Inhalt des Dokuments
(Möller , 4 SWS/6 LP, jeweils im WiSe)
LV-Nummer: 0434 L 900
VL Time: Mondays, 10.00-12.00, starts 23.10
VL Room: Auditorium 3, Geb. TEL 20. Etage
UE Time: Tuesdays, 14.00-16.00, starts 23.10, Mittag 
UE Room: PC-Pool TEL 106re
Please note that the type of examination of the module "Speech Signal Processing & Speech Technology" has changed from oral to Portfolio.
You have to register through ISIS to participate in the course. Information on how to register will be given in the first lecture.
Speech signals and speech sounds; human speech production; auditory perception; speech signal transmission and coding; speech recognition and speaker recognition; speech synthesis; spoken dialogue systems; multimodal dialogue systems.
Speech is the most important means of human communication, and more
and more it develops into an important modality for human-computer
interaction. Already systems work by speech recognition,
interpretation of linguistic content, control of dialoge flow,
generation of responses or production of speech signals. Beyond
that, the efficient transmission of speech is of utmost importance,
both in conventional transmission networks as well as in networks with
paket switching (eg. Voice over IP).
In the course of this lecture the basis for unterstanding and designing communication technology systems based on speech will be provided. Starting with the production and perception of natural human speech will shed light on many important characteristics of speech signals and requirements for their processing. Essential means for representing speech signals in the time and frequency domains will be laid out. On this basis, the functioning of important components of systems of speech technology will be explicated. Apart from efficient coding of speech, speech recognition, speech synthesis, as well as interaction withspeech processing systems (spoken dialogue systems, alternative term: voice user interfaces) will be central. Finally, improvement strategies for the smoother adaption of such systems to human communicative needs via multimodal means of input and output will be presented (multimodal dialogue systems).
The lecture has been developed with a focus on students of electrical engineering, computer engineering, as well as computer science. Above these, students from linguistics, communication sciences, engineering acoustics, sociology, human factors, as well as other departments are very welcome. Previous knowledge in speech signal processing or linguistics is not required.