Speech quality in telecommunication networks - sound demos

The sound files given in the table below demonstrate effects of telecommunication networks on speech quality. For two languages, Dutch and German, nine speech samples with different kinds and degrees of quality degradations are presented. The samples are classified into three groups of perceived overall quality, according to results of formal listening tests which were carried out within the HEARCOM project. "MOS" stands for Mean Opinion Score and denotes the mean quality rating over all listeners. For quality rating, a 5-point category scale was used, which ranged from 1 = bad quality to 5 = excellent quality. Intermediate categories are 2 = poor quality, 3 = fair quality, and 4 = good quality. The MOS values given in this table were obtained from normal hearing listeners.

  Dutch German
MOS around 3.5 Listen Listen Listen Listen Listen Listen
MOS around 2.5 Listen Listen Listen Listen Listen Listen
MOS around 1.5 Listen Listen Listen

Listen Listen Listen