Free electronics ,electrical,civil projects Report Proposal
Download Final-Year Project Report
A Novel Loudspeaker Equaliser
M. R. P. Thomas
download link:below
1 Introduction
The assessment of the ‘quality’ of a loudspeaker is open to a myriad of objective and subjective
measurements. Attempts to accurately and concisely model a loudspeaker as a linear circuit by R. H.
Small in his brilliant set of papers [18] [19] [20] (published in the early 70s) remain the standard by
which engineers characterise their drivers and enclosures.
It is generally accepted that the most important figure-of-merit is the loudspeaker’s frequency
response. This is rarely flat over the frequencies of interest (~50Hz – 20kHz) so equalisers are devised
for Hi-Fi applications to flatten the response of the system. Early equalisers used passive analogue
networks which were inaccurate as well as being prone to variations with temperature and age. More
recent developments have seen the use of DSP to replace the analogue networks, providing tighter
tolerances and no drift.Existing techniques do, however, contain two major flaws. First is the assumption that theloudspeaker’s frequency response is linear. It will be shown analytically and practically that at largevoice-coil excursions the driving force and damping are no longer proportional to input current.Second is the almost universal use of swept-sine methods to characterise frequency responses. Withswept sines, energy is concentrated over a very narrow bandwidth. It has been shown in [3] thatfrequency responses derived from spot tones can differ from those where the excitation energy isspread over a wider spectrum. Given that the energy in music and speech is rarely concentrated over a
small bandwidth, analysis employing spot tones should be avoided. Swept sines also don’t provide
phase information, but it can be shown that the phase response (and therefore the group delay) is not
necessarily linear and perhaps should be taken into consideration in the design of an equaliser.
The most popular techniques for determining frequency responses of linear systems will be described
in 2.3 and it will be argued that the best method for characterising the nonlinear response of a
loudspeaker is the Maximum Length Sequence (MLS) or M-Sequence Decorrelation technique. We
will discuss the theory and implementation issues with MLS in detail as they are an important tool in
this project.
Download full report
[PDF]A Novel Loudspeaker Equaliser-Project report
Download Final-Year Project Report
A Novel Loudspeaker Equaliser
M. R. P. Thomas
download link:below
1 Introduction
The assessment of the ‘quality’ of a loudspeaker is open to a myriad of objective and subjective
measurements. Attempts to accurately and concisely model a loudspeaker as a linear circuit by R. H.
Small in his brilliant set of papers [18] [19] [20] (published in the early 70s) remain the standard by
which engineers characterise their drivers and enclosures.
It is generally accepted that the most important figure-of-merit is the loudspeaker’s frequency
response. This is rarely flat over the frequencies of interest (~50Hz – 20kHz) so equalisers are devised
for Hi-Fi applications to flatten the response of the system. Early equalisers used passive analogue
networks which were inaccurate as well as being prone to variations with temperature and age. More
recent developments have seen the use of DSP to replace the analogue networks, providing tighter
tolerances and no drift.Existing techniques do, however, contain two major flaws. First is the assumption that theloudspeaker’s frequency response is linear. It will be shown analytically and practically that at largevoice-coil excursions the driving force and damping are no longer proportional to input current.Second is the almost universal use of swept-sine methods to characterise frequency responses. Withswept sines, energy is concentrated over a very narrow bandwidth. It has been shown in [3] thatfrequency responses derived from spot tones can differ from those where the excitation energy isspread over a wider spectrum. Given that the energy in music and speech is rarely concentrated over a
small bandwidth, analysis employing spot tones should be avoided. Swept sines also don’t provide
phase information, but it can be shown that the phase response (and therefore the group delay) is not
necessarily linear and perhaps should be taken into consideration in the design of an equaliser.
The most popular techniques for determining frequency responses of linear systems will be described
in 2.3 and it will be argued that the best method for characterising the nonlinear response of a
loudspeaker is the Maximum Length Sequence (MLS) or M-Sequence Decorrelation technique. We
will discuss the theory and implementation issues with MLS in detail as they are an important tool in
this project.
Download full report
[PDF]A Novel Loudspeaker Equaliser-Project report
No comments:
Post a Comment
its cool