Guias baseados em evidências sobre análise acústica de voz para fonoaudiólogos
Traditional acoustic measures ask explicit questions. AI models ask implicit ones—learned from data, encoded in training. As AI enters clinical voice assessment, understanding what question a model is actually asking becomes the central challenge.
F0 explains only 42% of perceived gender—resonance, intonation, and speech patterns account for the rest. Learn evidence-based pitch targets (180 Hz threshold), why formants require 20% elevation, hormone effects, WPATH SOC-8 guidelines, and how to track progress with acoustic measures.
Traditional perturbation measures aren't obsolete—but they have critical limitations. Learn when jitter and shimmer are valid, when they fail (severe dysphonia), why ASHA now recommends CPP as primary, and how Praat vs MDVP give different values.
Can you trust voice recordings from a smartphone? The Bridge2AI-Voice Consortium and 2020-2025 validation studies reveal which parameters work (F0, CPP) and which don't (shimmer, HNR). Learn the evidence-based protocols that make smartphone recordings clinically valid.
Single acoustic parameters can't capture the full complexity of voice quality. Learn how AVQI and ABI combine multiple measures into validated scores that correlate with perceptual ratings, plus validated cutoffs for Brazilian Portuguese.
CPP has emerged as the most reliable acoustic measure for detecting dysphonia. Learn why researchers and clinicians are moving beyond traditional jitter/shimmer measurements, and how to interpret CPP values in your clinical practice.
How to interpret wideband and narrowband spectrograms for clinical voice assessment. Patterns, pathologies, and practical applications.
Maximum Phonation Time and Dysphonia Severity Index explained. Recording protocols, normative data, and practical interpretation for clinical practice.
Estamos sempre buscando temas que ajudem fonoaudiólogos em sua prática clínica.
Sugerir um tema →Dr. Jorge C. Lucero
Professor de Ciência da Computação, Universidade de Brasília
Dr. Lucero tem mais de 30 anos de experiência pesquisando produção de voz e dinâmica de pregas vocais. Publicou mais de 50 artigos revisados por pares sobre ciência da voz e desenvolveu o SimuVox, uma ferramenta de síntese de voz baseada em física. Seu objetivo com o PhonaLab é tornar a análise profissional de voz acessível a todos os profissionais.
Conteúdo Educacional: Estas guias são fornecidas para fins educacionais e apoio à prática clínica baseada-em-evidências. O conteúdo não se destina a ser aconselhamento médico. Decisões clínicas devem ser tomadas por profissionais de saúde qualificados com base na avaliação individual do paciente.