Scribd
Upload
185 views4 pages

Measurement of Impulse Responses Using Alternative Methods

This document discusses and compares two techniques for measuring impulse responses: exponentially swept sines and maximum length sequences (MLS). It provides an overview of how each technique works, including the algorithms and assumptions involved. The document also describes an implementation of MATLAB programs for each technique. The results show that both methods produce estimated impulse responses that match the provided impulse responses being tested.

Uploaded by

Mudit Arya
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
185 views4 pages

Measurement of Impulse Responses Using Alternative Methods

This document discusses and compares two techniques for measuring impulse responses: exponentially swept sines and maximum length sequences (MLS). It provides an overview of how each technique works, including the algorithms and assumptions involved. The document also describes an implementation of MATLAB programs for each technique. The results show that both methods produce estimated impulse responses that match the provided impulse responses being tested.

Uploaded by

Mudit Arya
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 4

MEASUREMENT OF IMPULSE RESPONSES USING

ALTERNATIVE METHODS
DSP Lab Course WS 2013/2014
Authors:
Mudit Arya([email protected]),
Anupama Edakkattillam([email protected])
Supervisor: Christian Hofmann
Introduction
Measurement of impulse responses is a common task in audio signal processing. Usual
applications include measurement of speakers and measurement of room impulse responses.
Here two common measurement techniques are reviewed:

Exponentially swept sines


Maximum length sequences(MLS)

The SineSweep technique overcomes such limitations. It is based on the following idea: by using
exponential time growing frequency sweep, it is possible to simultaneously deconvolve the linear
impulse response of the system and to selectively separate each impulse response corresponding
to the harmonic distortion orders considered. The harmonic distortions appear prior to the linear
impulse response. Therefore, the linear impulse response measured is assured exempt from any
non-linearity and, at the same time, the measurement of the harmonic distortion at various orders
can be performed.
MLS uses pseudo-random white noise and the swept sine uses time varying frequency signals.
These methods are based on the assumption of perfect linearity and time-invariance of the
system, and give problems when these assumptions are not met. In particular MLS is quite
delicate; it does not tolerate very well nonlinearity or time-variance, and requires that the
excitation signal is tightly synchronized with the digital sampler employed for recording the
systems response.

Algorithm
SweptSine: First constructing a measurement signal x(n) for which a signal x-1(n) can be easily
determined that is inverse to x(n) in the sense that the convolution,

L1

x( n)
k=0

. x-1(n-k) = C . (n-no)

yields a scaled and time-shifted unit impulse. Then the convolution of the measured signal y(n)
and the inverse signal x-1(n) will give the scaled and time-shifted impulse response.
L1

C . h(n-no) =

y (k ) . x-1(n-k).
k=0

And the exponential sine sweep,


n

x(n) =

log (w 2 /w 1)
w 1.( L1)
sin (
.(e L1
1))
log ( w 2/ w 1)

with an instantaneous frequency increasing exponentially from w1 at n=0 up to w2 at n=L-1.


The inverse signal can be obtained by time reversal and amplitude scaling according to
-1

x (n) = x(L-1-n) .

n
w 2 (L1)
)
w1

which approximates the convolution eq. with no = L-1. The scaling factor can be found to be,
w1
1)
w2
w1
2( w 2w 1) log (
)
w2
. L(

C=

The approximation is in the sense that the convolution of x(n) and x-1(n) results in a unit impulse
band-limited to the range(w1,w2) (See Figure 2).
MLS Technique: This method is based upon the excitation of the acoustical space by a periodic
pseudo-random signal having almost the same stochastic properties as a pure white noise. The
number of samples of one period of an m order MLS signal is: L= 2m 1. With this technique the
impulse response is obtained by circular cross-correlation between the measured output and the
determined input (See Figure 1).

Implementation
A complete MATLAB program was developed in both Sweptsine and MLS technique. And
programs were run and modified when needed. The results from the programs have been
analyzed and saved.

Evaluation Results
Here are our experimental results from the developed program code.

Figure 1

MLS generated signal, provided impulse response, convolution output and estimated
impulse response

One can see from the above figure that the estimated impulse response is approximately same as
the impulse response under test.

Figure 2

Sweptsine input, provided impulse response, convolution output and estimated impulse
response.
One can see from the above figure that the estimated output response is approximately same as
the impulse response under test.

Literature
[1] M.Holters, T. Corbach, and U. Zlzer, "Impulse response measurement techniques and their
applicability in the real world", Proceedings of the 12th International Conference on Digital
Audio Effects (DAFx09),Como, Italy, 2009
[2] A. Farina, Simultaneous measurement of impulse response and distortion with a swept-sine
technique, Dipartimento di Ingegneria Industriale, Universit di Parma

You might also like