Return to search

Extraction of Loudspeaker- and room impulse responses under overlapping conditions

A loudspeaker is often considered to be a Linear Time Invariant (LTI) system, which can be completely categorized by its impulse response. What sets loudspeakers apart from other LTI-systems is the acoustical aspect including echoes, which makes it a lot harder to take accurate noise free measurements compared to other LTI-systems such as a simple RC circuit. There are two main challenges regarding loudspeaker measurement, the first is high frequency reflections of surrounding surfaces and the second is low frequency modal resonances in the room stemming from the initial echoes. A straightforward way of dealing with this issue is simply truncating the measured impulse response before the arrival of the first high frequency reflection. This is however not without its problems as this will result in high uncertainty for low frequency content of the measurement. The longer time until the first reflection is measured, the better the measurement. The ideal measurement would be a noise free environment with infinite distance towards the nearest reflective surface. This is of course not possible in practice, but this ideal environment can be simulated by using an anechoic chamber. This thesis investigates the possibility of creating pseudo anechoic measurements in a general room using optimization with information extracted from measurement data in combination with linear time-varying (LTV) filtering. Algorithms for extracting information such as time delay between reflections as well as compensation for distortion in the reflections have been developed. This information is later used to minimize a cost function in order to obtain an estimation of the loudspeakers' impulse response using multiple measurements. The resulting estimation is then filtered using the LTV filter in order to obtain the pseudo anechoic impulse response. This thesis investigates two different loudspeakers in two ordinary rooms as well as in an anechoic chamber, and evaluates the performance of the developed methods. The overall results seem promising, but due to some inconsistencies of the measurements taken in the anechoic chamber that changes the direct wave of the loudspeakers, the developed methods are unable to achieve a true anechoic impulse response.  It is concluded that to be able to achieve true pseudo anechoic results, measurements in rooms must better resemble the ones taken inside the anechoic chamber. This in combination with tuning the hyper parameters of the LTV filter looks promising to achieve pseudo anechoic impulse responses with high correlation to the true anechoic measurements.

Identiferoai:union.ndltd.org:UPSALLA1/oai:DiVA.org:uu-479971
Date January 2022
CreatorsGustafsson, Felix
PublisherUppsala universitet, Signaler och system
Source SetsDiVA Archive at Upsalla University
LanguageEnglish
Detected LanguageEnglish
TypeStudent thesis, info:eu-repo/semantics/bachelorThesis, text
Formatapplication/pdf
Rightsinfo:eu-repo/semantics/openAccess
RelationUPTEC E, 1654-7616 ; 22007

Page generated in 0.0027 seconds