLBA, ERIC, PubMed, and PsycINFO.
With the exceptions noted, this document was based on studies using transgender individuals rather than cisgender (gender-sex congruent) subjects. In some areas we could uncover almost no literature discussing transgender participants. These areas include nonverbal communication (e.g., nodding/shaking head, facial expressions); language and syntactic choices (e.g., more/less emotionally laden word choices, use of tag questions and qualifiers); and body movements, postures and gestures. While we do not dismiss the importance of these aspects of communication, the paucity of evidence using transgender subjects did not allow us to do more than mention these aspects in this document. Besides studies of cisgender participants, this companion document also contains instances of expert opinion about a given topic.
It should be noted that recent studies (both on cis-and transgender participants) highlight the context-and identity-driven variation of these communication parameters and shift the focus on the interaction of the (gendered and nongendered) identities in the communicative situation (cf. Grainger & Dunbar, 2009;Hogg, 1985;Podesva, 2007Podesva, , 2011.

LIMITATIONS
We are still in the early stages of understanding gender markers in voice and communication and how to affect their change in transgender people. The corpus of literature is small and many studies involve small numbers of participants and lack of control groups. Some of the literature is based on expert opinion rather than on well-designed studies. Oates reminds us that the evidence for effectiveness of voice therapy for transgender clients is weak, with 83% of the studies being "at the very lowest level on the evidence hierarchy. The remaining 17% of publications in this field provide only marginally stronger evidence" (Oates, 2012, p. 59). No studies compare specific interventions or the effectiveness of targeting particular parameters of voice and speech. Johnston (2007) notes that "the most compelling information will come from review articles, with or without quantitative metaanalysis" (p. 12). Until recently there have been few review articles that summarize the primary literature on transgender voice and communication.
Currently there are almost no studies that investigate effectiveness of voice-and-communication therapy protocols with transgender people. In the absence of evidence-based studies we have elected to discuss two kinds of protocols: those used with transgender people that include standard outcome measures and those modeled on evidence-based protocols from the voice therapy literature with cisgender people.
The literature we uncovered was found to be further biased in two ways: 1. Research and therapy have been conducted primarily with transsexual women, who are only one subgroup under the umbrella of those who identify as transgender. 2. The literature is chiefly concerned with voice. Oates (2012) references Pasricha et al. (2008), saying, "The voice was of primary concern [for transsexual women] and that it was the feature of their communication with which they were least satisfied" (p. 30).
This lack of strong evidence does not mean the practice is poor but simply that the field has not been fully and rigorously explored. It is important to remember that practice, especially in the medical field, always precedes evidence. In examining the literature on clinical practice, Oates (2012) noticed high levels of satisfaction among transgender clients; fairly strong consensus among speech-language therapists as to vocal parameters that are strong markers of speaker gender; and ways to approach voice therapy with this population.
A full review of the history of research in transgender voice and communication is beyond the scope of this document. Extensive reviews of the literature from 1978 to the present are found in Freidenberg (2002), Pickering and Baker (2012), and Oates and Dacakis (1983). Much of the early literature focused on single-case studies or small-group studies that report the outcomes of speech therapy and phonosurgery. Parameters of change targeted in speech therapy included pitch, resonance, inflections, loudness, articulation, and voice quality; speaking rate; and nonverbal communication (Bralley, Bull, Gore, & Edgerton, 1978;Coleman, 1983;G€ unzburger, 1989Oates & Dacakis, 1983;Spencer, 1988;Wolfe, Ratusnik, Smith, & Northrop, 1990). Oates and Dacakis (1997) expanded the discussion beyond research and therapy outcomes to discuss philosophy of treatment. They discuss the importance of working in a collaborative process with the client and outline the importance of designing a treatment plan that takes into account the client's changing gender identity, goals, and real life communicative situations.
As the understanding of issues in transgender voice and communication grew the literature became more refined. Comprehensive summaries of recent work are found in Dacakis et al. (2012); Gelfer and Tice (2013); Hancock and Garabedian (2013); and Pickering and Baker (2012).

ACROSS THE GENDER SPECTRUM
Many people identify as being under the gender nonconforming "umbrella" and there is great variation in the extent to which voice and communication changes are undertaken or desired by gender nonconforming individuals. Some gender nonconforming persons seek to develop two speech patterns (one more masculine and one more feminine) either because they identify as bigendered or because external pressures (family, employment, cultural community, friends) prevent living full time in a way that is consistent with their felt sense of self. Some people may have a sense of gender that is not at either pole of the cismale/cisfemale scale but is on the continuum of masculine and feminine. They would like a more flexible gender presentation to reflect this gender identity.
Further, most current transgender speech and voice protocols do not support bimodal speech as a treatment goal, based on the belief that to achieve maximal change it is necessary to have a consistent single speech pattern. Switching back and forth between two speech/voice patterns may be too difficult for some clients and inconsistent use decreases practice opportunities to acquire the new speech/voice habits. However, the human capacity to learn and speak more than two languages or dialects, develop a specific accent for an acting role, and develop a singing voice that is different from the speaking voice suggests it may be possible to develop bigender speech/voice. We encourage speechlanguage therapists to be open to this possibility and not to routinely exclude clients who have two speech/voice patterns as their treatment goal. We recommend that speech services be made available to the full spectrum of the gender nonconforming community.
As was discussed earlier, the majority of gender nonconforming clients seeking speech therapy services are transsexual women wishing to maximize their feminine voice and communication. Speech and voice feminization is widely recognized as a vital component of care for transgender women (Adler, Hirsch, & Mordaunt, 2012;Andrews, 1999;Kalra, 1977;Neumann & Welzel, 2004;Wollitzer, 1994). The weighting of studies in transgender communication is far from balanced, with almost all research being on transsexual women.
Much less attention has been paid to speech masculinization (Oates & Dacakis, 1997;van Borsel, Cuypere, Rubens, & Destaerke, 2000) and this lack of investigation may be obscuring concerns of transgender men. The study by van Borsel et al. (2000) of 16 female-to-male (FtM) transsexuals found that 88% of participants considered masculinization of communication to be as important or more important than sex reassignment surgery. While many transgender men find that they develop an acceptably masculine voice as a result of androgen therapy, some do encounter difficulties. We will discuss voiceand-communication change for transgender men in the section "Masculinizing Speech: Physiology and Identity."

Childhood and Adolescence
Speech and voice feminization in adolescents was first examined in 2012 in a single-subject study by Hancock and Helenius. They found that voice therapy based on principles established with adults was effective but caution that the implications of age on the therapy effectiveness have not been examined. As the need to transition is now being recognized earlier, more research into helping youth develop gender-appropriate communication is needed. We may wish to consider whether there is a role for the speech-language therapist in monitoring the communication needs of transgender adolescents.

Through the Gender Transition Process
Transsexual people's perception of both their gender identity and their voice may change as they move through transition. In her doctoral thesis, Byrne (2007) notes that some transsexual women feel less pressure to feminize their voice after gender reassignment surgery has been completed. She hypothesizes that once there is congruence between the physical body and gender identity, the need to sound completely female may be less important. It is important for speech-language therapists to remain sensitive to their clients' (possibly changing, evolving) wishes rather than follow their own assumptions about voice-and-communication goals.

Voice, Speech, and Sense of Self
Initial studies on the transgender voice looked chiefly at speaking fundamental frequency. More recently the scope of study has expanded to explore the intimate and complex relationship between voice/communication, sense of self, and function in the world. It is important to understand the client's perspective around speech and voice change (Andrews, 1999;Byrne, 2007;Dacakis, 2002;Dacakis, Davies, Oates, Douglas, & Johnston, 2013;Gelfer, 1999;McNeill et al., 2008;Oates & Dacakis, 1997;Owen & Hancock, 2010;S€ oderpalm et al., 2004). A study by McNeill et al. (2008) found that in male-to-female (MtF) transsexual clients, the happiness with voice was not directly related to speaking fundamental frequency. Another 2008 study looked at 12 MtF transsexuals and found that happiness with voice was not directly related to speaking fundamental frequency, but rather was significantly correlated with perception of femininity-both the participants' self-perceptions and the perceptions of listeners (McNeill et al., 2008). In 2010, in a study of 20 transgender women, Owen and Hancock confirmed this finding. In their study they found that the speaker's perception of her own femininity was strongly correlated to both the mean fundamental frequency of the voice and to her own feeling of femininity. Speakers who rated themselves as more feminine were also judged as more feminine by listeners. Another study by Hancock, Krissinger, and Owen (2011) extends the investigation, suggesting that exploring the client's own perception of the voice may be useful for evaluating the effectiveness of voice therapy. This is an important corroboration and verification that outcome measurements need to expand beyond the evaluation of acoustic parameters, such as speaking fundamental frequency.
In an attempt to gather information on the effect of voice in everyday life, Davies adapted the Vocal Handicap Index (Jacobson et al., 1997) to be trans-specific and created the Transsexual Self-Evaluation of Voice Questionnaire (Davies & Goldberg, 2006a). This questionnaire was further developed, and psychometric evaluation was undertaken by Dacakis, Davies, et al., resulting in the TVQ MtF (Transsexual Voice Questionnaire, Male to Female) (Dacakis et al., 2013). This is the first instrument with established psychometric properties that assesses a transsexual woman's experience of her own voice in real-life situations. The version for female to male transsexuals is still under development. The TVQ MtF is available as a free download from www.shelaghdavies.com or by contacting Georgia Dacakis at G.Dacakis@latrobe. edu.au. The questionnaire has been translated into Swedish, Portuguese, German, and Danish and these versions are also available on the website. Translations into a number of other languages are ongoing and, when finished, will be added to the website and available for downloading at no cost.

GENDERED VOICE AND COMMUNICATION
The following is a discussion of feminization in voice, speech, and communication. The process of masculinization is discussed in the section "Masculinizing Speech: Physiology and Identity."

Feminizing Voice, Speech and Other Aspects of Communication
Current evidence suggests that the largest feminization effects come from changing both the speaking fundamental frequency and the resonance of the voice. Other parameters that have been targeted for change are inflectional patterns and excursions, voice quality, speechsound articulation and duration, average speaking intensity, and speech rate. S€ oderpalm, Larsson, and Almquist (2004) established that MtF transsexuals were able to raise their speaking fundamental frequency without strain or vocal damage. This is reassuring but there is no consensus in the literature as to how high the speaking fundamental frequency (SFF) should be for a transgender woman's voice to be perceived as female. It is generally accepted that the SFF must move out of cismale speaking range (roughly 100-140 Hz) into either a gender neutral range (145-175 Hz) or cisfemale range (roughly 180-220 Hz). After that, opinion varies widely, and has been summarized by Hancock & Garabedian (2013). Studies have suggested a minimum SFF of 155 Hz (Wolfe et al., 1990), 165 Hz (Oates & Dacakis, 1983), 172 Hz (Owen & Hancock, 2010), and 180 Hz (Gorham-Rowan & Morris, 2006). Dacakis, Oates, and Douglas (2012) say:

Average Speaking Fundamental Frequency
Despite the fact that an increase of F0 (fundamental frequency) into the female F0 range may not routinely result in perceptions of the speaker as female, the evidence indicates that this is likely to increase listener perceptions of the voice as containing female attributes. (p. 167) It is perhaps not surprising that something as complex and close to the human heart as voice cannot be modified by simply altering one parameter. We also recall that contributions from the larynx (F0 and harmonics) are inextricably linked to contributions from the vocal tract (vowel formants, resonance effects) (Baken & Daniloff, 1991). Perceptually they are difficult to separate and indications are that they both must rise for the voice to be perceived as female.

Resonance
The word resonance has many meanings but in this context its meaning is highly specific. We refer to the filter function of the vocal tract on the sound wave produced by the larynx. Because of the interdependence of these organs, it is difficult to perceptually separate the contributions of source (larynx) and filter (vocal tract) in the gender perception of a voice (Carew et al., 2007). Research evidence shows that a voice may sound low even though the SFF is reasonably high and vice versa (Hillenbrand & Clark, 2009).
There is a 10% to 20% difference in length between the vocal tracts of cismen and ciswomen (Titze, 2000) and this difference in length produces acoustic differences in the sound wave that disseminates from the lips. Typically vocal tract resonances are higher for ciswomen and children and lower for cismen. These resonance effects are found in unvoiced as well as voiced sounds, as can be demonstrated by saying the voiceless fricative /s/ and protruding and retracting the lips.
Vowel formants-the first, second, and often third resonances of the vocal tract (F1, F2, and F3, respectively)-are frequently calculated to measure the effect of the vocal tract filter on the voice. While the first and second formants determine the vowel identity and quality, all three formants reflect the overall length of the vocal tract (Carew et al., 2007). Carew et al. (2007) developed a protocol of oral resonance therapy for transsexual women that used a slight retraction of the lips to shorten the resonance tract during speech. They found that all three vowel formants were higher, with F3 being most significantly affected. Participants and listeners both perceived the voices as more feminine after treatment. Interestingly, there was a concurrent spontaneous increase in fundamental frequency although this was not targeted in therapy. In an early study, Mount and Salmon (1988) found that a transsexual woman was able to raise her SFF to 210 Hz but was not perceived as female until she raised her vowel formants. Gelfer and Mikos (2005) and Pasricha et al. (2008) found that both SFF and vowel formants were important for gender perception. Finally, a study of ciswomen and cismen by Hillenbrand and Clark (2009) looked at the role of SFF and formant frequencies in distinguishing the voices of men and women. They found that shifting either SFF or formants alone was usually ineffective in changing the perceived gender of the speaker but when the speaker changed both together, the person was perceived to be the opposite gender 82% of the time. Although the researchers found that shifting the SFF was somewhat more important than shifting the formants, they concluded that both speaking fundamental frequency and vowel formants are needed to specify the gender of a speaker. This finding is further corroborated by the formant-and SFF-scaling research in Patterson, Smith, Dinther, and Walters (2007); ; and Smith, Patterson, Turner, Kawahara, and Irino (2005).

Inflections
The use of more upward gliding inflections and more frequent inflections (i.e., a more dynamic intonation contour) are generally considered to be feminine characteristics of voice production, while the use of a flatter speaking style has been considered to be more masculine (Pickering & Baker, 2012). Gelfer and Schofield (2000) found that using both a higher upper limit of SFF and a higher SFF facilitated the perception as female. However, two recent studies suggest that the use of a narrower inflectional range is perceived to be more feminine (Owen, 2009;Owen & Hancock, 2010). In 2014, Hancock, Colton and Douglas (2014) looked at intonation differences between transsexual women who passed as cisfemale and those who did not. The researchers found that the women who passed as cisfemale used more upward gliding intonations and a larger utterance semitone range than those who did not pass. The authors suggest that further, culturespecific study is needed of this important aspect of speech and voice production.

Voice Quality
In a study of ciswomen, van Borsel et al. (2009) found that breathiness was perceived to be a feminine characteristic of voice and suggested that this study needed to be replicated in transgender women. The adoption of a slightly breathy, or softer, quality is a frequent goal in many voice therapy protocols for transgender women (Davies & Goldberg, 2006a). The understanding is that this breathy quality may replicate the voice quality of ciswomen because of a small posterior gap between the vocal folds during vibration (Biever & Bless, 1989;Linville, 1992;S€ odersten & Lindestad, 1990). Palmer et al. (2012) undertook endoscopic and stroboscopic examination of nine transsexual women and noted that those who reported a "passing" female voice had a glottal closure pattern with a posterior glottal gap. In another study the mean SFF and semitone range, but not perturbations (noise to harmonics ratio, vocal turbulence index, jitter percent, shimmer percent), were strongly related to the perception of femininity (Owen & Hancock, 2010). Yet other studies found further, conflicting results on the production and perception of breathiness and creakiness in speech (Henton & Bladon, 1985;Henton, Bladon, Hyman, & Li, 1988;Kreiman, Gerratt, & Antonanzas-Barroso, 2007). More work is needed to confirm the effect of voice quality on the perception of voice in transgender women.

Articulation
Female speakers may speak with more precision than male speakers (Oates & Dacakis, 1983) and modification of articulatory patterns is frequently included in speech/voice therapy protocols for transgender women. In their study with transsexual women, Free and Dacakis found that speech was perceived to be more feminine when the consonants were rated as "more precise" rather than "less precise" Free & Dacakis, 2007).

Speaking Intensity
Women are generally considered to speak more quietly than men (Boonin, 2012). In one study of four transsexual women, the speaking loudness was the sole factor that accounted for gender perception (Holmberg et al., 2010). While this suggests that a moderate reduction in average speaking intensity, perhaps correlated with a slight increase in breathiness, may feminize the voice. Some transgender individuals who are selfconscious about speech may use insufficient vocal intensity as a means of minimizing attention to their voice and some transwomen may speak quietly to try to "soften" the voice (Dacakis, 2002). To be functional in all situations, particularly with background noise, transwomen need to learn how to produce a strong voice that still maintains a feminine vocal quality.

Speech Rate and Prosody
Studies on speech rate in cismen and ciswomen have been inconclusive . Van Borsel and De Maesschalck (2008) looked at the speech rate of 28 transsexual women and found no significant differences when compared with the rates of nontranssexual men and women. On the basis of these results it appears that speech rate is not a useful parameter to target in speech/voice change.
Oates remarks: "Communication management for [transgender/transsexual] clients may [also] require a focus on . . . vocabulary, pragmatic behaviours, conversational topics and style, and nonverbal communication" (Oates, 2012, p. 50). Although there is some psycholinguistic literature on cisgender speakers that supports this statement, we were not able to uncover literature using transgender speakers.

RESPONSIBILITIES OF THE VOICE-AND-COMMUNICATION PROFESSIONAL
Throughout the world, feminization/masculinization services are delivered primarily by speech-language therapists (speech therapy) and laryngologists (vocal surgery). While speech-language therapists do not need to be experts in every realm of transgender care to work with transgender clients, it is expected that those providing speech feminization/masculinization services have a basic understanding of the processes involved in gender transition, the potential impact of testosterone on FtM speech, and trans-specific psychosocial issues that shape clients' goals and treatment options. Understanding of basic sensitivity protocols such as use of preferred gender pronoun and name is essential but is outside the scope of this article.
To assist in greater understanding of transgender voice and communication issues, the client's permission should be sought to share anonymous assessment data with other speech professionals and for research purposes. Over the past 10 years the literature on transgender voice and communication has substantially increased but there is still a paucity of evidence in many areas, particularly in evidence-based clinical practice Hancock & Helenius, 2012). As with all research, it is important that transgender clients' involvement in research be fully voluntary-that is, it is not ethical to make service contingent on agreement to be studied.
As with other transgender care, we recommend that voice and communication services be offered in the context of a complete approach to transgender health that includes comprehensive primary care and a coordinated approach to psychological and social issues (Adler, Hirsch, & Mordaunt, 2012;Coleman et al., 2011;S€ oderpalm et al., 2004). Speech services must be individualized based on the individual's goals and identity, the risks and benefits of treatment options, and consideration of social and economic issues.
In the following two sections we provide an overview of existing good practice in the assessment and treatment of the transgender voice. While the two sections are largely geared toward the needs of transwomen, many of the considerations and concerns also apply to the FtM voice.

TRANS-SPECIFIC SPEECH/VOICE ASSESSMENT
Change in a transgender person's voice and communication starts with a thorough assessment. For a comprehensive discussion the reader is referred to Dacakis (2012a), "Assessment and Goal Setting Revisited" in Adler, Hirsch, and Mordaunt (2012).
Components of the initial assessment are 1. Establishing therapeutic rapport 2. Collecting the history a. Psychosocial history b. Voice use and vocal health c. Medical history 3. Clinical assessment of speech and voice a. Client's subjective assessment b. Acoustic/aerodynamic measurement c. Speech-language therapist's subjective analysis d. Developing goals and assessing potential for change e. Assisting the client in understanding therapeutic options The following recommendations relate to the general initial assessment that should be conducted as part of a therapeutic evaluation and treatment plan. Clients who wish to pursue pitch-elevating surgery must undergo additional evaluation prior to surgery; this is discussed in the section "Surgical Feminization of the Larynx and Voice."

Establishing Therapeutic Rapport
The relationship between client and clinician begins with the first interactions. In the initial session, the speech-language therapist is not only assessing the client, the client is also assessing the knowledge and supportiveness of the clinician. A relationship grounded in mutual respect, trust, and genuine care for the client's well-being facilitates open communication and encourages active engagement in therapy; conversely, it can be difficult to build therapeutic rapport if conflicts arise in initial sessions. Many transgender people have had negative experiences with ill-informed or unempathetic health professionals, and there may be wariness about entering unreservedly into a relationship around communication, which is by its nature highly personal. Because the assessment process sets the stage for all future interaction, it is extremely important to make the client feel respected and safe, and to create a feeling of positive anticipation for the therapy process. One report identified storage of information, privacy issues in setting appointment times, client name preference, use of the client's preferred pronouns, and therapist bias and judgments about gender transition issues that speech-language therapists need to consider in the intake process (King, Lindstedt, Jensen, & Law, 1999).

Collecting the History
Clients come to voice therapy with different backgrounds, goals and knowledge bases about voice and communication change, so the speech-language therapist should match the input to the needs of the client. Client history should include information about both transspecific concerns and general issues that may impact therapeutic options and potential outcomes. An initial open-ended question like, "What brings you to see me?" or "How can I help you?" may elicit a great deal of information. While some transgender individuals are very comfortable talking about their history, others are more private. In some cases it may be appropriate to revisit sensitive questions after therapeutic rapport has been well established.

Psychosocial History: Voice, Speech and Sense of Self
Our goal as speech-language therapists is to help the client develop voice and communication that is congruent with the person's own sense of gender. While many may wish to pass as cismale or cisfemale, others may feel more comfortable with a presentation somewhere between these polarities of gender. A sensitive exploration of what best fits the client's selfperception is necessary in determining goals in therapy. This question of "right fit" may be easy for the client to answer right away, or may be answered over time with experimentation, introspection, practice, and observation of role models. A self-evaluation questionnaire like the TVQ MtF (Dacakis et al., 2013: 21) can be useful in exploring and quantifying how well the voice fits the sense of self and how well it works in everyday life.
If during the assessment, the speech-language therapist uncovers evidence of a voice disorder, such as roughness, limited range, difficulty projecting, rapid vocal fatigue, or effortful voice production, the usual protocols of investigation should be followed, beginning with referral to an otolaryngologist.
In their review of 25 cases of voice-andcommunication change Hancock and Garabedian (2013) found that "28% presented with a voice disorder separate from gender presentation concerns." If a client has a voice disorder it is important to determine if it existed before any speech/ voice change was undertaken or if it resulted from attempts to change the voice. Previous attempts to feminize/masculinize speech should be investigated, including techniques used, duration of self-directed or professionally directed therapy, and the client's subjective feelings about the outcome (Adler & Antoni, 2012;Andrews, 1999;Dacakis, 2002;Gelfer, 1999;Perez, 2004). Modifying habitual pitch, adopting unfamiliar articulatory postures or producing the voice in nonhabitual ways can be fatiguing and potentially damaging for some people. Clients who developed dysphonia and vocal fatigue from previous attempts at voice feminization or masculinization may find this disappears when they adopt a more effective protocol. Others may require a traditional voice rehabilitation program before they undertake further voice change (Davies & Goldberg, 2006a).

Voice Use
A full history of the client's voice use should be collected, including professional and recreational use of the voice (e.g., involvement in singing/acting). This history is useful for exploring any areas of concern and determining whether previous training could be tapped during therapy. Teachers, singers, and clients with heavy voice demands or difficult speaking situations will need maximally efficient voice-production habits to sustain their gender-modified voice and avoid phonotrauma. As in traditional voice therapy, creating a list of voice-use situations is useful in generalizing the new voice into everyday life and creating a hierarchy of practice (Dacakis, 2012a;Davies & Goldberg, 2006a).

Medical History
In some clinics a laryngological examination is performed as a matter of course by an otolaryngologist. In others it is performed with a referral when there are symptoms of dysphonia or other signs of abnormal laryngeal function.
As in the assessment of any voice client, a full history of medical information relevant to the voice and speech should be recorded. This includes information about respiratory disease, upper respiratory tract infections, laryngopharyngeal reflux disease, thyroid function, head/ neck surgeries or trauma, hearing loss, and other medical information relevant to speech and voice change. This includes problems stemming from self-directed attempts to modify voice, high vocal load, or any other conditions that could impact speech (de Bruin, Coerts, & Greven, 2000;Gelfer, 1999;S€ oderpalm et al., 2004). To assist in the coordination of care, other health professionals involved in the client's general and trans-specific care should be noted (de Bruin et al., 2000;S€ oderpalm et al., 2004). All current medications (including feminizing or masculinizing hormones) should be recorded. As previously stated, clients who present with difficulty swallowing or symptoms that may suggest a voice disorder should be referred for laryngological examination (Dacakis, 2012a).
Lifestyle habits that may negatively impact speech and voice, such as smoking (tobacco, cocaine, marijuana) and more than moderate use of alcohol, should be explored. Because the stigma associated with substance abuse can make it difficult to get accurate information, a gentler question could be, "Have you ever used . . . ?" While female hormones have no effect on the voice of a transgender woman, some women report emotional changes that may affect the way they communicate so it can be useful to know if a transgender woman has been taking hormones and for how long. It is likely that androgens affect the voice of transgender men by thickening the vocal folds thus deepening the voice. Depending on the method of testosterone intake, this process can take up to 2 years to complete (cf. Papp, 2011a), so it is important to know the length of time a transgender man has been undergoing androgen therapy, the dose he is taking, and possibly also his recent serum testosterone levels.
There is some discussion around whether a speech-language therapist should ask if the client has undergone genital reassignment/ reconstruction surgery (GRS). While this procedure has no impact on speech/voice production it may affect a transperson's feelings about voice change, so it is important information for the speech-language therapist to have. Byrne (2007) notes that after gender reassignment surgery, some transwomen feel a comfort that their physical body is now congruent with their gender identity and developing a completely female voice is less important. However, some speech-language therapists may feel this is an intrusive, personal question and not necessary since it does not affect voice production. Further, not all transgender clients wish to undergo GRS, whether or not they live full time as women. Sensitivity and tact are important when approaching this question.
Trans-specific history should also include information about other feminization/masculinization treatments that relate directly to speech and voice (e.g., facial feminization surgery in MtFs, chest binding in FtMs) and any noted impact on speech following these treatments (Andrews, 1999;Gelfer, 1999;King et al., 1999;S€ oderpalm et al., 2004). It is not necessary to inquire specifically about trans-specific treatments that are unlikely to directly impact communication.

Clinical Assessment of Speech and Voice Parameters Associated With Gender
A complete clinical analysis should include a variety of investigations with input from both the client and the clinician. The objective is to gain a thorough understanding of how well the speech and voice currently function in everyday life, how well they reflect the client's own sense of gender, and what potential there may be for speech and voice feminization (Andrews & Schmidt, 1997;Byrne, 2003;Coleman, 1983;Dacakis, 2002;Dacakis et al., 2013;de Bruin et al., 2000;Gelfer, 1999;Gelfer & Schofield, 2000;Gold, 1999;Oates & Dacakis, 1997;Wollitzer, 1994).
The following is a summary of a comprehensive trans-specific speech and voice assessment. For a full discussion the reader is referred to Dacakis (2012a).

Client's Subjective Assessment
Clients' feelings about their current voice and communication and their goals for change are related to self-perception, feelings about the perceptions of others, and function in day-today life. These can be assessed both formally and informally.
Formal Evaluation. The Transsexual Voice Questionnaire (TVQ MtF ) has recently been developed by Dacakis et al. (2013). The FtM version is still under development. This questionnaire explores a transsexual woman's feelings about her voice and how well her voice functions in the real world. It can be used as an intake instrument and as an outcome measurement. The Transsexual Voice Questionnaire may be acquired free of charge at www.shelagh davies.com and from Georgia Dacakis. For translation, authorization from the authors is required and WHO protocols for translation must be followed. Please contact Georgia Dacakis at La Trobe University, Melbourne, Australia, for further information (g.dacakis@latrobe.edu.au).
Informal Evaluation. Asking clients for three things they would like to change about speech and voice or three situations in which they would like to sound more masculine or feminine gives a brief insight into the clients' goals and can be reviewed at the end of voice therapy as a rough outcome measurement (Davies & Goldberg, 2006a). Dacakis (2012a) suggests other useful probes such as asking if the client wants to change communication in all speaking situations or if the client has a particular vocal image in mind. Questions about previous attempts to change speech and voice can yield information about the client's familiarity with the process and his or her stimulability for change. It may also alert the client to any potentially harmful vocal hyperfunction resulting from undirected voice change (Adler, Hirsch, & Mordaunt, 2012;Davies & Goldberg, 2006a).

Acoustic Analyses
An assessment of the client's voice and speech gives a baseline against which to measure change and provides information about which changes would be most useful (Andrews & Schmidt, 1997;Kaye et al., 1993;King et al., 1999). The parameters of speech and voice relevant to gender perception of voice have been discussed earlier and will simply be mentioned here. They include the following: speaking fundamental frequency (SFF); speaking fundamental frequency range and variability; total physiological frequency range (or maximum phonation range, cf.); formants of vowels, in particular, the corner vowels /a/, /i/, and /u/; and average speaking intensity.
These may all be measured using a number of acoustics analysis software programs currently available. Some, such as Praat (http:// www.fon.hum.uva.nl/praat/) or SFS/WASP (http://www.speechandhearing.net/laboratory/ wasp.htm) may be downloaded free of charge. Some commercially available programs can be obtained through Kay-PENTAX, Tokyo, Japan, and LingWAVES, a program manufactured by WEVOSYS, Forchheim, Germany.
Speaking intensity may be measured using a sound level meter. Care should be taken to measure in dBC and to conform to the standard distance from lip to sound level meter, as this varies throughout the world (Gramming & Sundberg, 1988). The total physiological and intensity ranges may also be measured and displayed in a vocal range profile, or phonetogram. This graph displays vocal intensity, usually on the vertical axis, against fundamental frequency on the horizontal axis, providing a clear picture of the maximal extent of these vocal parameters (Sulter, Wit, Schutte, & Miller, 1994).
It is useful to record the fundamental frequency in sustained vowels and the speaking fundamental frequency in various tasks such as sustained vowels, reading, picture description, and spontaneous speech. A variety of samples give a broader base and differences among them may reflect a voice in transition. There is generally less mental energy required to sustain a vowel, read, or describe a picture than to generate language in spontaneous conversation. If the SFF is higher in reading than in conversation this may suggest that the client needs high levels of concentration on voice production and cannot maintain this when required to think of the words as well (Dacakis, 2012a;Davies & Goldberg, 2006a). However, the difficulty among these tasks may also reflect other factors such as educational level or comfort with oral reading.
Audio or Video Recordings. Following standard practice in a speech/voice evaluation, audio or video recordings of the assessment should be made. These are useful for analyzing current communication patterns and setting goals for therapy. They also serve as a baseline against which to measure change and (with the consent of the client) as a resource to train student speech-language therapists. Some speechlanguage therapists videotape the assessment session and review the tape with the client, looking at speech/voice parameters and also nonverbal communication features such as gestures, movement, and facial expressions. Other speech-language therapists prefer not to use video footage as clients report finding it intrusive and intimidating (Davies & Goldberg, 2006a). Both audio and video recordings allow comparisons to be made during trial therapy, as is discussed later.
Speech-Language Therapist's Perceptual Analysis. As in a traditional speech and voice assessment, an experienced speech-language therapist can supplement the objective analysis by making perceptual assessments of voice quality, intonation patterns, speaking style (legato vs. staccato), and speech sound articulation.

Trial Therapy to Assess Potential for Change
Once the assessment is complete the speechlanguage therapist can share the findings with the client, explain any technical details, and discuss possible parameters of change. This establishes a collaborative process, with client and clinician working together to establish goals.
At this point, trial therapy is useful to get a sense of how the voice and speech could change. For example, if the client wishes to feminize her voice, some words or short sentences can first be recorded at the habitual pitch. Then the speech-language therapist can assist the client to produce them efficiently at a higher pitch or at several higher pitches. The trials can be recorded and compared with the original one at the habitual pitch. This practice gives useful information about what feels comfortable and acceptable to the client and what may be possible. It also affords an opportunity to discuss the nature of the process of change. Change to any motor habit will feel unfamiliar and possibly uncomfortable simply because it is nonhabitual. However, this is a sign that things are moving and this movement is necessary to develop new habits. While the final decision about the nature of voice and communication change rests with the client, the clinician can assist this process by suggesting the client remain open to experimenting with what is unfamiliar (Adler, Hirsch, & Mordaunt, 2012;Davies & Goldberg, 2006a).

Assisting the Client in Understanding Therapeutic Options
Two options open to the transgender client wishing voice feminization are voice therapy or phonosurgery. They may be available singly or together and the availability of both varies widely throughout the world. The speech-language therapist should be familiar with options in his or her location, including the possibility of distance voice therapy through the Internet and be able to discuss the pros and cons with the client in a knowledgeable and unbiased way. Phonosurgery is discussed in more detail in the section "Surgical Feminization of the Larynx and Voice." Should the client choose voice and communication therapy, the speech-language therapist should communicate clearly about what is involved in the therapy program, including length of practice time, the nature of the exercises, and frequency of practice sessions. This allows the client to make adjustments to daily routines to accommodate practice.

Timing of Therapy
While there are no hard and fast rules about the timing of voice therapy, we should remember that a client's perception of voice may change as the person progresses through the transition process, so it is important to evaluate voice and communication near the beginning of any therapy program (Byrne, 2003(Byrne, , 2007Dacakis, 2012a). The ability to progress in a speech/voice therapy program can be affected by a variety of factors and a frank discussion about choosing the optimal timing of a therapy program is useful. The impact of any planned surgeries on the timing of voice and communication therapy should also be considered.
Some psychosocial factors to investigate are the following: 1. Motivation for change: For a client beginning a new job in his or her desired gender, the stress and motivation can be very high. Others who talk only with supportive family and friends may not feel the same need to work hard on speech/voice change. 2. The amount of time spent in the desired gender role: Gelfer (1999) and Gold (1999) suggest that a transsexual woman living full time is likely to do better in voice therapy. However, the transition process is a personal journey that varies widely among clients. Some clients may wish to feminize/masculinize the voice before they begin to present publicly at all. Others may feel more comfortable with a gender-neutral presentation while they work on their voice and communication. As with all aspects of the transgender process, the client's wishes should be respected. 3. Availability of practice opportunities: Does the client have sufficient time to devote to a structured exercise program? Is the client socially isolated or have a wide network of family, friends, and acquaintances with whom to practice?

Mental Health and Stability
Transitioning is, by definition, a time of great upheaval and psychosocial stress. Before entering speech/voice therapy a client should reflect on whether they have the time and stability in life to undertake it. Issues of mental illness or substance abuse will obviously affect compliance with any therapy program. If psychosocial issues are significantly impacting treatment, referral to a trans-competent mental health professional may be useful.

Minimal Standards
To date there are no minimal standards for the assessment of voice and communication in transgender clients. The field of practice is recent and the evidence base, still weak. It is recommended that once the evidence is sufficiently strong international minimal standards be developed. The establishment of minimal standards not only supports best practice among clinicians around the world but also, when standard assessment procedures are used, facilitates the collection of data. However it must be stressed that clients need to give their explicit permission for this data collection and distribution.

COMPONENTS OF A PROGRAM TO FEMINIZE VOICE AND COMMUNICATION
In this section we will recount evidence-and practice-based approaches to feminizing voice and communication. For information on masculinization of voice and communication see the section "Masculinizing Speech: Physiology and Identity."

Preparing for the Process of Voice-and-Communication Feminization
Most expert speech-language therapists agree that feminizing voice and communication is a long process requiring considerable work on the client's part (Davies & Goldberg, 2006a). While therapy outcomes cannot be predetermined, the estimated amount of daily practice time and expected duration of the course of therapy should be discussed, as should the factors that can influence the course of therapy (Byrne, 2003). The process requires altering deeply ingrained communication habits and behaviors, which can be challenging to modify. It may be useful to use the Stages of Change model (Prochaska, DiClemente, & Norcross, 1992;Zimmerman, Olsen, & Bosworth, 2000) or other behavioral change tools to assist in anticipating and addressing barriers to implementing change. If pitch-changing surgery is sought, there should be discussion of the parameters of speech and voice that may still need work after surgery, as well as an estimation of the healing time involved and the time required to stabilize the new pitch (Dacakis, 2002;Neumann & Welzel, 2004;Wagner et al., 2003).
Some clients are interested primarily in an assessment of their speech and a professional opinion on what would be involved in changing it. Information about therapeutic options can help with decisions regarding the timing of gender transition. Specific objectives relating to voice modification depend on what is feasible to produce without strain, what fits with the client's selfimage, and how important passing is to the client (some clients may be comfortable with gender-neutral speech, while others will want to aim for a voice that is perceived by listeners as cismale or cisfemale).

Education and Information
Clients undergoing speech feminization need information in a number of areas, such as: how the voice is produced; physiological differences in biological male and female voice production systems; how these differences affect voice production and listener perception; gender-salient parameters of voice and communication specific to each client's (linguistic and cultural) community; treatment options, outcomes and risks; and strategies to prevent vocal fatigue or vocal injury.
While some transgender individuals are extremely well informed, others have no knowledge or have been exposed to inaccurate information via the Internet or peer groups.

Pretherapy Voice Conditioning
In a pilot project Dacakis (2012b) used Stemple's Vocal Function Exercises (Stemple, Lee, D'Amico, & Pickup, 1994) with a group of 19 participants recruited from a voice therapy waiting list. This voice exercise program included four sessions over a 7-week period. There were statistically significant improvements in a number of voice parameters and the clients perceived the therapy to be effective. Initial results suggest that this kind of pretherapy conditioning may be useful, particularly for distance therapy, as it is easy to monitor over online audiovideo telecommunication software, such as Skype, for transwomen not yet presenting as female and for women on a voice therapy waiting list or without ready access to weekly therapy.

Treatment Format: Individual and Group Sessions
Traditionally voice and communication therapy has been delivered in a 1:1 setting to facilitate the personalized intervention necessary to modify and monitor change in target behaviors. Determining appropriate average speaking pitch, working with aspects the client finds especially challenging, and dealing with dysphonia are all best conducted in a 1:1 setting. However, there can be advantages to conducting therapy, at least partially, in a group setting. A group format is ideal for participants to share observations, insights, and practical advice. A group can facilitate peer support and reduce self-consciousness that a client may experience when working alone with a therapist (Davies & Goldberg, 2006a). Information about voice care and voice change can also be given efficiently in a group.
Some programs incorporate both group and individual formats. Examples are found in Davies and Goldberg (2006a) and in Mordaunt (2012). Group therapy may be used before 1:1 therapy for education purposes and to introduce participants to the process of change or after 1:1 therapy to offer support and further practice opportunities. In Davies and Goldberg (2006a), group therapy and 1:1 sessions are interspersed.

Length of Treatment Time
The length of treatment for transwomen generally ranges from a minimum of 15 hr to a maximum of 1 year of weekly sessions. Studies investigating optimal length of therapy time have shown mixed results. Some studies suggest that longer time in therapy correlates with higher SFF (Dacakis, 2000;Hancock & Garabedian, 2013). Two other studies demonstrated improvement after intervention of 5 weeks and 8 weeks, respectively (Carew et al., 2007;Gelfer & Tice, 2013). In the Gelfer study these changes lasted, at least in part, for up to 15 months after treatment.
Optimal treatment time for each individual will vary, depending on the degree of change sought, the client's natural abilities, the amount of time and energy the client devotes to practice, and the presence of co-occurring psychosocial difficulties. A short program may promote increased focus and motivation while a longer course of treatment allows more time for habituation and generalization.
Psychosocial adjustment is an important part of changing speech and voice. Participants may require time to get in touch with what sort of voice best matches the person within. This is by necessity a process that takes time and professional input to assess what is desirable and possible. Many transgender individuals begin with the goal of having a pitch that is unrealistically high (MtF) or low (FtM); only with experimentation and practice will it become apparent that this is probably not achievable, necessary, or even desirable. Additionally, it can take time to feel that an altered voice is an authentic expression of self rather than an artificial "mask."

Speech and Voice Therapy Protocols
Speech-language therapists are in the difficult situation of having to provide service in speech/voice feminization in the absence of evidence-based protocols. In this section we have chosen protocols that include standard outcome measures and those modeled on evidence-based protocols from the voice therapy literature with cisgender people.
Examples of two complete protocols for speech/voice feminization may be found in Voice and Communication Therapy for the Transgender/Transsexual Client: A Comprehensive Clinical Guide (Adler, Hirsch, & Mordaunt, 2012) and in Davies and Goldberg (2006a). These protocols should be used advisedly as their efficacy and effectiveness have not been established in the literature.

Perceptual-Motor Learning Principles
In their recent book, Vocology, Titze and Verdolini Abbot (2012, pp. 217-238) discuss perceptual-motor learning principles and how they apply to the remediation of voice disorders (pp. 217-238). These apply equally well to voice and communication feminization and the reader is referred to this book for a comprehensive discussion of the principles underlying speech and voice change. The principles include the following: using target sensations or images as input, rather than instructions; the need for attention and concentration on producing and reproducing the target sensation; progressing from simple tasks to more complex ones; and performing and learning behaviors together (such as breathing and voicing) if they should be performed together.

Vocal Stretching, Vocal Conditioning
Voice training protocols frequently begin with vocal flexibility exercises.

Vocal Function Exercises
Stemple's Vocal Function Exercises (Stemple et al., 1994) have been used to facilitate a wider pitch range and train efficiency of voice production (Dacakis, 2012b;Gelfer & Van Dong, 2013). Dacakis used them for pretherapy voice conditioning and found they were beneficial. Gelfer and Van Dong used vocal function exercises along with their standard voice feminization program. They found that while listeners rated posttreatment voice recordings as more feminine, the addition of vocal function exercises did not improve outcomes, as compared with previous literature.

Semi-Occluded Vocal Tract Techniques
Semi-occluded vocal tract (SOVT) techniques may be used to facilitate voice production in the upper part of the pitch range; train a "head voice" or "mixed registration"; encourage efficient vocal fold vibration; and stimulate a forward sensation of the voice, or a high voice placement (Verdolini-Marston, Burke, Lessac, Glaze, & Caldwell, 1995). SOVT techniques include phonating into straws or tubes, lip or tongue trilling, the use of nasals /m/, /n/, voiced fricatives, e.g., /z/, /v/, and closed vowels, e.g., /u/, /i/. Two exercises are often used: a pitch glide throughout the middle and upper pitch range and a messa di voce exercise, going from soft to loud and then soft again (Titze, 2000;Titze & Verdolini Abbot, 2012). Prolonging the voice at a higher pitch through a semioccluded vocal tract may develop efficiency of voice production at a higher nonhabitual pitch (Davies, 2012;Davies & Goldberg, 2006a).

Raising the Speaking Fundamental Frequency
These techniques have been used by experienced speech-language therapists to train an efficiently produced higher average speaking pitch.
1. Using trial therapy in the assessment session the clinician and client find a speaking pitch that is either in gender neutral or low cisfemale range and can be produced without strain. This is used as a training pitch and is prolonged using semi-occluded-vocal-tract techniques. Speech sounds, words then longer utterances are intoned then modulated in a natural way, always keeping the voice easily produced without the need for excessive vocal effort (Dacakis, 2012a;Davies & Goldberg, 2006a;Gelfer & Van Dong, 2013;Gelfer & Tice, 2013). 2. The client starts near the habitual speaking pitch and slowly raises the average speaking pitch throughout the therapy sessions. This continues until both the client and the clinician feel the pitch is appropriate and manageable (Mordaunt, 2012). 3. The speech-language therapist asks the client to sing "do re mi" then hold the pitch of "mi" and use that as a target for a higher, easier voice (Bennett, 2010). 4. The client uses a pitch well into cisfemale range, such as 220 Hz, as a target for pitch training (Perez, 2004).

Raising Vocal Tract Resonance Frequencies
1. Instruct the client to use a forward tongue carriage and slightly spread lip posture during speech. This technique was found to be effective in a clinical study of 10 transsexual women (Carew et al., 2007). 2. Use /i/ as a target mouth shape for high, forward tongue placement in other vowels (Mordaunt, 2012

Language and Discourse Pragmatics
While there are strong social stereotypes about gender norms and language (e.g., use of slang, intensifiers, and tag questions), genderassociated norms of language and discourse are so dependent on an ever-shifting social context that findings from studies conducted in past decades may not be reflective of current trends in the client's speech community (Freidenberg, 2002;Oates & Dacakis, 1983). Additionally, there is strong interplay between gendered language norms and norms relating to culture, class, and age (Hasegawa & Hata, 1995;Linville, 1998;Moran, McCloskey, & Bady, 1995;Morris & Brown, 1994), so norms appropriate for one client would not be appropriate for another. If there are habits relating to vocabulary and syntactic choices, such as the use of modifiers, qualifiers, indirect versus direct speaking style, that the client finds discomforting or the clinician feels may contribute to perceptions that don't fit the client's self-image, we recommend that the speechlanguage therapist offer feedback in these areas, encouraging the client to weigh research findings or the clinician's suggestions against personal experience.
We recommend that modification of language and discourse be based on the client's own observations of gender markers in the specific environmental context of concern to the client (e.g., work, home, cultural community, social setting). Clients with strong beliefs about "appropriate" language may benefit from the speech-language therapist's assistance to explore stereotype versus actual observed behavior of peers.
A discussion of gender-dependent communication variables in cisgender people, together with exercises, can be found in chapters 13

Nonverbal Communication: Facial Expressions, Posture, and Movement
Norms relating to posture, gestures, and other nonverbal aspects of communication are strongly influenced by many things including cultural, class, and age norms. These behaviors also constantly change depending on many variables such the speaker's mood and conversational partner or situation. It is not within the normal scope of practice of a typical speechlanguage therapist to provide a detailed assessment of nonverbal communication behaviors. If a client wishes input that the clinician feels unable to give, it may be possible to refer to a trans-competent-speech-language therapist who has training in nonverbal communication.
Other options include a discussion of the nonverbal parameters of communication, together with structured observations to be conducted within the client's community. This can help foster awareness and sensitivity to subtle changes in nonverbal communication, which could be more useful than learning and following rigid patterns of behavior.
A discussion of nonverbal communication and therapy protocols is found in Adler, Hirsch, and Mordaunt (2012): Hirsch and Van Borsel (chapter 15) and Boonin and Hirsch (chapter 16). These chapters are based on the authors' experience and observations and on literature of cisgender men and women.

Habituation and Generalization
The ability to use a feminine voice in everyday life is a significant achievement and also the goal of many transwomen who embark on speech/voice change. Some strategies can promote carryover into everyday life: practicing words that are typically part of daily conversation (e.g., hi, bye, yes, no); in clinical practice of conversational speech, focusing on situations/topics related to the client's life and engaging in role plays suggested by client to match real-life situations that pose the most difficulty (e.g., job interview, coffee shop interaction; e.g., Hooper, 1985); experimenting with emotional intensity by practicing sentences expressing feelings like joy, sorrow, irritation or anger; and mindful practice outside the clinic setting in a wide variety of real-life settings, depending on comfort level The process of habituation of any new motor act takes mindful practice over a period of time (Adams, 1971). Maintaining change outside the clinical setting is challenging and until recently has not been measured. In 2007, Ma et al. used Soundswell's computerized real-time phonetogram Phog 1.0 (Hitech Development AB, T€ aby, Sweden) to record voice in everyday situations among cisgender subjects with dysphonia (Ma et al., 2007). In 2009, S€ odersten, Hertega rd, Larsson, and Holmberg (2009) researched the use of a portable dosimeter called the Ambulatory Phonation Monitor (APM) in one MtF client to monitor voice use outside the clinic. When her fundamental frequency decreased under a certain threshold level a tactile feedback signal was given to remind her about raising her fundamental frequency. Also the portable Voxlog system can provide tactile feedback when the speaker exceeds or goes below a certain fundamental frequency threshold level. This could be useful for transgender speakers. Dacakis et al. (2012) also suggest such a device may prove useful in a clinical context. In North America the dosimeter that is commercially available is the Ambulatory Phonation Monitor (APM), developed collaboratively by Kay PENTAX and speech-language therapists from Massachusetts General Hospital.

Follow-Up Sessions
Follow-up sessions after the initial treatment has finished, or facilitated support groups for ongoing practice, may be useful in maintaining and entrenching changes made in therapy. Clinically supervised follow-up also provides an excellent opportunity to gather much-needed data about the effectiveness of a program over time.

Follow-Up Sessions (Group/Individual)
There is as yet no empirical evidence regarding the optimum frequency for follow-up sessions, the optimum content, or the criteria for termination. In the absence of data, we suggest that refresher sessions could be initially offered 3 months after treatment and then at 4-to 6month intervals or as the clinician and client deem appropriate.
Follow-up sessions can include a discussion of successes, problems, strategies, and difficulties the client has experienced since the end of therapy; a review of the core exercises of the program (to ensure the client is practicing correctly and to determine if the exercises are still appropriate); and time to address any concerns that have arisen since the end of treatment. Ideally, follow-up would include re-evaluation of the same parameters measured in the pretreatment assessment, both to assess the maintenance of the desired changes and also to evaluate the effectiveness of refresher sessions.
If the initial therapy was provided in a group setting, a group setting is a natural forum for refresher sessions. As with group format for initial therapy, group format for refresher work offers valuable opportunities for clients to compare experiences. In our experience this can be most useful and encouraging, especially for those in the early stages of their transition. Individualized follow-up may be more appropriate than the group format if the client has numerous concerns or unusual concerns that require individual attention or if the client feels uncomfortable in a group setting.

Client-Run Voice and Communication Support Groups
Self-help groups may be useful for transgender individuals who have completed clinical treatment and are seeking peer support to maintain or strengthen voice and communication changes. Client-run voice and communication groups can provide motivation to maintain practice, a forum to practice and to share ideas and concerns, and an opportunity to socialize and undertake specific role playing. Client-run groups can also foster the client's sense of ownership and control of speech and voice production so that he or she feels less dependent on the therapist.
In any self-help group there is a danger that an individual may inappropriately assume a professional clinical role. In a communication group, this could be circumvented by providing group facilitation training to members, by having the speech-language therapist as guest visitor from time to time, and by having self-help sessions along with therapist-run refresher sessions.

IMPROVING ACCESS TO SERVICE Distance Assessment and Therapy
At the time of writing, online applications and forums such as the ones provided by Skype and Google Hangouts offer opportunities for interactive distance voice and communication therapy. These recent technological advances have the potential to greatly expand access to voice and communication services-one of the main difficulties experienced by many transgender people. However, speech-language therapists should be mindful of possible concerns with a lack of confidentiality inherent in Internet communication.
In her presentation at the WPATH 2011 symposium, Dacakis (2011) outlined a protocol for this kind of service delivery. Video conferencing is being used with increasing frequency in many health care settings. It has the potential to provide good visual and audio contact between speech-language therapists and geographically isolated clients and could also serve as a training modality for speech-language therapists in the care of transclients.

Improving Accessibility and Utility to Clients With Access Barriers
The transgender community is tremendously diverse, and protocols must be flexible enough to address this diversity of service needs and issues relating to access. In the transgender voice and communication literature reviewed for this project, there was little discussion of modification to address the needs of clients who have difficulty accessing the typical setting or format of voice and communication services-including individuals who have speech, hearing, cognitive, or learning disabilities, are not highly fluent in the primary language spoken where they live, or are not literate or who are geographically isolated or cannot leave a residential facility, such as a prison or long-term care home. Without empirical evidence to guide practice, we offer the following suggestions based on our experience providing services to a diverse range of transgender clients.
For individuals who do not speak the speechlanguage therapist's language, an information package could be translated into a variety of languages. Providing voice and communication services in another language is not possible unless the speech-language therapist speaks the client's language well, as subtleties of inflections, inflectional range, word stress, and semantic and syntactic choices require a thorough knowledge of the language. The only direct therapeutic input that could perhaps be given would be in changing the speaking fundamental frequency and formant characteristics. Speech-language therapists who are multilingual could be supported to take transspecific training, perhaps working in consultation with a more trans-experienced speech-language therapist to provide service in the client's primary language. If the client is partially fluent in the primary language of their community, wishes speech therapy in that language, and will be using it in everyday life, therapy delivered in that language can be beneficial as the client has the opportunity of learning more-feminine/masculine patterns of speech as they acquire the language. For individuals who are only partially fluent in the language used in therapy, the therapeutic process will likely be longer and will require much more 1:1 input.
Transgender clients with speech or hearing disabilities who are able to attend voice and communication therapy sessions may find great benefit from using visual input during speech therapy. This has been used with good success with other populations (e.g., using palatography and ultrasound to work with phonological disorders in people with hearing impairments). For transgender clients, there are a number of software programs that record fundamental frequency and allow the creation of a "model voice wave." The speech-language therapist can record a desired average speaking pitch or an intonation pattern and the client could then use the visual input to copy it; alternatively, the speech-language therapist could record the lowest (MtF)/highest (FtM) desirable frequency and the client could use the visual input to keep the speaking pitch above (MtF) or below (FtM) this line.
Online and/or offline free pitch-measuring programs can be useful for clients' home practice. Speech-language therapists should be alert for other emerging applications for electronic communication devices. Some currently available programs include the following: If a client has cognitive or learning disabilities, depending on the nature of the disability it may be useful to include a loved one or care aide in the therapeutic process. This person could help the client establish a regular practice schedule and give input to the exercises, under the guidance of the speech-language therapist. A different format may be useful for the client who has difficulty processing the information necessary to change speech habits. Rather than using an approach that requires introspection (e.g., "How does that sound? Am I feeling my voice in my face?"), the speech-language therapist may be more directive in determining which exercises would most useful and could be performed appropriately by the client; the clinician and client together would draw up a practice schedule and the client would simply practice the motor movements outlined. Individualized attention is likely to be more effective than group work for providing the client with more intensive input. To be successful, this kind of format would require regular clinical intervention and support outside the therapy room.

Self-Guided Speech Feminization
There are many videos, websites, CDs, and other materials available for self-guided speech feminization. We cannot comment on the efficacy or safety of these materials, but we are concerned that (a) many are not produced by speech-language therapists and (b) there are risks associated with attempting to change the voice without professional assistance. Speech feminization/masculinization involves substantial changes in habitual production and so has the potential to cause a voice disorder or aggravate an existing one. It is strongly recommended that anyone seeking to feminize or masculinize speech first be assessed by a speech-language therapist, that a speech-language therapist be involved in monitoring progress, and that a speech-language therapist be consulted if there are any symptoms of vocal fatigue or negative changes to vocal quality. Additionally, we recommend that consumers be cautious of any materials promoting a rigid set of speech norms, as speech is highly individualized and culturally driven and cannot be guided solely by a set of generic rules.
Over the past 10 years, study and practice in the area of transgender voice and communication has expanded and deepened. We look forward to greater understanding and better practice as this field continues to grow.

SURGICAL FEMINIZATION OF THE LARYNX AND VOICE
Transgender people wishing to feminize their voice or appearance through surgical means may choose to undergo two different kinds of surgical procedures: 1. A thyroid chondroplasty, which is a cosmetic procedure that reduces the size and angle of the thyroid cartilage to resemble the flatter, smaller cisfemale thyroid cartilage. It has no feminizing effect on the voice (Sandhu, 2007). 2. Phonosurgery, which can raise the speaking fundamental frequency and eliminate lower frequencies in the cismale speaking range (Spiegel, 2006).

Thyroid Chondroplasty
A thyroid chondroplasty, also called a tracheal shave, laryngeal shave or Adam's apple reduction, may be performed alone or in addition to a phonosurgical procedure. This surgery may be performed by an otolaryngologist or plastic surgeon. The surgeon makes the incision in a crease at the front of the neck, directly over the larynx. In Sandhu's opinion "Unless there is a history of keloid scar formation, the incision is not easily visible after a few weeks. The advantage of this technique is that it allows for the dimensions of the larynx to be measured accurately for maximal and safe removal of the laryngeal prominence" (Sandhu, 2007, p. 196). An alternative scar location is a small incision high on the midline of the neck. Although the advantage is that the scar is hidden under the chin, the scar cannot be placed in a natural crease and the patient is more likely to develop adhesions between the skin and the larynx so that there is "puckering" of the skin with swallowing. Sandhu (2007) notes there may be a small risk of detaching the vocal cords anteriorly, with an associated risk to the voice and airway. Spiegel (2006) notes that during this procedure it is important to preserve the 138 Davies, Papp, and Antoni anterior attachment of the true vocal folds, otherwise an undesirable lowering of pitch can result. A summary of outcome data is found in Matai, Cheesman, and Clarke (2003).

Voice Feminization Surgery
As was discussed in the section "Average Speaking Fundamental Frequency," a higher speaking fundamental frequency (SFF) is considered to be one of the most important components of a feminine voice and transwomen may undergo phonosurgery to raise both the SFF and frequency range, as well as to eliminate access to lower frequency ranges (Remacle et al., 2011;Spiegel, 2006;Wagner et al., 2003). Van Borsel found that the "perception of femaleness significantly correlated with the average fundamental frequency" (Adler & van Borsel, 2006). However, as was also discussed earlier, there are a number of voice characteristics that are believed to combine to produce a feminine voice. These include resonance characteristics, articulatory contacts, voice quality, and intonation contours. Adler and van Borsel (2006) noted that while a pitch-raising surgery may be a viable option to increase the speaking pitch in MtFs, this surgery alone may not be sufficient to produce a voice that is perceived as totally female.
A number of reasons support the use of phonosurgery to feminize the voice in transgender women. The availability of speech-language therapists with this particular clinical competency may be limited in the area where the individual lives. Treatment outcomes may be influenced by the client's commitment to therapy, level of practice outside of the therapy sessions, social confidence level, and emotional health. Whilst some clients show an aptitude for voice work and possess a "good ear," which aids their skill acquisition, other clients find voice work challenging on a physical and/or emotional level. Some clients struggle with vocal fatigue, and to maintain a raised pitch, even with sustained voice therapy, progress may be limited (Adler, Constansis, & van Borsel, 2012). Antoni (2007) notes, "Limited data exist regarding the formal outcomes of vocal surgery with transsexuals" (p. 144). There is a paucity of outcome data for pitch-elevating surgery, particularly longitudinal data to monitor outcomes over time. In general, professional opinion is mixed about voice surgery, with some clinicians stating that it is not a viable treatment option (Andrews, 1999;Koufman, 2007), and others recommending and others recommending that voice surgery be considered a treatment of last resort for MtFs who have not experienced satisfactory increase in pitch following speech therapy (Lawrence, 2004;Oates & Dacakis, 1997;Wagner et al., 2003). Certainly the reported negative effects (compromised voice quality, diminished vocal loudness, adverse impact on swallowing or breathing, sore throat, wound infection, and scarring) and variable outcomes are a concern (Brown, Perry, Cheesman, & Pring, 2000;Dacakis, 2002;Koufman, 2007;Lawrence, 2004;Neumann & Welzel, 2004;Oates & Dacakis, 1997;Thomas, 2003;Wagner et al., 2003;Yang, Palmer, Murray, Meltzer, & Cohen, 2002).
However, some proponents of voice surgery suggest that surgery can protect the voice from damage caused by strain to elevate pitch through nonsurgical means (Brown et al., 2000;Neumann & Welzel, 2004;Yang et al., 2002). Given this, while there are clear risks of voice surgery and the decision to pursue it should be carefully considered, the decision about risk-benefit ratio and preferred technique is best left to the patient, with input from both a trans-experienced surgeon and a trans-experienced speech-language therapist.

Presurgical Assessment
In addition to the standard screening conducted prior to any surgery (assessing for risks relating to anesthesia, infection, etc.), assessment prior to pitch-elevating surgery should include an anatomical and functional assessment of the larynx, the client's subjective assessment of the client's voice, an acoustic analysis, audio or video recordings, and the clinician's perceptual analysis. 2 Particular attention should be paid to the measurement of speaking fundamental frequency and frequency range and of voice quality (Neumann & Welzel, 2004;Yang et al., 2002). Any dysphonia present before surgery should be addressed through standard voice therapy protocols prior to voice surgery. Care should be taken to ensure that the patient understands the risks and anticipated outcome of the technique that will be used.
After finding that some subjects have strained and unnaturally elevated voices following surgery, attributed to habitually speaking at an artificially elevated pitch for sustained periods of time prior to surgery, one surgical group reported testing for ability to phonate at a pitch within the masculine range as part of preoperative consultation. Clients who are unable to do this were felt to have the equivalent of a muscle tension dysphonia and were referred for preoperative voice therapy to recover the ability to produce relaxed phonation (Yang et al., 2002).
Estrogen is associated with risk for deep vein thrombosis and pulmonary embolism. If the patient will be immobilized for a prolonged period during or following surgery, consultation with the prescribing physician is necessary to discuss the advisability of tapering estrogen use before surgery.
Smoking increases the risk of complications from anesthetic and impairs healing, and there is evidence that smoking following voice surgery can negatively impact voice quality and pitch (Wagner et al., 2003). Patients should be made aware of the risks associated with smoking and of smoking cessation resources and and should be strongly encouraged not to smoke prior to or immediately following surgery.

Postsurgical Care
Postsurgical care depends on the specific surgical technique employed. The surgeon should review aftercare instructions with the patient as part of informed consent prior to surgery, as well as provide written information to be consulted before and after surgery. The surgeon should also be easily accessible for questions relating to postoperative complications. If surgery is performed away from the patient's hometown, the patient's local primary care provider should consult with the surgeon to determine appropriate follow-up.
Follow-up evaluation by the surgeon (or postoperative care coordinator if the surgeon is in a different location) is recommended at 1 week, 4 weeks, 12 weeks, and 6 months after surgery to monitor healing and recovery. Immediately following surgery, temporarily decreased pitch, diminished voice quality, and edema were commonly reported, with spontaneous recovery in most cases. Less-common complications that required medical intervention included mild emphysema, neck abscess, negative response to the sutures/plates used in cricothyroid approximation (requiring removal of the material), and loosening of the sutures used in cricothyroid approximation (requiring additional surgery) (Neumann & Welzel, 2004;Wagner et al., 2003).
For most techniques, it is recommended that patients not use their voice at all for 1 to 7 days after surgery and then to use the voice cautiously until any discomfort (e.g., due to postoperative edema) has passed (Brown et al., 2000;Neumann & Welzel, 2004). For the more invasive, combined thyroid cartilage and vocal fold reduction, 2 weeks vocal rest is suggested (Kunachak, Prakunhungsit, & Sujjalak, 2000). Following cricothyroid approximation, steam inhalation may be recommended to hydrate and lubricate the vocal cords to promote healing (Brown et al., 2000).
Speech therapy is recommended following surgery to help the patient adapt to and stabilize the new voice (Neumann & Welzel, 2004;Wagner et al., 2003). If pitch-elevating surgery was performed before other components of speech had been satisfactorily altered, other aspects of voice such as resonance, intonation, and articulation may also need to be addressed through speech and voice therapy (Dacakis, 2002;Neumann & Welzel, 2004).

Voice Feminization Surgeries
With one exception, the thyrohyoid approximation, the voice feminization surgery described in the literature addresses only pitch change.

Pitch-Elevating Surgical Procedures
Cricothyroid approximation (CTA) is the most common voice feminization surgery and the only one that has a body of outcome data. It is also one of the few procedures that does not involve surgical alteration of the vocal folds themselves. In this procedure, "the cricoid cartilage is moved backward and upward and the thyroid cartilage is . . . moved forward and downward. Then the two cartilages are closer, and the vocal cord tension is increased." (Wagner et al., 2003(Wagner et al., , pp. 1158(Wagner et al., -1159. This mimics the contraction of the cricothyroid muscle that tenses and elongates the vocal folds, causing the pitch to increase. The cartilages are then secured in place by stitches and/or metal plates. This procedure, in effect, puts the vocal folds on permanent stretch. The surgery is carried out through a small skin excision over the larynx and may be performed at the same time as a thyroid chondroplasty.
In theory, CTA is reversible, but in some cases scar tissue has permanently fused the cartilages together. Where this fusion has not happened, the stitches/plates can loosen over time, causing the cartilage to move back to its original position and the pitch to drop. Because the vocal folds aren't directly surgically changed by this technique, it may be possible to perform additional surgery if the CTA is not sufficient.
While several studies have described an increase in SFF after CTA (Brown et al., 2000;Kanagalingam et al., 2005;Neumann, Welzel, & Berghaus, 2002;Neumann & Welzel, 2004;Yang et al., 2002), the amount of the pitch increase was variable and the subjective outcomes to the surgery were mixed. Yang et al. reviewed 20 patients and found a mean change in SFF of 57 Hz (from a mean of 145 Hz preoperatively to 202 Hz postoperatively). In 2004, Neumann & Welzel (2004) looked at 67 patients and found that a year post-CTA, 30% of the voices were within cisfemale pitch range and 32% were in a gender neutral range.
Patient satisfaction with post-CTA voice was also mixed. In Yang et al. (2002), 58% of the patients expressed satisfaction and 33% expressed dissatisfaction; 42% described their postoperative voice as "rough or hoarse." When describing how well their voice worked on the phone, 31% said they were never perceived as male, 47% reported being occasionally taken as male, and 22% said they were frequently considered to have a masculine voice. At a follow-up a year or more after surgery, 29% of the subjects reported dysphagia with liquids. A review of CTA surgery by Kanagalingam et al. (2005) found that 79% of patients were satisfied with the results and 71% had increased fundamental frequency both in reading and speaking.
The variability of outcomes after a CTA may result from anatomical differences in the cricothyroid joint itself (Storck et al., 2011;Windisch, Hammer, Prodinger, Friedrich, & Anderhuber, 2010). In 2011, Storck and colleagues sought to determine why a CTA worked well in some patients and not others and to see if there was a way of predicting how successful a CTA would be. Using high-resolution computerized tomography and threedimensional imaging software, they examined 24 cadaver larynges of transsexual women who had undergone CTA. They found that the anatomical structure of the CTJ [cricothyroid joint] influences directly (1) the position of the effective rotation axis and (2) the elongation of the vocal folds. "Based on the location of this effective rotation axis, we could distinguish three groups. In group I (N D 13), the rotation axis was located in the lower third; in group II (N D 5), it was located in the middle third; and in group III (N D 6), it was located in the upper third of the cricoid. The elongations of the vocal fold were 12%, 8%, and 3%, in groups I, II, and III, respectively" (Storck et al., 2011, p. 632).

Shortening the Vocal Folds
Other phonosurgeries are available to transgender women; however, these are more invasive than the cricothyroid approximation and there is sparse literature on both the surgical procedures and the outcomes.
There are a number of reported surgical techniques used to shorten the vocal folds, allowing them to vibrate at a higher frequency (e.g., Gross, 1999). One such procedure is called the anterior commissure advancement. The epithelial tissue of the anterior third of the true vocal folds is ablated, then this tissue is sutured together to form a nonvibrating anterior web (Remacle et al., 2011). An example of a variation of this procedure, in which no sutures are used, is described on the Yeson Voice Center website (http://www.yesonvc.net/disease/ feminization_surgery.asp).
Another example of a voice feminization laryngoplasty technique is described by Thomas and Macmillan (2013). In this procedure, a 10-to 14-mm wide segment of the anterior thyroid cartilage is removed. The true and false vocal folds are split in the anterior midline and the anterior portions of both are excised. The true vocal folds are then sutured together to create a new anterior commissure and the thyroid alae are brought back together to create a smaller larynx.

Other Procedures
In some more recent patients of Dr. Thomas, an additional surgery, a thyrohyoid approximation, was performed. The thyroid cartilage was raised and sutured to the hyoid bone to diminish, as much as possible, the thyrohyoid space. The goal in this procedure is to create more feminine resonance characteristics by shortening the vocal tract.
In discussing outcomes Thomas and Macmillan (2013) note that the feminization laryngoplasty elevated SFF into cisfemale range and resulted in a permanent loss of bass notes. They also provide references for online voice recordings of patients who had undergone both of these surgeries. However, we could uncover no further literature discussing these surgical procedures and would advise caution in undertaking a highly invasive surgery for which there is, as yet, little outcome data.
Another phonosurgery available to transgender women is Laser-Assisted Voice Adjustment or LAVA. In this procedure a CO2 laser is used to vaporize portions of the vocal folds, resulting in a decrease in mass and increase in stiffness. The scarring resulting from this procedure produces a significant rise in pitch (Orloff et al., 2006). Orloff et al. conducted a retrospective review of 31 patients, with a mean follow-up time of 23 weeks. While pitch increases averaging 26 Hz were reported, the evaluation also showed decreased vocal quality, loudness, and vocal range. In 2010, Kocak et al. described the use of a modified LAVA procedure in a small study of 3 patients whose voices were not improved by cricothyroid approximation. After the vocal folds had been vaporized with a CO2 laser they were tensed medially and stabilized with sutures. The authors reported higher SFF and improved patient satisfaction after surgery (Kocak et al., 2010).
In summary, there is a variety of phonosurgical procedures available to transgender women wishing to increase the pitch of the voice. However, variable outcomes and paucity of outcome data suggest that these procedures be approached with caution and that they should be coupled with voice therapy to address other gendered aspects of speech and voice production (Dacakis, 2002(Dacakis, , 2012bOrloff et al., 2006;Spiegel, 2006).

MASCULINIZING SPEECH: PHYSIOLOGY AND IDENTITY Introduction
The lacuna in both the global research efforts on and therapeutic treatment of the FtM vocal changes does not make it possible to provide a summary of current good practice. Instead, this section of the companion document attempts to summarize and interpret the results of the existing evidence-based research and give an overview of gendered voice and speech issues that a speech practitioner working with FtM people needs to be aware of.
Although the effects of sex hormones on the human voice have been widely documented, up until recently very few studies addressed transgender individuals directly. Instead, practitioners had to rely on studies that explored the vocal maturation of biological males (e.g., Titze, 1994), women's voices changing due to endocrine problems or the treatment of their symptoms (e.g., Pattie, Murdoch, Theodoros, & Forbes, 1998), and the ageing voice (e.g., Titze, 1994).
A small number of recent studies have begun to document the effects of exogenous androgen hormones in transgender men, or FtM transsexuals (Adler & van Borsel, 2006;Constansis, 2008;Damrose, 2009;Papp, 2011a;van Borsel et al., 2000;van Borsel, Cuypere, & van den Berghe, 2001). Although other researchers have presented pitch change data for transitioning transgender men, longitudinal studies are still scarce, especially ones that interpret the changes using a psycho-acoustic pitch-scale for accurate within-and between-subject comparisons. As it will be shown below, besides the changes in the speaking fundamental frequency (SFF), it is important to also monitor the changes in the available maximum phonational range (MPR)-that is, the distance between the lowest and highest phonational frequencies (cf. p. 49). In addition to documenting the amount of change during the first year on testosterone, it is important to characterize the timing of the changes and following the changes for possibly multiple years. The motivation behind this is that some of the most frequent questions from FtMs commencing testosterone therapy address (a) how soon the changes are noticeable, (b) how low their SFF may reach, (c) how long it takes for the mean fundamental frequency to reach its lowest and settle, and (d) how their singing voice might be impacted by the changes.
Several newer studies on the FtM voice challenge existing assumptions about the FtM vocal transitioning process. Specifically, results in Papp (2011a) and Zimman (2012) do not support some theories of transitioning (e.g., Heath, 2006) that view the FtM transitioning process as striving for an unambiguously masculine end result. Similarly, Papp's and Zimman's results are not consistent with theories of gender that define gender in terms of a binary opposition of male and female categories and of transgenderism as the "switching" from one feature set to another. Instead, their data suggest that many transgender men view gender as a continuum and they are often comfortable with (a) moving in general toward a more masculine, as opposed to unambiguously masculine, self-representation and (b) keeping their vocal gender fluid and adjusting features of it to circumstances, which may even call for demasculinizing vocal descriptors.
The difference between habitual and "passing" pitch is of special interest in transsexual and transgender populations. In the case of MtF transsexuals who went through puberty as males, the question is how speech-language therapists can safely retrain speakers to use a potentially very different range in their repertoire for everyday communication without voice quality deteriorating or becoming aphonic, without losing intonational variety and projecting power, and without developing vocal pathologies. The little documentation available on the hormonal transitioning of FtMs usually shows substantial lowering of the fundamental frequency (Adler, Constansis, & van Borsel, 2012;Papp, 2011a;Zimman, 2010). There is little information, however, on whether speakers' actual habitual pitch lowers in tandem with the predicted habitual or optimal speaking pitch (usually calculated with reference to the pitch floor or the Maximum Phonation Range). Negotiating the distance between a "passing" pitch versus the actual speaking fundamental frequency is very likely a task for the speechlanguage therapist. As there is evidence that listeners are able to locate a speaker's SFF within their phonational range without having been exposed to the speaker's entire range (Bishop & Keating, 2012;Honorof & Whalen, 2005), speech practitioners need to be aware that the gendered perception of speech is not a trivial matter that purely depends on the SFF and speaking range of the SFF but depends on a host of other acoustic descriptors that we still know fairly little about.
Related to the accurate measurement of the acoustic correlate of the speaking pitch, the fundamental frequency, is the issue of the psychoacoustic representation-that is, the comparability of the values, both within and between clients. It is often convenient to measure absolute frequencies in Hz when pitch or change in pitch is studied. But while the production of frequency is linear by nature, the perception of frequency is largely logarithmic. As a result, logarithmic or other nonlinear auditory scales of frequency representation are required to accurately characterize perceptual changes in the speaking fundamental frequency. One such recommended scale is the semitone (ST) scale. On one hand, being one of the few commonly used scales both in clinical practice and musicology, the results are easily interpretable for the widest possible audience. Additionally, the ST scale is the only scale with only one agreed-upon formula that has remained virtually unchanged in the past many centuries, unlike the Bark, Mel, and Equivalent Rectangular Bandwidth (ERB) scales (Miller, 2002).

Speaking Fundamental Frequency
There is substantial evidence that androgens in adolescent cismen, ciswomen with hyperandrogenism (excessive production and/ or secretion of androgens in women), and MtF clients who went through adolescence as males not only lower the SFF, but they do so permanently. The change in SFF is interpreted to be caused by the increase of vocal fold mass with possible contribution from vocal fold stiffness as well. The irreversibility of this change has been amply documented in the literature on cismen, ciswomen with hyperandrogenism, and MtF transsexuals. As an example, case studies of females in androgen therapies establish that this noticeable deepening of the habitual pitch happens in the first 3 months of the therapy, and is permanent and irreversible even on cessation of the androgens (Baker, 1999;Barbieri, Evans, & Kistner, 1982;Boothroyd & Lepre, 1990;Mercaitis, Peaper, & Schwartz, 1985;Newman & Forbes, 1993;Pattie et al., 1998;Wardle, Whitehead, & Mills, 1983).
Evidence on SFF changes in FtM individuals is more scarce though. Van Borsel et al. (2000) and Adler, Constansis, and van Borsel (2012) present longitudinal data of two FtMs, with data points collected about every 2 months. Damrose's (2009) subject was a 33-year-old semi-professional jazz singer, recorded monthly for 20 months. Papp (2011a) recorded seven subjects fortnightly during the first year of testosterone therapy and measured SFF and formant changes. Zimman (2012) recorded 10 transgender men eight times during a window of about 40 to 60 weeks, most of them in the first or second year of transitioning. The changes in SFF in the above studies seem to be following an S-curve: after the initiation of the androgen therapy there is a marked drop in SFF that plateaus out after a number months. The majority of the changes are completed within a year, but Papp (2011a) observed further lowering during the second year on testosterone, and also as a result of oophorectomy (the removal of the ovaries) and the onset of menopause. Papp also found previously undocumented patterns in which some transgender men show a reverse J-pattern in pitch changes: after the initial drop and plateau, the habitual pitch slightly rises. As an upward shift is deemed impossible on purely physiological bases, Papp hypothesizes that this is due to transgender men fine-tuning their habitual SFF to match their gender identity and express nuances (gender expression and sexual orientation) within their gender.
In Papp's experience some transgender men have actively or passively resisted speaking in a lower (chest) voice. In LW's account, when he felt a strange "rumbling" in his chest he usually recalibrated himself to a higher pitch, unless it was a formal situation demanding authoritative behavior, in which he maintained the low pitch. As another example, TJ would often ask whether his voice had changed during the course of the androgen therapy-by that he meant an observable change in the mean fundamental frequency, which he himself did not perceive as changing. He was indeed correct: over 1 year on testosterone his mean habitual pitch changed from 200 to 170 Hz, which is a change of 3 ST. At the same time, however, his pitch floor lowered from 130 to 80 Hz, or more than 8 ST. If his habitual pitch had recalibrated to a predicted 4 ST over his pitch floor he would have been speaking at around 100 Hz, which is an unambiguously masculine pitch. When asked if he tried to play with his voice and speak lower, TJ's answer was that he had tried and since "it sounded weird" he reverted (close) to his pretestosterone habitual pitch.
It has been observed that some transgender men have problems "passing" as males well into their time on testosterone when their predicted optimal pitch based on their available pitch floor (cf. Awan, 1993;Britto & Doyle, 1990;Cooksey, 2000;Boone, McFarlane, Berg, & Zraick, 2009) and/or their habitual pitch was unambiguously masculine (Nygren, 2014;Papp, 2011a;S€ oderpalm et al., 2004;van Borsel et al., 2000). These findings make a strong case for including a speech-language therapist in the routine medical care of transgender men. The role of a speech-language therapist could include both mapping the habitual speaking fundamental frequency and intensity within the available pitch range of transgender men and facilitating behavioral changes or making referrals to other professionals, such as endocrinologists and ear-nosethroat specialists (ENTs), as a trained specialist is able to make distinctions between transgender men reporting temporary, developmental phenomena versus potentially acute problems relating to the habitual speaking fundamental frequency.
One type of problem encountered in practice is when transgender men exhibit little or no lowering of the habitual fundamental frequency and the pitch maxima, even though they have been on testosterone for over a year. This is likely a physiological-hormonal problem, which needs to be investigated with the help of a speech-language therapist, an ear-nose-throat specialist, and/or an endocrinologist. Another kind of problem is reported by individuals who demonstrably have access to cismale-normative fundamental frequencies (which they may or may not know about) but do not seem to be able to speak at those frequencies habitually. These transgender men would then need to be retrained to realize the existence of and then adjust to frequency ranges that are conducive to presenting as the desired gender. Naturally, a speech-language therapist would not need to work with the third group of individuals who, while having access to a cismale SFF, exhibit higher-than-cismale habitual frequency values and are happy with the way they are read (be it gay, queer, androgynous, effeminate, or on occasion, female).

Maximum Phonation Range
Maximum phonation range is the physical range where an individual is able to phonate audibly (cf. Hallin, Fr€ ost, Holmberg, & S€ odersten, 2012;Holmberg et al., 2010;Sanchez, Oates, Dacakis, & Holmberg, 2014;Sulter et al., 1994). At its simplest it is used as a diagnostic tool to document changes in the voicing source, but given the importance of the speaking fundamental frequency in trans populations, it is very useful to monitor the voice changing as a result of testosterone therapy. The lowest available frequency (also known as the Minimum F0) is elicited by asking the client to glide or scale down as low as they can and sustain the lowest note. The highest available frequency is usually elicited by asking the client to glide or scale up as high as they can, with no regard to voice quality. Musically untrained clients usually find it easiest if they imagine they are on a roller coaster ride and they imitate the sharp rises and deep falls of the roller coaster with their voice. It is up to the speech-language therapist to record and auditorily evaluate the client's production during the assessment/elicitation session in order to get the true extremes of the phonatory range. Then the results of the acoustic measurement should be expressed in a psychoacoustic scale (e.g., semitones, ERB) and it could be recorded as one of the indicators of vocal change.
The correct interpretation of the MPR results is crucial in giving good feedback to transgender men. Results in Damrose (2009) and Adler, Constansis, and van Borsel (2012) seem to document a marked change in the maximum phonation range (MPR) of transgender men transitioning on testosterone. They conclude that as a result of testosterone therapy the three transgender men's MPR has decreased when expressed in Hz. Building on their work, Papp (2011a) converted their original results into semitones and found that, by the end of the documented vocal transitioning, the MPR of Damrose's subject shrank by 1.3 ST, but the MPR of the two subjects in Adler et al.'s study not only did not shrink but in fact expanded by a trivial 0.3 ST. A net change in the range of 1 ST over 12 months is a perceptually insignificant change that is well within the ranges of documented time-of-day fluctuations (Lehto, Laaksonen, Vilkman, & Alku, 2008;van Mersbergen, Verdolini, & Titze, 1999), longitudinal fluctuations over a month (Gelfer, 1989), and on par with within-day versus across-day differences (Murry, Brown, & Morris, 1995). As such, it indicates no perceptual narrowing of the available pitch range. On the contrary, the consistency of the size of the ranges is quite remarkable considering the time span of over one year between the first and last measurements. This reanalysis of the findings of previous studies indicates only a relocation of the entire pitch range to a lower fundamental frequency interval.
The results in Papp (2011a) also lend further support to the temporariness of the range reduction: transitioning did indeed lead to loss of range (as measured in semitones) around 150 to 200 days into the testosterone therapy, but by the end of the first year on testosterone most subjects not only regained but actually largely surpassed their original range. As pointed out in Adler, Constansis, and van Borsel (2012), this "bottleneck" is due to the steadily lowering pitch floor, which is concurrent with a temporary lack of access to the upper range.
Papp's overview of the changes of the pitch descriptors during the first year of testosterone therapy shows that the sentential mean fundamental frequency decreased substantially in some subjects. On average the 1-year SFF means are about 3 to 11 ST lower and the SFF minima about 6 to 12 ST lower than the pretestosterone values. SFF maxima changed differently among subjects. Some speakers lost about 8 to 9 ST from their upper range, while others gained 4 to 10 ST in the upper range by the end of their first year on testosterone. In Papp's study, two speakers experienced a slight decline in pitch range (3-4 ST), while the others experienced that their pitch range widened over the first year on testosterone by more than an octave (12C ST). A hitherto unexamined finding in Papp (2011a) is that the location of the mean SFF value within the MPR also changed. Speakers' mean SFF before transitioning was higher than the predictable 3 to 4 ST above the pitch floor. This distance grew in the case of every transgender man in this study. In the case of two speakers, this means that they spoke a full octave (12 ST) above their pitch floor. This "buoyancy" of the mean pitch over the pitch floor is observable in the subjects in the study by Adler, Constansis, and van Borsel (2012) as well, whose mean SFF was about 7 to 8 ST over their pitch floor.
Introspection from Papp's study participants seems to raise the point that the speed, but not necessarily the magnitude, of the mean SFF decrease might correlate with the amount or ratio of hormones in the body. Subjects who took larger amounts of testosterone, or whose estrogen production was (virtually) nonexistent (cf. postmenopausal and postoophorectomy transgender men) felt they demonstrated a faster, more hyperbolic change. This, however is not borne out (yet) by empirical evidence: Nygren (2014) found no or only weak correlations between hormone levels and fundamental frequency values over the first 18 months of the testosterone therapy.

Surgical Lowering of the Fundamental Frequency
While it is not common, surgical intervention is available to lower the speaking fundamental frequency using techniques developed by Isshiki, Taira, and Tanabe (1983). The vocal fold relaxation technique has been used in cases of mutational falsetto and hypogonadic males as well as transgender men (Hoffman et al., 2014;Kocak et al., 2008;Remacle, Matar, Verduyckt, & Lawson, 2010).

Vocal Tract Resonances
The growth of the vocal tract length in biological males is usually accounted for by the growth spurt during adolescence and the descent of the larynx. However, a change in transgender men's height or laryngeal descent is not reported in the studies so far. This suggests that no physiological (as opposed to behaviorally conditioned) changes are to be expected as a result of the androgen therapy. However, the results of Papp's longitudinal study (2011) reveal previously undocumented lowering in the vowel formant values over the first year of testosterone therapy. The first three formant values, which had been previously not considered a factor during the vocal transitioning, were significantly lower during the first year of hormone therapy. This trend was lent further support by Zimman (2012).
There are a few physiological changes in the bony or cartilaginous structures around the vocal tract in transgender men. Antoszewski, Zadzinska, and Foczpanski (2009) found that the size of FtM transsexuals' maxillary and mandibular canines and the first molars is between those of cismales and cisfemales.
Informal introspection by transgender men in Papp (2011a) indicates that the angle between the posterior part of the mandible and the lower border of the ramus changes noticeably during transitioning, creating a more masculine jawline. Some transgender men experience the growth of the bony protrusion of an Adam's apple, which may be accompanied with other laryngeal cartilage shifts and growths. All these phenomena indicate that these structures in and around the vocal tract may remain sensitive to androgens comparatively late in life.
Papp's findings based on two subjects followed for multiple years strongly suggest that besides physiological forces, behavioral modifications may also lie behind the observed formant changes. This is evidenced by steady progress toward more-masculine formant values, followed by the reversal of the values a number of years into the androgen therapy, much like the inverse J-curve observed in the SFF of these men. This increased ability to make behavioral choices, which is facilitated by the physiologically altered structures, provides further evidence of the relative flexibility of speech features that were previously thought of as largely fixed. The slow fluctuation and sometimes reversal of the values further support the idea that speakers continually fine-tune features as they repeatedly relocate themselves along the gender continuum.

Binders
Transgender men who have gone through hormonal adolescence as females often wear a compressing layer of clothing in order to create a more masculine looking chest. The binders may range from ACE bandage and (multiple, layered) sports bras to tight tops and specially-made binders covering the whole torso. Introspection among transgender men holds that the strong chest compression that especially larger-chested transgender men need to employ often leads to shortness of breath, inadequate breath support for voicing and physical exercise, and, in especially severe cases, fainting and cracked ribs.
While wearing binders is a non-negotiable part of many no-surgery or pre-surgery transgender men, it might be useful to help clients monitor their pulmonary/spirometric values (such as descriptors of the (forced) inspiratory and expiratory vital capacity maneuvers, and total lung capacity) by comparing in-binder and no-binder tidal breathing, especially at the beginning of the social transitioning, which often coincides with the onset of habitually wearing a binder. As proper pulmonary support is necessary for good laryngeal valving efficiency; also phonation intensity should be checked in transgender men-that is, whether they can project in a sustained manner.

Hormones
Hormone therapy might affect ventilatory responses to obstructive lung problems (such as asthma, chronic obstructive pulmonary disease (COPD), bronchitis), as testosterone, progesterone and estrogen are known to alter ventilatory control. Moreover, gender differences exist in lung function and metabolic demand, both of which may affect the ventilatory compensatory responses to airway obstruction. We recommend surveying and monitoring a history of pulmonary or ventilatory problems in clients on hormones.

Sleep-Disordered Breathing
A lack of consistent results is found in studies of ventilatory responses in men and women. However, there are a number of observed risk factors that were not well understood prior to the publication of population-based studies that included both genders. Although the gender differences are less striking than once thought, there still appears to be a higher prevalence of both sleep-disordered breathing (SDB) and obstructive sleep apnea (OSA) in men than in (premenopausal) women. OSA is a relatively common condition often characterized by snoring, recurrent episodes of cessation of breathing, disrupted sleep, and daytime somnolence. Associated serious complications are hypertension, increased risk of heart disease, stroke, and increased susceptibility to industrial and motor vehicle accidents (Park, Ramar, & Olson, 2011;Ramar & Guilleminault, 2007).
The reason for the male predominance remains uncertain. Because testosterone contributes to fat deposition in the neck and upper body and reshapes the mandible and possibly the pharyngeal cavity, it may contribute to the development of OSA through reduced or altered upper airway size. In support of the hypothesis connecting testosterone with OSA is the increased prevalence, induction, or worsening of OSA in these groups: postmenopausal women Shahar et al., 2003;Young, Finn, Austin, & Peterson, 2003); women with high endogenous levels of testosterone and/or polycystic ovarian syndrome (PCOS) (Dexter & Dovre, 1998;Fogel et al., 2001;Vgontzas et al., 2001); women with exogenous testosterone administered (Johnson, Anch, & Remmers, 1984);and hypogonadal, elderly, or obese men through administration of testosterone (Hoyos, Killick, Yee, Grunstein, & Liu, 2012;Killick et al., 2013;Liu et al., 2003;Matsumoto et al., 1985;Sandblom et al., 1983;Schneider et al., 1985).
Most of these studies controlled for the effects of age and weight/BMI, as both have been found to show positive correlation with the prevalence of OSA. For documented and possible risk factors of OSA see Ramar and Guilleminault (2007); for further testosteroneinduced risk factors consult the WPATH Standards of Care E. Coleman et al. (2011).
Because OSA may be worsened or unmasked by androgen therapy in FtM individuals, and since it may be a risk factor for a number of further health problems, we recommend that the speech-language therapist should regularly ask the clients about changes in quality of sleep and snoring.

SUMMARY AND OUTLOOK
As shown above, during the process of vocal masculinization, hormonal and psychosocial forces often work in tandem in order to create a voice that represents not only the true gender but the true complete identity of transgender men, which includes expressions of age, ethnicity, sexual orientation, and fluidity in and around gender. As such, voice and speech goals may change during transitioning, whether the transgender man is taking testosterone or not. In order to gauge the goals and the vocal satisfaction, it is recommended to survey the clients before, during, and after the therapy provided by the speech and language therapist. The survey should preferably include (among other information) client expectations, vocational demands on voice, current satisfaction with voice and communication, age, menopause/ ooporectomy status, amount of testosterone taken and method of administration, current SFF descriptors, use of chest binders, snoring and sleep-disordered breathing, and if the client is comfortable with sharing such information, the current gender identity and/or sexual orientation. This information, taken together with culturally appropriate sociolinguistic usage will help the speech-language therapist recommend pitch, formant, breathiness, and creakiness goals that give a voice to the person within.
While transitioning on androgens is very frequent among FtMs, it is not the only way to reshape the voice and speech by any means. Many no-hormone or prehormone transgender men may seek out a speech-language therapist's help in order to reshape their voice and speech. According to Papp's large-scale survey (Papp, 2011b), a large proportion of transgender men have started smoking or have considered starting smoking in order to lower their SFF. While evidence suggests (e.g., Gilbert & Weismer, 1974) that smoking might indeed lower the SFF, speech-language therapist should promote alternative, sustainable, nondamaging ways to reshape a transgender man's voice.
Finally, as suggested above, there is much to learn about the perception of the transmale voice, both accompanied by or in absence of visual clues. Observations by transgender men, such as "I sound weirdly nasal," "People keep asking if I have a cold," "People say I sound like a young gay man, but I'm straight and in my 40s," indicate outstanding issues. These issues were quantifiable in Nygren's (2014) dissertation: She found that in her pool of 50, for nearly a quarter of transgender men, voice and communication therapy was needed due to problems with vocal fatigue and instability, insufficient lowering of pitch and projecting power, and problems with the voice sounding too young.
These issues might include the source-filter mismatch in some transgender men possibly caused by the combination of cismalelike SFF and cisfemale-like resonances; altered resonances possibly due to changing bone and cartilage structures; or temporary dysphonia during the first year on testosterone while the SFF, much like in cismale adolescence, is still in flux.
All these further emphasize the importance of (gender(ed)) sensitivity, good training in speech-language therapists, and continued focused research in order to better serve this invisible, but sizeable, population.

SUMMARY OF RECOMMENDATIONS
Voice and Communication Intervention for Gender Nonconforming Individuals 1. Transgender voice and communication services should be offered in the context of a complete approach to transgender health that includes comprehensive primary care and a coordinated approach to psychological and social issues. 2. In working with gender nonconforming clients, the speech-language therapist's primary goal is to help the client develop voice and communication that more closely approximates the client's sense of self. 3. Feminizing/masculinizing voice involves nonhabitual use of the voice-producing mechanism. To prevent the possibility of vocal damage, professional evaluation and assistance are essential. 4. Self-guided voice and communication change without professional supervision is not recommended. Clients intending to pursue self-guided voice change should be encouraged to, at a minimum, have an initial professional assessment and then to consult with their primary care provider if they develop symptoms of vocal fatigue or negative changes to vocal quality. Self-help voice and communication groups should have appropriate clinical support.

Clinical Competence
1. Voice and communication professionals working with transgender individuals must have a basic understanding of transgender health (including hormonal and surgical feminization/masculinization) and trans-specific psychosocial issues; they must be familiar with basic sensitivity protocols such as use of preferred gender pronoun and name. 2. Gender nonconforming individuals who are seeking voice and communication services for reasons other than speech feminization/masculinization can be treated by trans-sensitive speech-language therapists using standard voice and communication protocols. voice and communication feminization/ masculinization requires additional clinical expertise and special clinical protocols.

Treatment Decisions
1. The client is responsible for treatment decisions, supported by the clinician's informed professional opinion, assessment data, and any allies the client wishes to be involved. 2. To support fully informed treatment decisions, clients should be informed of the following: a. potential risks and benefits associated with treatment options b. estimated duration of treatment; factors that can influence the outcome or duration of treatment 3. Existing protocols for voice and communication feminization should be reviewed and considered when developing individualized treatment plans. As there are no established protocols for speech masculinization, FtMs seeking this service should be informed that the protocol is a trial. 4. While modification of existing protocols is encouraged, all treatment plans (including those using new or experimental techniques) are expected to be based on a clearly articulated, logical, and valid clinical rationale. Departure from existing protocols should be explained as such to the client as part of fully informed consent and should be documented in detail to facilitate evaluation. Assessment 1. Assessment prior to voice and communication feminization/masculinization should include the following: a. psychosocial, voice use, voice health, and medical history b. clinical assessment of speech and voice including: 1. the client's subjective assessment 2. instrumental measurement 3. the clinician's subjective analysis 4. an assessment of potential for change 5. assistance with understanding therapeutic options 2. As there is evidence that behavioral changes (of pitch, inflection, resonance, etc.) may degrade over time, periodic reevaluation is recommended following treatment with further clinical assistance as needed.
Voice and Communication Therapy 1. Voice and communication therapy should be individualized based on each person's goals and identity, the risks and benefits of treatment options, and consideration of social and economic issues. 2. Rather than adopting a rigid and artificial set of voice and communication norms, it is recommended that clients be assisted to develop an individualized and contextspecific set of norms based on communication patterns in their own social, cultural, work, and home environments. 3. It is clinically optimal to be able to offer both individual sessions and group treatment, with the proportion of time in each format depending on the client's therapeutic needs and goals.
Pitch-Elevating Surgery 1. As there is no professional consensus regarding the effectiveness and riskbenefit ratio of pitch-elevating surgery, care should be taken to ensure that clients are fully informed of potential risks, postoperative care requirements, and possible outcomes (including decreased pitch). 2. Assessment by both a laryngologist and speech-language therapist is recommended prior to surgery. 3. Prior to surgery, the laryngologist should discuss after-care instructions with the patient and provide written after-care instructions. 4. Voice therapy should be offered following phonosurgery to help the patient adapt to and stabilize the new voice.
Outcome Evaluation 1. Outcomes should be rigorously evaluated and documented. 2. At minimum, the baseline assessment should be repeated immediately following the end of therapy. Ideally, re-evaluation would take place at 6 months, 1 year, 5 years, and 10 years after treatment. 3. Evaluation should include client satisfaction with the treatment outcome and with the quality of care provided, as well as perceptual and objective measures of voice and communication change. 4. Informal or formal sharing of outcome data (with colleagues, at conferences, in publications, etc.) must occur only if the client has provided fully informed and voluntary written consent.

Research
1. There is a paucity of data relating to speech feminization/masculinization. Further research in this area is eagerly anticipated. 2. To ensure that participation in research is voluntary, services should not be offered solely as part of a research protocol.

NOTES
1. For the rest of the document we will use speechlanguage therapist (SLT) and clinician as umbrella terms for speech-language therapists and pathologists. Other speech and voice specialists, such as singing teachers and speaking voice trainers may also be involved in discrete areas of the voice and communication needs of the transgender population.