PESQ (Perceptual Evaluation of Speech Quality) is an objective metric that predicts human perception of speech quality by comparing original and processed signals.
What is an example of PESQ?
A clean reference recording and its enhanced version are scored using PESQ to evaluate how close the enhancement is to the original.
How does PESQ work?
PESQ models human auditory perception, analyzing distortions and timing differences to produce a score typically between 1 (bad) and 4.5 (excellent).
How does ai-coustics use PESQ?
Here at ai-coustics, we use PESQ to objectively measure and compare the quality of enhanced speech produced by our models.
