Home | Contact Us | FAQ | Search & Site Map | Link to Us
Sign In | Join | Other 45 Sites in Network
Home
Discussion Groups
Biology
BiologyBotanyMicrobiologyEntomologyEvolutionPaleontology
Chemistry
General ChemistryAnalytical ChemistryElectrochemistryOrganic Synthesis
Earth Science
GeologyMineralogyOceanographyMeteorologyEarthquakes
Physics
General PhysicsResearchRelativityParticle PhysicsElectromagnetismFusionOpticsAcousticsNew Theories

Natural Science Forum / Physics / Acoustics / March 2006



Tip: Looking for answers? Try searching our database.

room impulse response and polynomial facoring

Thread view: 
Enable EMail Alerts  Start New Thread
Thread rating: 
jeremyscerri@gmail.com - 05 Mar 2006 20:54 GMT
i am doing some research on acoustics and particularly the nature of
room reverberation. this led me to some harsh number crunching using
matlab.
i am working with polynomials having hundreds of terms and very high
powers in the range of 800 to 2000.
i am particularly interested in the position of the roots relative to
the unit circle. Thus i am using roots() a matlab funtion to compute
the roots of the polynomial. This seems to be working fine (altough i
presume i have rounding errors).
Now my intention is to position all roots inside the unti circle by
computing the reciprocal of the roots outside and then recompute the
coefficients of the new polynomial using the poly() function
This however is leading to a false result. The functions roots() and
poly(), seem to work well when i tested them with small polynomials as
shown below:

polycoeff =

   1.0000    0.6000    0.4000    0.2000    0.1000

>> polyroots=roots(polycoeff)

polyroots =

  0.1340 + 0.5544i
  0.1340 - 0.5544i
 -0.4340 + 0.3450i
 -0.4340 - 0.3450i

>> originalpoly=poly(polyroots)

originalpoly =

   1.0000    0.6000    0.4000    0.2000    0.1000

Is this a problem of rounding? I even tried using some other code i
found on matlab's file exchange poly1() and roots1() but still same
false result occur. The coefficients of the resulting polynomial range
from very small 10^-40 to 10^128 which is not the case with my
polynomials which have the largest coeff at 1. Any help greatly
appreciated.

jeremy
knight427@gmail.com - 09 Mar 2006 15:07 GMT
Hi Jeremy,

Your question sounds more purely mathematical and/or programming in
nature.  I expect you might find more help at another group.  However,
I am curious how the roots of a polynomial relate to a room impulse
response.  I guess I must be dense since I'm the only one asking.

Best,
Derrick
Salmon Egg - 10 Mar 2006 07:41 GMT
On 3/9/06 7:07 AM, in article
1141916875.646000.28030@e56g2000cwe.googlegroups.com, "knight427@gmail.com"

> Hi Jeremy,
>
[quoted text clipped - 5 lines]
> Best,
> Derrick

Any function can be approximated by a polynomial. Usually that is not a good
way to go. You would probably be better off using sines and cosines to
approximate what goes on in a rectangular room. Of course, that has to be
done for every frequency.

Bill
-- Ferme le Bush
Don Pearce - 10 Mar 2006 07:44 GMT
>On 3/9/06 7:07 AM, in article
>1141916875.646000.28030@e56g2000cwe.googlegroups.com, "knight427@gmail.com"
[quoted text clipped - 13 lines]
>approximate what goes on in a rectangular room. Of course, that has to be
>done for every frequency.

Last time I looked, there were quite a few of those. ;-)

d

Pearce Consulting
http://www.pearce.uk.com
jeremyscerri@gmail.com - 11 Mar 2006 09:58 GMT
i am taking the z-transform of my room impulse response, that i why i
am ending up with a polynomial. I need to find the inverse of it since
one of my requirments is to cancel out the reverberation. For inversion
i need to make it stable first by reflecting all zeros inside the unit
circle.

jeremy

> >On 3/9/06 7:07 AM, in article
> >1141916875.646000.28030@e56g2000cwe.googlegroups.com, "knight427@gmail.com"
[quoted text clipped - 20 lines]
> Pearce Consulting
> http://www.pearce.uk.com
Don Pearce - 11 Mar 2006 10:15 GMT
>i am taking the z-transform of my room impulse response, that i why i
>am ending up with a polynomial. I need to find the inverse of it since
[quoted text clipped - 3 lines]
>
>jeremy

Your room doesn't have an impulse response. Every single point in your
room has its own impulse response when stimulated from every other
point in the room. If the stimulus is distributed, as is the case with
reverberation, there is no polynomial which solves it - you will find
no convergence.

Sorry, but this is a lost cause.

d

Pearce Consulting
http://www.pearce.uk.com
Salmon Egg - 12 Mar 2006 05:28 GMT
On 3/11/06 1:58 AM, in article
1142071101.317027.226970@p10g2000cwp.googlegroups.com,

> i am taking the z-transform of my room impulse response, that i why i
> am ending up with a polynomial. I need to find the inverse of it since
> one of my requirments is to cancel out the reverberation. For inversion
> i need to make it stable first by reflecting all zeros inside the unit
> circle.

If you are trying to actively cancel room reverberation throughout the
entire room, be prepared for failure. There are just too many modes to
control them all. I can understand possible cancellation at a small number
of points such as where microphones may be, but even then, I do not know how
you would distinguish desired signal from what you want to cancel. The best
I can offer is to use an anechoic chamber so as to eliminate reverberation
entire.

Bill
-- Ferme le Bush
jeremyscerri@gmail.com - 14 Mar 2006 17:20 GMT
yes, i agree that impulse response will be changed by any change in
geometry in room or position of mic's or speakers, what i will be doing
is measuring the impulse reponse using MLS techniques regularly.

jeremy
 
Sign In
Join
My Latest Posts
My Monitored Threads
My Blog
My Photo Gallery
My Profile
My Homepage

Start New Thread
Enable EMail Alerts
Rate this Thread



©2009 Advenet LLC   Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.