The intelligibility of pointillistic speech

Area of Science:

• Auditory Perception
• Signal Processing
• Speech Science

Background:

• Auditory perception research often requires highly controlled stimuli.
• Existing speech stimuli can be susceptible to degradation, limiting experimental control.
• A need exists for robust speech stimuli for multitalker listening studies.

Purpose of the Study:

• To describe a novel processed speech method.
• To evaluate the discriminability and robustness of this processed speech.
• To explore its potential applications in auditory research.

Main Methods:

• Speech was processed into synchronous sinusoidal pulses across frequency.
• Experiments varied frequency channel parameters and pulse duration.
• A condition involved random removal of processed speech components.

Main Results:

• The processed speech was highly discriminable in closed-set identification.
• The processed speech exhibited significant resilience to random component removal.
• Parameter variations (channels, duration) influenced discriminability.

Conclusions:

• The described processed speech offers high discriminability and robustness.
• This method provides a controlled stimulus for auditory research.
• It is particularly suitable for multitalker listening experiments requiring stimulus control.