Abschlussarbeit

Zurück zur Liste

Adaptive Sekunkärfeldanalyse für die aktive Geräuschunterdrückung

AutorInnen Ureta Staackmann, A.
Jahr 2012
Art der Arbeit Toningenieur-Projekt
Themenfeld Audiosignalverarbeitung
Abstract Headphones with ANC (Active Noise Control) have at least one microphone for each side, left and right respectively, delivering information about the undesired signal to be eliminated. An anti-noise signal can be generated (with an inversion of the disturbing signal's phase) in order to inhibit as much as possible the disturbing signal. From the superposition law, it follows a cancellation of the disturbing signal by means of summation of both the latter and the anti-noise signal generated by the adaptive filter. The whole system is controlled by the LMS algorithm that will adapt the filter coefficients according to the disturbing signal and its incidence angle amongst others. This algorithm may require, depending on the structure used, different reference signals. The reference signal describing the disturbing signal can be provided by two additional external microphones, one for each side. This signal can also be estimated in case no external microphones are available. The reference error signal provided by the inner microphones is on the other hand absolutely necessary in order to have feedback about the residual noise, I.e. the remains of the disturbing signal after the noise cancellation process. These reference signals must possess the same phase (time displacement) in order for the algorithm to work properly. They must hence be delayed according to the latency of the system which is the time needed for the AD and DA conversions, respectively. Also the transmission path between the loudspeaker and the listener's ear, the so called secondary path, will vary every time the listener puts on the headphones and with it the group delay as well. Compensating the influence of the secondary path represents an- other challenge for ANC systems. Therefore it is necessary to artificially delay the reference signals corresponding to the given situation. Whereas the latency of a system does not change over time, the group delay of the secondary path may constantly vary affecting directly the system's performance. In this project, the effects of a constant changing secondary path as well as the estimation accurateness needed for it on a ANC system are to be studied. Moreover an adaptive identification method of the secondary path is to be developed. In addition to the known adaptive methods for a secondary path analysis, a new approach relying on a data base of measured secondary paths is to be implemented.
URL http://phaidra.kug.ac.at/o:11452
BetreuerInnen Flock, M.