Call for Participants: Audio Quality Evaluation for Generative Deep Neural Models
Luca Comanducci
Hi everyone!
We would like to invite you to participate in a listening test to help us evaluate the audio quality of different generative deep neural models. The study aims to compare how well these models generate realistic audio foley sounds, such as footsteps, dog barks, and rain. During the test, you will listen to audio samples generated by advanced deep learning models and rate their quality and temporal fidelity compared to reference recorded sounds.
The experiment will last about 10 minutes, and no personal data is required. You can complete the test online at your convenience, but we suggest using headphones and being in a quiet environment for a better evaluation experience. Your feedback will be invaluable in helping us assess the performance of these generative models.
The listening test is available at the following link: http://51.21.36.86:5000/. If you have any questions or need further information, please do not hesitate to contact marcofuri...@mail.polimi.it.
Thank you for your participation!
Best regards,
51.21.36.86
webMUSHRA by International Audio Laboratories Erlangen
Luca Comanducci
Marco Furio Colombo
Francesca Ronchini
Image and Sound Processing Group
Politecnico di Milano
Via Ponzio 34/5
20133, Milano, Italy