Background: Patients with facial paralysis often report frustration with communication; however, there are limited data regarding intelligibility deficiencies. Objective: To compare speech intelligibility in patients with severe and non-severe facial paralysis, and in patients with or without synkinesis. Methods: Video and audio data were reviewed retrospectively. Groups were stratified as follows: Group A - severe paralysis (Sunnybrook 0-20) without synkinesis; Group B - non-severe (Sunnybrook >20) paralysis without synkinesis; and Group C - non-severe paralysis with synkinesis. Intelligibility was assessed by lay-people and a speech and language pathologist (SLP) using the Frenchay Dysarthria Assessment Version 2 (FDA-2). A receiver operating characteristic (ROC) curve was used to determine a Sunnybrook cutoff for intelligibility. Results: Eighty cases were reviewed with mean age 55.6, 53.8% female. 25.0% were in Group A, 30.0% Group B, and 45.0% in Group C. Lay-people rated 15.0% and the SLP rated 28.7% as having intelligibility deficiency. An ROC curve demonstrated that patients with Sunnybrook ≤18.5 were more likely to have intelligibility abnormality. Conclusion: Patients with Sunnybrook ≤18.5 are more likely to demonstrate intelligibility deficiency. Clinicians with a more trained ear are more likely to identify intelligibility abnormality compared with lay-people. Those with synkinesis are more intelligible compared with those without it.