Skip to main content
. 2024 Jul 4;67(11):4176–4185. doi: 10.1044/2024_JSLHR-24-00045

Figure 3.

The image displays the effect size in terms of cohen\u2019s d for 14 speech mode variables. The speech mode can be conversational or read. The description lists the mean value followed by the minimum and maximum values of each variable. 1. Unadapted: 1.2, 0.7, and 2.1. 2. Personalized: 0.9, 0.5, and 1.6. 3. SNR: 0.4, 0, and 1.4. Word Count: 1.7, 1.3, and 2.4. Duration: 2.1, 1.7, and 2.7. TTR: 0.9, 0.6, and 1.4. Severity: 0.2, negative 0.3, and 0.6. Intelligibility: 0.3, negative 0.1, and 0.7. Articulatory: 0.4, 0, 0.9. Phonatory: 0.3, negative 0.2, and 0.7. Resonatory: negative 0.1, negative 0.5, and 0.4. Respiratory: 0.3, negative 0.2, and 0.7. Speaking Rate: negative 0.5, negative 1.1, and negative 0.1. RMS: negative 1.3, negative 1.8, and negative 0.7.

Effect sizes analysis of variables thought to impact automatic speech recognition accuracy. SNR = signal-to-noise ratio; TTR = type–token ratio; RMS = root-mean-square of signal amplitude.