Automatic Evaluation of Musical Sound Separation Quality
This paper addresses the problem of evaluating effectiveness of musical sound separation algorithms. The standardized procedure for evaluating separation quality does not exist. The most convincing and typical way to do this is by carrying out subjective listening tests. However, subjective tests need a solid statistical validation, which means that many experts should take part in such tests, the room characteristics should be adequate, and what is also important, such tests are time consuming. Thus this paper attempts to show that it is possible to carry out the evaluation tests in an automatic way, by employing an Artificial Network System (ANN), which is further justified by experts’ opinion.
Click to purchase paper or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!
This paper costs $20 for non-members, $5 for AES members and is free for E-Library subscribers.