|Anforderungen||Android programming skills|
In a telephone service the speech quality is affected by various degradations. These are often caused by properties of the communication channel, such as missing packets or distortions caused by a low bitrate codec. To measure the performance of a telephone service, it is important to know how the quality was perceived by the service user. Traditionally, to measure the quality of transmitted speech, subjective experiments with naive participants are conducted. Since subjective methods require a significant effort to prepare, instrumental signal-based models have been established. These models make a prediction of the perceived quality by a comparison of the reference and the degraded speech signal.
The goal of the thesis is to develop a system that allows for measuring the speech quality of mobile voice networks. To realise this, on the sender side an answering station, with which a speech sample can be send from an Android smartphone through the voice channel, needs to be developed. On the receiver side, an app that automatically calls the answering station and records the voice channel is required.
The theses may be written in German or in English.