Ez egy előző félévben kiírt, archivált téma.
In speech technology, there are numerous models for representing the human speech signals, which are for example the basis of cell phone codecs. A group of speech codecs, the so called sinewave coders are based on the idea that the periodic parts of speech can be split as the sum of a finite number of sine signals having different frequency and phase. The task of the student is to learn about available sine coders (e,g, HMPD, https://github.com/covarep/covarep/tree/master/vocoder/hmpd), and to extend them with the following: error free handling of creaky voice, continuous modeling of intonation, appropriate modeling of voiced and unvoiced sounds, and potential applications in speech synthesis. This topic is recommended for students with scientific interest. The topic contains theoretical challenges, which are suitable for TDK conference and other competitions. The area is suitable for both computer science and electrical engineering students, having practical challenges as well. English knowledge is required!