Telephone Transmission Impact on Synthesized Speech: Quality Assessment and Prediction

ISSN： 1610-1928

Source： Acta Acustica united with Acustica, Vol.90, Iss.1, 2004-01, pp. : 121-136

Disclaimer: Any content in publications that violate the sovereignty, the constitution or regulations of the PRC is not accepted or approved by CNPIEC.

Previous Menu Next

Abstract

Modern telephone networks introduce a number of perceptively diverse degradations which impact the overall quality of transmitted speech. In this paper the question is addressed whether the overall amount of degradation is similar for synthesized compared to naturally produced speech. Prototypical speech samples are first impaired by different degradations in a controlled way, using a transmission simulation model. The samples are then judged upon by test subjects in an application-oriented listening-only scenario. It turns out that noise-type degradations exercise about the same quality impact on naturally produced and synthesized speech. On the other hand, the impact of low bit-rate codecs is different for the two types of stimuli. This finding is explained with an "artificiality" dimension which can be seen as an additional impact for naturally produced speech, and which is already present in the synthesized speech samples.