What is SIGMOS?
SIGMOS (Signal Mean Opinion Score) is a subjective quality metric where human listeners rate the clarity and naturalness of audio signals. Like other MOS-style metrics, it measures perceptual quality rather than how useful it is to machines.
What is an example of SIGMOS?
A panel of listeners scores enhanced speech samples on a 1–5 scale, with higher scores indicating better quality.
How does SIGMOS work?
It relies on human evaluation rather than algorithms, making it a gold standard for assessing perceptual audio quality.
How does ai-coustics use SIGMOS?
At ai-coustics, we use SIGMOS to evaluate perceptual enhancement models like Rook, where the goal is to sound clear and natural to human listeners. For machine-oriented models like our Quail family which is tuned for ASR and voice agent reliability, we use other metrics, like Word Error Rate.
