Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

PESQ between noisy and enhanced version of speech. #12

Closed
BilalDendani opened this issue Nov 3, 2019 · 2 comments
Closed

PESQ between noisy and enhanced version of speech. #12

BilalDendani opened this issue Nov 3, 2019 · 2 comments

Comments

@BilalDendani
Copy link

Hello everyone,
If I do not have a clean version of speech .
Can I compute the PESQ between the noisy original speech and the enhanced speech?
what does that mean ?
Best Regards.

@vBaiCai
Copy link
Owner

vBaiCai commented Nov 3, 2019

If you don't have a reference signal (clean speech), you cannot calculate the PESQ. Of course, you can calculate the PESQ between noisy and enhanced signals. But, I think it is some type of undefined behavior.

@BilalDendani
Copy link
Author

Thank you for your answer.
Consider if we have just the noisy data set. We want apply some techniques of speech enhancement. We know that the PESQ measure of speech quality corresponds to the correlation between reference original clean version and the degraded version. In my case, I do not have clean data set. Can I consider to compute PESQ measure as the value of correlation between original noisy and the enhanced version to deduce if my Speech Enhancement technique does improve the speech quality.

@vBaiCai vBaiCai closed this as completed Dec 28, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants