Este trabalho apresenta uma análise completa da qualidade de voz esperada em serviços de VoIP quando prestados sobre a plataforma WLAN IEEE 802.11g. Diferentemente da maior parte da literatura, este estudo complementa as informações de atraso, variação do atraso e taxa de perda dos pacotes, com um modelo analítico de avaliação de qualidade de voz proposto pela ITU-T, o Modelo E. Desta forma, todos os fatores modelados como perceptíveis pelos usuários do serviço telefônico são considerados simultaneamente, gerando, por fim, resultados correspondentes à escala de avaliação subjetiva da qualidade de voz MOS, que qualifica a opinião média em: “muito satisfeitos”, “satisfeitos”, “alguns insatisfeitos”, “muitos insatisfeitos”, “quase todos insatisfeitos” e “não recomendado”. O código do simulador NS-2 foi expandido para prover todas as funcionalidades requeridas neste trabalho. Com ele, diferentes cenários de aplicação foram simulados utilizando três dos codificadores de voz mais utilizados atualmente, o G.723.1, o G.729a e o G.711 (com e sem PLC). A funcionalidade centralizada da camada MAC IEEE 802.11, o PCF, também foi estudada, tendo os parâmetros referentes ao período entre repetições do CFP e à máxima duração do mesmo analisados e ajustados de modo a prover o melhor desempenho. Experimentos similares foram realizados para a função DCF, permitindo então uma comparação direta entre os dois modos de operação do nível MAC. Além disso, ensaios considerando outros tráfegos menos prioritários (estações de dados) foram realizados com o intuito de caracterizar a capacidade dos sistemas em diferenciar ambos os tráfegos. / This work is intended to provide a complete evaluation of the voice quality obtained in IEEE 802.11g WLAN environments. In contrast to the major literature, this dissertation complements the traditional network metrics, as delay, jitter and loss rate, with an analytical model for voice quality evaluation created by ITU-T, named E Model. With this model, all factors able of affecting the user’s perception of quality in a telephony service are simultaneously considered to reflect the subjective voice quality scale MOS, which rates the mean opinion as follow: “very satisfied”, “satisfied”, “some user satisfied”, “many users dissatisfied”, “nearly all user dissatisfied” and, “not recommended”. An extension to the NS-2 simulator source code has been developed to provide all the features required by this study. With this ‘customized simulator’, several scenarios have been analyzed considering three of the most important vocoders in the market today, G.723.1, G.729a, and G.711 (with and without PLC). Moreover, the centralized scheme for the IEEE 802.11 MAC, called PCF, has been studied to show the effect of the CFP repetition interval and the CFP maximum duration on its upper bound limit of simultaneous voice calls. Similar experiments have been carried out for the DCF MAC functionality, allowing these medium access modes to have their performance directly compared against each other. Finally, scenarios with low priority traffic (data traffic) have been created to help describing the ability of both MAC functions to deal with traffic differentiation.

Tal och röst styrs av komplicerade processer, där vitt skilda funktioner används i dess produktion och perception. Hur tal och röst uppfattas skiljer sig på flera sätt från den objektiva mätbara akustiska talsignalen. Grundtonens, F0’s, styrka och det man kallar talarformanten är två parametrar som påverkar hur rösten låter. Studien hade som syfte att undersöka om grundtonens styrka och med eller utan talarformant påverkar vilka stimuli otränade röstbedömare tycker bäst om. För att undersöka detta skapades ett syntetiskt /a/ efter formantfrekvenser för svenska vokaler (Fant, 1959), som varierades med utan eller med talarformant och med sju olika grundtonsstyrkor, totalt fjorton stimuli. Tolv kvinnor och åtta män, fick lyssna på /a/:na och skatta dem från det /a/ som de gillade mest till det /a/ som de gillade minst. Studiens frågeställningar bekräftades, variationer av grundtonens styrka och med eller utan talarformant påverkade vilka stimuli otränade röstbedömare föredrog, och män och kvinnor föredrog olika stimuli. De två /a/:na med svagast och det /a/ med starkast grundton, gillades minst av hela gruppen. De fyra /a/:na däremellan tyckte informanterna om ungefär lika mycket. Studerade man var för sig vad män och kvinnor skattade blev resultatet annorlunda. Kvinnor föredrog stimuli med talarformant mer än de utan talarformant, medan män föredrog stimuli med stark grundton. Ytterligare studier kunde vara intressanta att göra, för att undersöka om stimulus egenskaper med avseende på talarformant och grundtonsstyrka har någon effekt på hur tal uppfattas och tolkas. / Speech and voice is governed by complicated processes, where a variety of different functions are used in its production and perception. Speech and voice differs in many ways from the objective measurable acoustic speech signal. The intensity of the fundamental frequency, F0, and what is called the speaker formant are two parameters that affect how voice is perceived. The purpose of the study was to investigate if the varied intensity of F0 and with or without speaker formant affect what stimuli untrained voice assessors like the most. To do so, a synthetic /a/ was created, formed by formant frequencies for Swedish vowels (Fant, 1959), varied with with or without speaker formant, and seven different intensities of the fundamental frequency, creating a total of 14 stimuli. Twelve women and eight men were asked to listen to and grade the stimuli, from the /a/ that was liked the most to the one that was liked the least. The expectations of the study were confirmed, the variation of the intensity of F0 and with or without speaker formant, did affect which stimuli the listeners preferred, and also confirmed that men and women preferred different stimuli. The two /a/’s with the weakest and the /a/ with the strongest intensity of F0, were least liked. The four /a/s with varying intensity of F0 in between, were liked approximately the same. If men and women ratings were studied separately from one another, the result differed. Women preferred stimuli with speaker formant more than the ones without, whereas men more appreciated stimuli with strong intensity of F0. If the study was the be retested and the results would be confirmed, further tests would be of interest to do, to investigate if the differences of preferences affect how speech is perceived and interpreted.

This research had like objective to investigate the vocal and laryngeal effect and sensations that appeared from to the execution of the voiced tongue vibration technique (VTVT) applied in three series of fifteen repetitions, in the maximum time of phonation with habitual tone and intensity, with pauses of thirty seconds of passive rest between each serie. For that, had taken part in the study 24 people, of the female sex, ranging from 25 to 41 years old, without vocal complaints. All those persons were submitted to a perceptive-auditive and acoustic of voice analysis, through the Multi-Dimensional Voice Program (MDVP) - Model 5105 e Multi- Speech - Model 3700, of Kay Elemetrics; and a valuation of the larynx, by means of video laryngostroboscopic exam, before and after the execution of VTVT. The results of the appreciation were analyzed of quantitative and qualitative form, by means of statistic examinations, being compared with the existent bibliography about the subject. The analysis of the results had become evident that it was statistically significant, after the execution of the VTVT: the improvement of the kind of voice; of the focus of vertical resonance; of the vocal quality; the increase of the fundamental frequency; the improvement of the parameters of the spectrographic evaluation, in the Large Band and Narrow Band filters; the maintenance of the parameters of the laryngeal images (glottal closure, constriction of the laryngeal vestibule, amplitude and symmetry of the vibration of the vocal folds); the betterment of the medial constriction of the vestibule, according to the increase of time of the execution of the VTVT; and the predominance of the positive sensations;. The data of this study also had permitted to conclude, that the VTVT presents modifications about the glottal source and about the resonant filter / Esta pesquisa teve como objetivo investigar o impacto vocal e laríngeo e as sensações surgidas frente à execução da técnica de vibração sonorizada de língua (TVSL), aplicada em três séries de quinze repetições, em tempo máximo de fonação com tom e intensidade habituais, com intervalos de 30 segundos de repouso passivo entre cada série. Para isso, participaram do estudo 24 sujeitos, do sexo feminino, com média de idade de 25,41 anos, sem queixas vocais. Todos indivíduos foram submetidos à análise perceptivo-auditiva e acústica da voz, por meio dos programas Multi-Dimensional Voice Programa (MDVP) - Model 5105 e Multi-Speech - Model 3700, da Kay Elemetrics e avaliação da laringe, por meio do exame de videolaringoestroboscopia, antes e após a execução da TVSL. Os resultados das avaliações foram analisados de forma quantitativa e qualitativa, por meio de testes estatísticos, sendo comparados com a bibliografia existente sobre o assunto. A análise dos resultados evidenciou que foi estatisticamente significativo, após a execução da TVSL: a melhora do tipo de voz; do foco de ressonância vertical; da qualidade vocal; o aumento da freqüência fundamental; a melhora de parâmetros da avaliação espectrográfica, em filtros de Banda Larga e Banda Estreita; a manutenção dos parâmetros das imagens laríngeas (fechamento glótico, constrição do vestíbulo laríngeo, amplitude e simetria de vibração das pregas vocais); a melhora da constrição medial do vestíbulo, conforme o aumento do tempo de execução da TVSL; e o predomínio de sensações positivas. Os dados desta pesquisa permitem concluir, que a TVSL apresenta modificações sobre a fonte glótica e sobre o filtro ressonantal

A avaliação perceptivo-auditiva tem papel fundamental no estudo e na avaliação da voz, no entanto, por ser subjetiva está sujeita a imprecisões e variações. Por outro lado, a análise acústica permite a reprodutibilidade de resultados, porém precisa ser aprimorada, pois não analisa com precisão vozes com disfonias mais intensas e com ondas caóticas. Assim, elaborar medidas que proporcionem conhecimentos confiáveis em relação à função vocal resulta de uma necessidade antiga dentro desta linha de pesquisa e atuação clínica. Neste contexto, o uso da inteligência artificial, como as redes neurais artificiais, indica ser uma abordagem promissora. Objetivo: Validar um sistema automático utilizando redes neurais artificiais para a avaliação de vozes rugosas e soprosas. Materiais e métodos: Foram selecionadas 150 vozes, desde neutras até com presença em grau intenso de rugosidade e/ou soprosidade, do banco de dados da Clínica de Fonoaudiologia da Faculdade de Odontologia de Bauru (FOB/USP). Dessas vozes, 23 foram excluídas por não responderem aos critérios de inclusão na amostra, assim utilizaram-se 123 vozes. Procedimentos: avaliação perceptivo-auditiva pela escala visual analógica de 100 mm e pela escala numérica de quatro pontos; extração de características do sinal de voz por meio da Transformada Wavelet Packet e dos parâmetros acústicos: jitter, shimmer, amplitude da derivada e amplitude do pitch; e validação do classificador por meio da parametrização, treino, teste e avaliação das redes neurais artificiais. Resultados: Na avaliação perceptivo-auditiva encontrou-se, por meio do teste Coeficiente de Correlação Intraclasse (CCI), concordâncias inter e intrajuiz excelentes, com p = 0,85 na concordância interjuízes e p variando de 0,87 a 0,93 nas concordâncias intrajuiz. Em relação ao desempenho da rede neural artificial, na discriminação da soprosidade e da rugosidade e dos seus respectivos graus, encontrou-se o melhor desempenho para a soprosidade no subconjunto composto pelo jitter, amplitude do pitch e frequência fundamental, no qual obteve-se taxa de acerto de 74%, concordância excelente com a avaliação perceptivo-auditiva da escala visual analógica (0,80 no CCI) e erro médio de 9 mm. Para a rugosidade, o melhor subconjunto foi composto pela Transformada Wavelet Packet com 1 nível de decomposição, jitter, shimmer, amplitude do pitch e frequência fundamental, no qual obteve-se 73% de acerto, concordância excelente (0,84 no CCI), e erro médio de 10 mm. Conclusão: O uso da inteligência artificial baseado em redes neurais artificiais na identificação, e graduação da rugosidade e da soprosidade, apresentou confiabilidade excelente (CCI > 0,80), com resultados semelhantes a concordância interjuízes. Dessa forma, a rede neural artificial revela-se como uma metodologia promissora de avaliação vocal, tendo sua maior vantagem a objetividade na avaliação. / The auditory-perceptual evaluation is fundamental in the study and analysis of voice. This evaluation, however, is subjective and tends to be imprecise and variable. On the other hand, acoustic analysis allows reproducing results, although these results must be refined since the analysis is not precise enough for intense dysphonia or chaotic waves. Therefore, the will to develop measurements allowing reliable knowledge related to vocal function is not new on this research and clinical actuation field. In this context, the use of artificial intelligence such as neural networks seems to be a promising research field. Objective: to validate an automatic system using artificial neural networks for evaluation of vocal roughness and breathiness. Methods: One hundred fifty (150) voices were selected from from Clínica de Fonoaudiologia da Faculdade de Odontologia de Bauru (FOB/USP) database. These voices presented variation from neutral to intense roughness and/or breathiness. Twenty-three of them were excluded since they did not match inclusion criteria. Thus, 123 voices were used for analysis. The procedures include use of auditoryperception based on two scales: visual analog scale of 100 mm and four points numerical scale. Additionally, the characteristics of voice signals were extracted by Wavelet Packet Transform and by analysis of acoustic parameters: jitter, shimmer, derivative amplitude and pitch amplitude. Validation of classifying system was carried out by parameterization, training, test and evaluation of artificial neural networks. Results: In the auditory-perceptual evaluation, excellent interrater (p=0.85) and intrarater (0.87<p<0.93) agreement were obtained by means of Intraclass Correlation Coefficient (ICC) testing. The artificial neural network performance has achieved the best results for breathiness in the subset composed by parameters jitter, pitch amplitude and fundamental frequency. In this case, the neural network obtained a rate of 74%, demonstrating excellent concordance with auditory-perceptual evaluation for visual analog scale (0.80 ICC) and mean error of 9 mm. As for roughness evaluation, the best subset is composed by Wavelet Packet Transform with 1 resolution level, jitter, shimmer, pitch amplitude and fundamental frequency. For this case, a 73% rate was achieved (0.84 ICC) and mean error of 10 mm was obtained. Conclusion: The use of artificial neural networks for roughness and breathiness evaluation present high reliability (ICC&gt;0.80), with results similar to interrater agreement. Thus, the artificial neural network reveals a promising method for vocal evaluation, bringing objective analysis as a strong advantage.

<p>Predmet doktorske disertacije je utvrđivanje stanja glasa budućih eltinih vokalnih profesionalaca (studenata glume), upoređivanje sa stanjem glasa studenata nevokalnih<br />zanimanja, utvrđivanje metodičkih postupaka za razvoj glasa, kao i utvrđivanje uticaja glasovnih vežbi primenom &quot;WIC&quot; tehnike na stanje glasa kod ispitanika kojima je potvrđen vokalni poremećaj.</p> / <p>This PhD thesis aims to establish the state of voice in future elite vocal professionals (drama students), to compare it with the state of voice in students of non-vocal professions, to determine the methodic procedures for the voice development, as well as to ascertain the effects of vocal exercises by application of &ldquo;WIC&rdquo; technique on the state of voice in respondents who were found to suffer from the vocal disorder.</p>

Real-time voice measurements were performed to assess whether there are significant problems with 802.11b wireless networks regarding real-time voice communication. We present an analysis of how the 802.11b MAC protocol and diverse environmental conditions affect the quality of real-time voice in terms of loss, delay, and jitter. We also reveal practical issues of wireless monitoring with passive sniffers for this type of analysis. The results obtained in our measurements show that in the majority of the experiments the quality was good, but under some circumstances the requirements for an acceptable voice communication were not met. / Realtidsröstmätningar gjordes för att testa om det finns problem med 802.11b trådlösa nätverk beträffande realtidsröstkommunikation. En analys presenteras av hur 802.11b MACs protokoll och olika tillstånd i omgivningen påverkar kvaliteten på realtidsrösten i form av förluster, fördröjningar och jitter. Även praktiska angelägenheter om trådlös övervakning med passiva sniffers visas. De erhållna resultaten visar att i en majoritet av fallen var kvaliteten acceptabel, men under vissa förhållanden blev inte kraven för röstkommunikation uppfyllda.

LTE (Long Term Evolution) a été développé et normalisé par le 3GPP (3rd Generation Partnership Project). C’est un réseau à commutation de paquets. Cela signifie que la voix sur LTE (VoLTE) est un service de VoIP avec les exigences de qualité de service garantis au lieu de transmettre dans un réseau à commutation de circuits tels que les systèmes existants (2G/3G). VoLTE est déployé dans un réseau entièrement IP combinée avec IMS (IP Sous-système Multimédia). De ce fait, le déploiement de VoLTE est assez complexe et comment assurer la qualité de transmission de la voix sur les réseaux LTE est un très grand défi. Ainsi, il faut plusieurs solutions différentes pour renforcer et améliorer la qualité de transmission de la voix sur les réseaux LTE. Dans cette thèse, nous présentons des solutions en vue d’améliorer la qualité de transmission de la voix sur les réseaux LTE pour les services audio à bandes étroites et larges. Pour cela, il nous faudra différents facteurs complets en solutions. L’un d’eux est QoE (Qualité de l’Expérience) qui est une nouvelle tendance. Et afin de déterminer la perception des utilisateurs pour le service en temps réel tel que VoLTE, nous utilisons le E-model étendu et le WB (large bande) E-model pour des services audio à bandes étroites et larges respectivement. Les solutions proposées ici portent principalement sur des éléments clés dans les réseaux LTE, tels que le codage par chaine, MAC (Contrôle d’Accès Moyen) des systèmes de planification et la qualité de voix du moniteur décrits comme suit. Tout d’abord, des algorithmes améliorés pour renforcer le codec de la chaine LTE (codeur et décodeur) ont été proposés. Pour améliorer le codeur de chaine LTE, un algorithme d’adaptation conjointe a été déployé. Le but de cet algorithme est de minimiser la redondance générée par codage en chaine avec une légère réduction de la perception de l’utilisateur. Ensuite, afin d’améliorer le décodeur par chaine LTE, un algorithme amélioré Log-MAP a été présenté. Cet algorithme vise à obtenir la performance BER (Bit Error Rate) qui est le plus proche du Log-MAP avec une complexité de calcul réduite par rapport à l’état de l’art. Deuxièmement, la chaine et les systèmes QoS de planification améliorés de la perception de l’utilisateur et du mode de priorité VoIP ont été proposés. Ces planificateurs sont déployés à la fois pour les utilisateurs d’audio à larges et à étroites bandes. Les résultats numériques montrent qu’ils surpassent plusieurs planificateurs en vedette tels que FLS, M-LWDF et EXP/PF en termes de retard, de taux de perte de paquets, de débit cellulaire, d’indice et de l’équité et d’efficacité spectrale dans presque tous les cas. Enfin, pour assurer la qualité vocale de transmission sur le réseau LTE, la prédiction de la satisfaction des utilisateurs est essentielle. Pour cette raison, nous présentons deux modèles non intrusifs pour mesurer la qualité de la voix sur les réseaux LTE. Ces modèles sont utilisés pour les utilisateurs d’audio à bandes étroites et larges bandes. Les modèles proposés ne se réfèrent pas au signal original. Par conséquent, ils sont très appropriés pour prédire la qualité de l’appel vocal sur les réseaux LTE / LTE (Long Term Evolution) has been developed and standardized by 3GPP (3rd Generation Partnership Project). It is a packet-switched network. This means voice over LTE (VoLTE) is a VoIP service with the guaranteed QoS requirements instead of transmitted in a circuit-switched network such as the legacy system (2G/3G). Since VoLTE is deployed in an All-IP network combined with IMS (IP Multimedia Subsytem), thus, the VoLTE deployment is quite complex and how to ensure voice transmission quality over LTE networks is a very big challenge. Therefore, there needs to be many different solutions to enhance and improve voice transmission quality over LTE networks. In this dissertation, we present solutions to enhance and improve voice transmission quality over LTE networks for both narrowband and wideband audio services. In order to do that, there needs to be many different factors complemented in solutions. One of them is QoE (Quality of Experience) which is a new trend. And in order to determine user perception for real-time service such as VoLTE, we use extended E-model and WB (Wideband) E-model for narrowband and wideband audio services, respectively. The proposed solutions in this thesis mainly focus on key elements in LTE networks such as channel coding, MAC (Medium Access Control) scheduling schemes and monitor voice quality described as follows. First, enhanced/improved algorithms for enhancing LTE channel codec (coder and decoder) have been proposed. In order to enhance LTE channel coder, a joint source-channel code rate adaption algorithm has been deployed. The goal of this algorithm is to minimize redundancy generated by channel coding with a slight reduction of user perception. Next, in order to enhance LTE channel decoder, an improved Log-MAP algorithm has been presented. This algorithm aims at obtaining BER performance that is closest to the LOP-MAP with the computational complexity reduced in comparison with state-of-the-art. Second, channel- and QoS-Aware scheduling schemes with the enhancement of user perception and VoIP priority mode have been proposed. These schedulers are deployed for both narrowband and wideband audio users. The numerical results show that they outperform several featured schedulers such as FLS, M-LWDF, and EXP/PF in terms of delay, packet loss rate, cell throughput, fairness index, and spectral efficiency in almost cases. Last, in order to ensure voice transmission quality over LTE network, prediction of user satisfaction is essential. For this reason, we present two object non-intrusive models for measuring voice quality in LTE networks. These models are used for narrowband and wideband audio users. The proposed models do not refer to the original signal, thus, they are very suitable for predicting voice call quality in LTE networks

This study investigates how body position affects the relative contribution of the rib cage to speech breathing and voice quality. The frequently used upright and supine body positions are expanded with the addition of an inverted position as a condition. Five participants performed two speech production tasks across the three body positional conditions. The estimation of the relative contribution of the rib cage to speech breathing was obtained through respiratory inductance plethysmography (RIP). Phonation type was estimated through the usage of the α-ratio, obtained with a neck-surface accelerometer. The results between the participants suggest a pattern regarding levels of the relative contribution of the rib cage in relation to phonation type, which offers opportunities for further investigation in future research. / Den här studien undersöker hur kroppsposition påverkar det relativa bidraget av bröstkorgen för talandning och röstkvalité. Den frekventa användningen av upprät och supin kroppsposition utvidgas med en inverterad position som villkor. Fem forskningspersoner genomförde två talproduktionsuppgifter genom de tre kroppspositionsvillkoren. Uppskattningen av det relativa bidraget av bröstkorgen för talandning var erhållen genom respiratory inductance plethysmography (RIP). Fonationstyp var uppskattad genom användningen av α-ratio, erhållen med en ytligt hals-placerad accelerometer. Resultaten mellan deltagarna föreslår ett mönster för nivåer av det relativa bidraget av bröstkorgen relaterat till fonationstyp, som ger möjligheter för vidare undersökning i framtida forskning.

This thesis explores the use of the Estill Voice Model, in particular six voice qualities (Speech, Twang, Falsetto, Cry/Sob, Belt and Opera) and their permutations, to define character, character traits and emotions. Traits and emotions that specific voice qualities can influence are, but are not limited to, location, age, background, socioeconomic status, genre, intelligence, nationality, class, culture, gender, promiscuity, disposition, pain and revelations. In particular, this thesis explores the use of voice qualities to show specific human qualities of the character "George" from Sondheim and Lapine's "Sunday in the Park with George" and the people he imitates in his painting by letting the characters' given circumstances (textual and subtextual), the way other actors portray the characters and the director's and musical director's input inform the choices in voice quality. By using the specific technical aspects of the Estill Voice Training System and combining them with the limi

INTRODUÇÃO: Os resultados do aumento da frequência fundamental após diminuição do comprimento da porção vibratória das pregas vocais pela realização de sinéquia na região anterior da glote são variáveis, e não é estabelecida a extensão ideal de tal sinéquia nas técnicas cirúrgicas para aumentar a frequência vocal. Comparou-se neste estudo a elevação da frequência fundamental do som produzido por laringes humanas excisadas de cadáveres após realização de diferentes extensões de sinéquia glótica anterior, correspondentes a ¼ (25%), 1/3 (33%) e ½ (50%) do comprimento da prega vocal. Além disso, verificou-se a correlação entre as frequências fundamentais anterior e posteriores aos experimentos com tais diferentes extensões de sinéquia glótica anterior para formular modelos matemáticos que estimem a frequência fundamental obtida após o procedimento. Os padrões de vibração das pregas vocais foram avaliados qualitativamente por videolaringoscopia e videoquimografia digital com câmera de alta velocidade. MÉTODOS: Realizou-se estudo experimental com 21 laringes humanas excisadas de cadáveres masculinos adultos. O som glótico foi artificialmente produzido com a colocação de ar comprimido pela traqueia, passando pelas pregas vocais. A frequência fundamental e os parâmetros da videoquimografia digital foram obtidos previamente (controle) e após realização de pontos que mimetizaram uma sinéquia glótica anterior de ¼, 1/3 e ½ do comprimento das pregas vocais em todas as laringes, simulando uma das técnicas cirúrgicas para aumento da frequência vocal, a glotoplastia de Wendler. RESULTADOS: A frequência fundamental média foi distinta entre as diferentes extensões de sinéquia glótica anterior (P < 0,001), e observou-se aumento progressivo, significativo a cada ampliação da extensão da sinéquia na glote (P < 0,05), tanto em hertz quanto em semitons. Os modelos de regressão que estimam a frequência fundamental após os procedimentos mostraram melhor coeficiente de determinação (r2) para extensões menores de sinéquia. Não se verificou aperiodicidade das vibrações em nenhum experimento, nem se constatou padrão de mudança de fechamento glótico ou de simetrias de fase e amplitude. CONCLUSÕES: A frequência fundamental do som aumentou com diferença significativa a cada ampliação da extensão da sinéquia anterior na glote, sem alterações significativas nos parâmetros qualitativos da onda mucosa. Esses resultados sugerem que a medida da extensão da sinéquia glótica é relevante para o resultado final na elevação da frequência fundamental, e os dados apresentados podem ser uma referência inicial para o prognóstico do resultado cirúrgico / INTRODUCTION: After anterior glottic web formation to decrease the length of the vibrating portion of the vocal folds, the degree of increase in fundamental frequency is variable. The ideal extent of such a web in surgical approaches aimed at raising the vocal pitch has not been established. This study compared the increase in the fundamental frequency of sound produced by excised human larynges after the formation of anterior glottic webs of varying extents (25%, 33%, and 50% of the length the vocal fold). In addition, the correlation between the fundamental frequencies before and after the experiments with the various extents of anterior glottic web formation was verified in order to create mathematical models designed to estimate the change in fundamental frequency obtained after the procedure. The vibration patterns of the vocal folds were evaluated qualitatively by laryngoscopy and by kymography with a highspeed digital camera. METHOD: We conducted an experimental study on 21 normal larynges excised from adult male human cadavers. Laryngeal vibration was artificially produced and was recorded by high speed videoendoscopy and digital kymography for analysis. The fundamental frequency and digital videokymography parameters were obtained before (as a control) and after the formation of anterior glottic webs occupying 25%, 33%, and 50% of the vocal folds, simulating a surgical technique to increase the vocal frequency, Wendler glottoplasty. RESULTS: The mean fundamental frequency differed among the various extents of anterior glottic web formation (P < 0.001), and there were significant increases from one extent to the next (P < 0.05), in hertz as well as in semitones. Regression models to estimate post-procedure fundamental frequency showed better coefficients of determination (r2) for smaller web extents. There was no aperiodic vibration in any of the experiments. The changes in glottal closure/phase and amplitude symmetry did not follow any particular pattern. CONCLUSIONS: The fundamental frequency increased significantly at each increase in the extent of the anterior glottic web, without significant changes in the qualitative parameters of vocal fold vibration. These results suggest that the extent of the glottic web influences the final result, in terms of the magnitude of the increase in the fundamental frequency, and the data presented here could serve as an initial reference for the postoperative prognosis