School of Health Sciences

Speech Acoustics & Perception Research Lab

Research lab in Speech Acoustics & Perception is housed in the Department of Speech and Hearing at Cleveland State University. This laboratory comprises of a team of speech-language pathologists and speech scientists who conduct research on basic and applied speech and hearing issues.

What do we do?
Research projects involve cross-language issues in speech production and perception, with a current emphasis on foreign-accentedness

Research Objectives

Research in this laboratory investigates cross-language aspects of speech production and perception. Current projects address foreign-accented English with the goal to determine factors that relate to the production and perception of “accentedness”.

Methodology in this laboratory mainly comprises of speech acoustic analyses, categorical perception tasks, lexical-decision paradigms, and lab-induced speech training.

The larger goals of research in this laboratory aim to increase theoretical understanding of speech perception, as well as yield practical variables for application in speech-pathology and ESL teaching programs.

Instrumentation and Methodology

High-fidelity speech recording
Participants' speech is recorded in controlled sound-treated environment using high-fidelity recording devices;to be later digitized, edited, analyzed and/or played to listeners.

Directional and omnidirectional microphones; DAT recorders/CD recorders.

Digitization and editing of recorded samples
The recorded sample is converted from analog to digital form in order to allow further processing of the signals. Additionally, the recordings are “cleaned up”, i.e. removal of unnecessary fillers, repetitions and the recorded chunks are split into unfinished sound files.

DAT/CD recorders, connected to computers with sound editing programs such as Sound Forge, PRAAT, Cool Edit, Wave, etc.

Acoustic Analysis
The individual digitized sound files are now analyzed for various acoustic parameters depending on the experiments' rationale (e.g. duration, frequency, amplitude, formant history analysis, LPC analysis, etc.)

Software programs such as CSL, Multispeech, PRAAT, BLISS, etc.
high fidelity speech recording

Acoustic Manipulation
Depending on the experiments' rationale, the individual digitized sound files could also be manipulated or edited in different ways i.e. smearing, adding noise, splicing segments, changing acoustic parameters of duration, frequency, amplitude, sampling rate, etc.
Software programs such as CSL, Multispeech, PRAAT, BLISS, etc.

Speech Perception Tasks
Subjects participate in a variety of listening experiments e.g. perceptual identification and discrimination, rating, lexical-decision, priming, etc.
Software programs such as SuperLab and BLISS