Descriptive analysis of processed speech quality was carried out by semantic differentiation and external preference mapping was used to relate the attributes to overall quality judgements. Clean and noisy speech samples from different speakers were processed by various processing chains representing mobile communications, resulting in a total of 170 samples. The perceptual characteristics of the test samples were described by 18 screened subjects and the final descriptive language with 21 attributes were developed in panel discussions. The scaled attributes were mapped to overall quality evaluations collected from 30 screened subjects by partial least square regression. The Phase II ideal point modelling was used to predict the quality with an average error of about 6 % and to study the linearity of the attributes.
(See document for exact affiliation information.)
Click to purchase paper as a non-member 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 Join the AES. If you need to check your member status, login to the Member Portal.
