Correlation of Instrumental Voice Evaluation with Perceptual Voice Analysis Using a Modified Visual Analog Scale

Abstract

Various rating scales have been used for perceptual voice analysis including ordinal (ORD) scales and visual analog (VA) scales. The purpose of this study was to determine the most suitable scale for studies using perceptual voice analysis as a gold standard for validation of objective analysis protocols. The study was carried out on 74 female voice samples from 68 dysphonic patients and 6 controls. A panel of 4 raters with experience in perceptual analysis was asked to score voices according to the G component (overall quality) of the GRBAS system. Two rating scales were used. The first was a conventional 4-point ORD scale. The second was a modified VA (mVA) scale obtained by transforming the VA scale into an ORD scale using a weighted conversion scheme. Objective voice evaluation was performed using the EVA^® workstation. Objective measurements included acoustic, aerodynamic, and physiologic parameters as well as parameters based on nonlinear mathematics (e.g., Lyapunov coefficient). Instrumental measurements were compared with results of perceptual analysis using either the conventional ORD scale or mVA scale. Results demonstrate that correlation between perceptual and objective voice judgments is better using a mVA scale than a conventional ORD scale (concordance, 88 vs. 64%). Data also indicate that the mVA scale described herein improves the correlation between objective and perceptual voice analysis.

Reliability and Validity of the Italian Version of the Consensus Auditory-Perceptual Evaluation of Voice (CAPE-V)

“…The CAPE-V also includes two unlabeled scales; the clinician may use them to rate additional prominent attributes required to describe a given voice. Furthermore, the CAPE-V uses a visual analog scale labeled with visual markers for ‘mild', ‘moderate' and ‘severe' under each line, instead of a four-point equal-appearing interval scale [12,13]. Moreover, the CAPE-V prescribes specific tasks (vowel prolongations, specific sentences designed to elicit various laryngeal behaviors and running speech).…”

The Effect of the Speech Task Characteristics on Perceptual Judgment of Mild to Moderate Dysphonia: A Methodological Study

“…These observations contrast with methodological studies of de Krom [10] and Revis et al [11] who did not report any discrepancy in the evaluations of the perceptual voice quality according to the task. These divergent results observed among these studies could be the consequence of discrepancies in the composition of their voice samples (especially the severity of the dysphonia) since it has been demonstrated that the assessment of patients with higher dysphonia scores may reduce the variability of the score severity related to the task [12, 13]. Among the trials using sustained vowels, some authors removed the onset/attack of the signal while others assessed the perceptual voice quality on the entire signal.…”

“…Among the trials using sustained vowels, some authors removed the onset/attack of the signal while others assessed the perceptual voice quality on the entire signal. Regarding the fact that the voice onset may contain important data for the perceptual assessment [12, 28], removing of the voice onset could reduce the severity of the perceptual grade attributed to the voice sample. In this, paper, although we removed the onset of the sustained vowel, we show higher scores of GRBASI components when based on sustained vowel than when based on connected speech.…”

“…These voice samples can be used in another perceptual tool such as the Consensus Auditory-Perceptual Evaluation of Voice, which includes different tasks, i.e., vowels, connected speech, and conversation [9]. The sustained vowel has the advantage of being easier to elicit and still being unaffected by the articulation of the subject [10, 11] but this production model is unnatural and not representative of the daily speaking voice [12]. In contrast, continuous speech is more natural but some speaking characteristics (articulation, speech rate) may substantially influence the assessment of the voice quality [13].…”

Auditory-Perceptual Voice Characteristics in Pre-Operative Patients Undergoing Thyroid or Parathyroid Surgery

“…Auditory perceptual rating scales both describe vocal quality and quantify the severity of voice by assessing specific perceptual attributes heard in the auditory signal [16]. Limitations of auditory perceptual analysis include subjectivity of ratings, poor inter- and intrarater reliability, the influence of listener experience, and the unstable internal standards that may vary over time for comparison of samples [8,17]. Despite the fact that auditory perceptual analysis lacks perfect validity or reliability, it is still a key standard method for making judgements about vocal impairments [18].…”