A psychoacoustic approach to dysphonic voice quality perception

描述（由申请人提供）：语音障碍通常会导致患者、临床医生和对话伙伴注意到的语音质量变化，而语音质量的改善是治疗的关键结果。然而，我们对人们如何感知语音质量的了解有限。这限制了我们准确量化或描述质量变化的能力，例如由于疾病或治疗引起的变化。这个延续项目结合了声音科学、言语科学、听力科学和工程学的概念和技术来解决这个问题。一般来说，研究首先在实验室中获得语音质量感知的高精度测量。然后，这些数据被用于开发准确反映听众数据的语音质量感知的数学模型。为了在人对语音质量的判断和模型输出之间获得紧密的匹配，使用听觉处理模型来获得声音信号的内部表示。然后从这种内部听觉表征中捕获特定的测量，并用于对语音质量的感知建模。通过简单的刺激（如“hot”中的元音/a/），完成了对两个不同的语音质量维度（呼吸度和粗糙度）的感知判断、声学信号到其内部表示的转换以及语音质量模型的一般形式的方法。在拟议的工作中，这些方法将进一步发展，以建立一个全面理解语音质量感知的框架，并使其能够转化为临床实践。这些方法将(1)用于解释多个共同发生的语音质量维度；(2)适用于更自然和复杂的刺激（多个元音和音节）；(3)利用了解其他语音质量维度（应变）。(4)为了提高模型的准确性，扩大模型对严重不发音声音（如II型和III型）的适用性，将开发估计不发音声音“音高”和“音高强度”的方法，并将其纳入相关模型。(4)为了以提高临床效用的方式增强测量方案，将模型输出转换为与每个语音质量维度的感知幅度直观相关的尺度。这将创建一组易于使用和解释的直观的语音质量指标。(5)最后，通过初步的临床研究，评估三种模型和指标在常规临床评估中的可行性。

DESCRIPTION (provided by applicant): Voice disorders often lead to changes in voice quality noticed by patients, clinicians, and conversation partners, and improvement in voice quality is a critical outcome of treatment. However, we have limited knowledge of how people perceive voice quality. This has restricted our ability to accurately quantify or describe changes in qualit, such as due to a disease or when resulting from treatment. This continuation project combines concepts and techniques from voice science, speech science, hearing science, and engineering to address this problem. In general, the research proceeds by first obtaining high-precision measures of voice quality perception in the laboratory. These data are then used to develop mathematical models of voice quality perception that accurately reflect listeners' data. To obtain a close match between human judgments of voice quality and model output, models of auditory processing are used to obtain an internal representation of the voice acoustic signal. Specific measures are then captured from this internal auditory representation and used to model the perception of voice quality. Methods for obtaining perceptual judgments of single voice quality dimensions, the transformation of the acoustic signal to its internal representation, and the general form of the voice quality models have been completed for two different voice quality dimensions (breathiness and roughness) using simple stimuli (vowel /a/ as in "hot"). In the proposed work, these approaches will be developed further to establish a framework for comprehensive understanding of voice quality perception and to enable translation to clinical practice. These approaches will be (1) used to account for multiple, co-occurring voice quality dimensions; (2) applied to more natural and complex stimuli (multiple vowels and syllables); and (3) leveraged to understand other voice quality dimensions (strain). (4) To increase model accuracy and to expand their applicability to severely dysphonic voices (e.g. Type II and Type III), methods to estimate the "pitch" and "pitch strength" of dysphonic voices will be developed and incorporated into relevant models. (4) To enhance the measurement schemes in a manner that improves clinical utility, model output will be transformed to a scale that is intuitively relted to the perceptual magnitude of each voice quality dimension. This will create a set of intuitive voice quality metrics that are easy to use and interpret. (5) Finally, the feasibility of using thee models and metrics in regular clinical assessment will be evaluated through an initial clinical study.