With use of standard audiometric procedures, the hearing thresholds of patients with known diabetic retinopathy did not differ significantly from those of a control population group. However, use of a more subtle psychoacoustic tool--a filtered speech task--showed a definite difference in hearing acuity between the two groups.