The 7 Objective Metrics We Conducted for the Reconstruction and Resynthesis Tasks | HackerNoon
Briefly

from Hackernoon7 months ago
"For VC, we used two subjective metrics: naturalness mean opinion score (nMOS) and voice similarity MOS (sMOS) with a CI of 95%; and three objective metrics for naturalness: UTMOS, character error rate (CER) and word error rate (WER)."
Hackernoonhttps://hackernoon.com/the-7-objective-metrics-we-conducted-for-the-reconstruction-and-resynthesis-tasks
"We conducted seven objective metrics for reconstruction and resynthesis tasks: a log-scale Mel error distance, perceptual evaluation of speech quality (PESQ), and others, showcasing a comprehensive evaluation framework for our models."
Hackernoonhttps://hackernoon.com/the-7-objective-metrics-we-conducted-for-the-reconstruction-and-resynthesis-tasks
Read at Hackernoon
#speech-synthesis #voice-conversion #evaluation-metrics #neural-models #zero-shot-learning
[
|
]