Synthetic Speech Quality Evaluation (Hearing-Impaired Listeners)

Clinical Trial

Offered by: Nemours Children's
Location: Delaware Valley

What is the trial about?

The purpose of the study is to investigate how hearing-impaired listeners perceive synthetic speech of various qualities. Results from this study will be used to assess the feasibility of using text-to-speech "voices" in the design of software tools for aural rehabilitation (AR) applications.

Who can participate?

Hearing-impaired adults over 18 years of age who have bilateral cochlear implants or a unilateral implant with non-functional hearing in the contralateral ear can be in the study. All participants must be native speakers of American English. Participants must also sign a release form authorizing the researchers at Nemours to request their hearing history and audiometric information from their doctor.

What is involved?

One 1-hour listening session.

After signing a consent form, we will ask you to sign a release form authorizing the researchers at Nemours to request your hearing history and audiometric information from your doctor. Unilaterally implanted participants will be screened for eligibility (no useful hearing in their non-implanted ear). All eligible participants will then receive instructions on the task and a short practice session to make sure they understand the task. For the practice and full listening task, you will sit in a sound-dampened room and listen to short semantically unpredictable (nonsense) sentences over loudspeakers. The sentences will be produced with synthetic (computer-generated) speech. Some of the sentences may be difficult to understand. You will hear each sentence only once. After each sentence, you will have as much time as you need to type the sentence as you heard it. When you are finished typing, you will go to the “Next” button to hear the next sentence. For each sentence, we will measure the difference between what you type and what the sentence was supposed to be. The practice session will consist of 10 sentences. For the full listening task, you will hear 100 sentences. The session will take about one hour to complete.